Quantitative Trait Loci and Genes Affecting Beef Tenderness

Quantitative trait loci and genes affecting beef tenderness

Lei-Yao Chang

A thesis submitted to the University of Adelaide in fulfilment of the requirement of the degree of Doctor of Philosophy

The University of Adelaide School of Animal and Veterinary Sciences May 2012

DECLARATION

I declare that this thesis is a record of original work and contains no material that has been accepted for the award of any other degree or diploma in any university or other tertiary institution to Lei-Yao Chang. To the best of my knowledge and belief, this thesis contains no material previously published or written by any other person, except where due reference is made in the text.

I give consent to this copy of my thesis, when deposited in the University Library, being made available for loan and photocopying, subject to the provisions of the

I also give permission for the digital version of my thesis to be made available on the web, via the University’s digital research repository, the Library catalogue, the

Australasian Digital Theses Program (ADTP) and also through web search engines, unless permission has been granted by the University to restrict access for a period of time.

Lei-Yao Chang

June, 2012

Table of Contents

Declaration ...... ii Table of Contents…………………………………………………………… ...iii Appendices ...... viii Index of Figures ...... x Index of Tables ...... xv Index of Models ...... xxii List of Abbreviations ...... xxiii Dedication ...... xxv Acknowledgements ...... xxi Abstract ...... xxviii

Chapter 1: Literature Review ...... 1 1.1 Introduction ...... 2 1.2 Components of muscle ...... 3 1.2.1 Muscle fibres ...... 5 1.2.2 Muscle fibre types ...... 7 1.2.3 Connective tissue ...... 8 1.3 Measuring tenderness ...... 12 1.4 Tenderisation ...... 14 1.4.1 Rigor mortis ...... 14 1.4.2 Proteolytic enzymes ...... 15 1.5 Intrinsic factors affecting tenderness ...... 19 1.5.1 Sarcomere length ...... 19 1.5.2 pH decline and ultimate pH ...... 19 1.5.3 Effect of muscle fibre types on ageing ...... 23 1.5.4 Effect of muscle bundle size on tenderness ...... 24 1.5.5 Intramuscular fat ...... 25 1.5.6 Growth rate ...... 26 1.5.7 Genetic effects ...... 26 1.6 Extrinsic factors affecting tenderness ...... 28 1.6.1 Electrical stimulation ...... 28 1.6.2 Carcass suspension and Tendercut® ...... 28 1.6.3 Exogenous meat treatments ...... 29

iii

1.6.4 Animal treatment effects ...... 29 1.7 Conclusion ...... 31

Chapter 2: Resources and trait derivation ...... 32 2.1. JS Davies Gene Mapping Cattle Project ...... 34 2.2. Tenderness traits ...... 36 2.2.1 Shear force ...... 36 2.2.1.1 Multi-trait analysis of shear force ...... 37 2.2.1.2 Results of multi-trait analysis of shear force ...... 39 2.2.2 Ageing rate ...... 42 2.2.2.1 Analysis of ageing rate ...... 43 2.2.2. 2 Results from analysis of ageing rate ...... 45 2.3 Relationships between tenderness traits ...... 49 2.3.1 Relationship between compression in ST muscle and tenderness ... 51 2.3.2 Relationship between hydroxyproline content in ST muscle and tenderness ...... 51 2.3.3 Relationship between cooking loss and tenderness ...... 52 2.3.4 Relationship between pH and tenderness ...... 52 2.3.5 Relationship between muscle weights and tenderness ...... 53 2.4. Discussion ...... 54 2.4.1 Adjusted shear force ...... 54 2.4.2 Ageing rate ...... 55 2.4.3 Related tenderness traits ...... 56 2.5. Conclusion ...... 58

Chapter 3: QTL Mapping ...... 59 3.1 Introduction ...... 60 3.2 Methods for quantitative trait loci (QTL) mapping ...... 61 3.3 Results ...... 65 3.3.1 Quantitative trait loci for adjusted shear force ...... 65 3.3.1.1 Effect of CAPN1-SNP316 and SNP530 genotypes on adjusted shear force QTL ...... 68 3.3.1.2 Effect of CAPN1-SNP316 genotype on adjusted shear force QTL ...... 71 3.3.1.3 Comparison of the QTL for adjusted shear force ...... 72 3.3.2 Quantitative trait loci for ageing rate ...... 73 3.3.2.1 Effect of MSTN F94L genotype on ageing rate QTL ...... 74 iv

3.3.2.2 Effect of MSTN F94L, CAPN1-SNP316 and SNP530 genotypes on ageing rate QTL ...... 76 3.3.2.3 Effect of MSTN F94L genotype and CAPN1-SNP316 genotype on ageing rate QTL ...... 78 3.4. Discussion ...... 78 3.4.1 Potential candidate genes for adjusted shear force ...... 88 3.4.2 Potential candidate genes for ageing rate ...... 90 3.4.3 Conclusion ...... 92

Chapter 4: Candidate Genes ...... 93 4.1 Introduction ...... 94 4.2 Materials and Methods ...... 95 4.2.1 Optimization of PCR ...... 96 4.2.2 Electrophoresis ...... 99 4.2.3 Staining ...... 99 4.2.4 Purification for PCR products ...... 100 4.2.5 Sequencing ...... 100 4.2.6 Genotyping ...... 101 4.2.7 Prediction of protein structure ...... 102 4.2.8 Genomic sequence alignments ...... 102 4.3 Results ...... 102 4.4 DNA Variants ...... 105 4.4.1 MYL7 variants ...... 106 4.4.2 MYO1G variants ...... 106 4.4.3 MBNL3 variants ...... 107 4.4.4 CAPN4 variants ...... 108 4.4.5 CAPN5 variants ...... 110 4.4.6 LOXL1 variants ...... 113 4.5. Discussion ...... 114

Chapter 5: Association Studies ...... 119 5.1 Introduction ...... 120 5.2 Methods ...... 126 5.2.1 Statistical analysis of DNA variant associations ...... 132 5.2.2 Trait summary ...... 132

5.3 Results ...... 135 5.3.1 Effects of CAPN major variants on tenderness related traits ...... 136 5.3.2 Effects of candidate gene variants on tenderness related traits ...... 140 5.3.2.1 Effects of candidate gene variants on cooking loss ...... 140 5.3.2.2 Effects of candidate gene variants on pH ...... 142 5.3.2.3 Effects of candidate gene variants on adjusted shear force . 145 5.3.2.4 Effect of candidate gene variants on ageing rate ...... 146 5.3.2.5 Effect of candidate gene variants on compression ...... 150 5.3.2.6 Effect of candidate gene variants on hydroxyproline in ST muscle ...... 151 5.3.2.7 Effect of candidate gene variants on muscle weights ...... 152 5.3.3 Effects of candidate gene variants on tenderness related traits dependent of myostatin and calpain 1 gene variants ...... 155 5.3.3.1 Effect of candidate gene variants on adjusted shear force with CAPN1 genotypes ...... 155 5.3.3.2 Effect of candidate gene variants on ageing rate with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes ...... 158 5.3.3.3 Effect of candidate gene variants on compression with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes ...... 162 5.3.3.4 Effect of candidate gene variants on hydroxyproline in ST muscle with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes ...... 165 5.3.3.5 Effect of candidate gene variants on muscle weights with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes ...... 165 5.3.3.6 Effect of candidate gene variants on cooking loss and pH with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes ...... 168 5.3.4 Simultaneous effects of candidate gene variants on tenderness ...... 172 5.4 Discussion ...... 176

Chapter 6: Gene Interactions ...... 186 6.1. Introduction ...... 187 6.2 Methods ...... 188 6.3 Results ...... 190 6.3.1 Candidate gene interactions with MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530 on tenderness traits ...... 192 6.3.1.1 Adjusted shear force ...... 192 6.3.1.2 Ageing rate ...... 197 6.3.1.3 Compression ...... 201

6.3.1.4 Hydroxyproline content of ST muscle ...... 203 6.3.1.5 Muscle weights ...... 205

6.3.2 Interactions between all candidate genes including 3 major DNA variants (MSTN F94L, CAPN1-SNP316 and SNP530) on adjusted shear force ...... 212 6.3.2.1 LD adjusted shear force ...... 213 6.3.2.2 ST adjusted shear force ...... 213 6.4 Discussion ...... 215

Chapter 7: General Discussion ...... 221 7.1 Introduction ...... 222 7.2 QTL mapping for tenderness ...... 223 7.3 Breed effects on tenderness ...... 227 7.4. DNA variants affecting tenderness ...... 229 7.5. Signalling pathways affecting tenderness ...... 233 7.6. Role of myostatin on tenderness ...... 238 7.7. Muscle specific effects on tenderness ...... 239 7.8. Collagen effects on tenderness ...... 241 7.9 Epistatic effects on tenderness ...... 241 7.10 Future research ...... 245 7.11 Conclusions ...... 246

References ...... 249

Appendices ...... 286

vii

Appendices

Appendix 1: Sample preparation and measuring tenderness ...... 287 Appendix 2: QTL mapping results for adjusted shear force without CAPN1 .. 288 Appendix 3: QTL mapping results for adjusted shear force with CAPN1-SNP316 ...... 290 Appendix 4: QTL mapping results for adjusted shear force with CAPN1-SNP316 and CAPN1-SNP530 ...... 292 Appendix 5: QTL mapping results for ageing rate without MSTN F94L and CAPN1 ...... 294 Appendix 6: QTL mapping results for ageing rate with MSTN F94L ...... 296 Appendix 7: QTL mapping results for ageing rate with MSTN F94L and CAPN1- SNP316 ...... 298 Appendix 8: QTL mapping results for ageing rate with MSTN F94L and CAPN1- SNP316 and CAPN1-SNP530...... 300 Appendix 9: QTL mapping results for adjusted shear force with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (for each family) ...... 302 Appendix 10A: QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (for each family) ...... 304 Appendix 10B: Ageing rate QTL graphs ...... 306 Appendix 11: 50x TAE buffer formula...... 308 Appendix 12: Purification kit protocol (Ultra Clean PCR Clean-up Kit, MoBio) ...... 308 Appendix 13: Purification of sequencing PCR products ...... 309 Appendix 14: PCR conditions for primer sets ...... 310 Appendix 15: Candidate genes and map locations ...... 313 Appendix 16: Candidate gene variants: locations and sequences ...... 314 Appendix 17A: Genotype numbers for single candidate gene variants ...... 315 Appendix 17B: Genotype numbers for significant pairs of candidate gene variants ...... 316 Appendix 18: Effect of candidate gene variants on shear force day 1 of ageing ...... 319 Appendix 19: Effect of candidate gene variants on shear force day 26 of ageing ...... 323 Appendix 20: Effect of candidate gene variants on shear force day 1 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes ...... 325 Appendix 21: Effect of candidate gene variants on shear force day 26 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes ...... 329

viii

Appendix 22: Effect of candidate gene variant interactions on shear force day 1 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes .. 332 Appendix 23: Effect of candidate gene variant interactions on shear force day 26 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes ...... 339 Appendix 24: Variant interactions for LD adjusted shear force ...... 344 Appendix 25: Variant interactions for ST adjusted shear force ...... 349 Appendix 26: Trait variation explained by candidate genes and interactions .. 354 Appendix 27: Additive effects of candidate genes for different traits ...... 357

Index of Figures

Figure 1.1 : Top down-view of skeletal muscle ...... 6

Figure 1.2 : Schematic diagram of the muscle filaments...... 16

Figure 1.3 : pH/temperature window model ...... 22

Figure 2.1 : Experimental design of the JS Davies Cattle Gene Mapping Project ...... 35

Figure 2.2 : Differences in ageing between muscles (LD, ST) and breeds (XJ, XL) ...... 41

Figure 2.3 : Components of variance in shear force ...... 41

Figure 2.4 : Relationship between ST (wbst_adjusted) and LD adjusted shear force (wbld _adjusted) ...... 42

Figure 2.5 : Scatter plot for adjusted shear force and ageing rate in ST muscle . 48

Figure 2.6 : Scatter plot for adjusted shear force and ageing rate in the LD muscle ...... 48

Figure 2.7 : Relationship between ageing rates of ST and LD ...... 49

Figure 3.1 : QTL on BTA 5 for adjusted shear force of LD and ST muscles, without and with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model ...... 66

Figure 3.2 : QTL on BTA 18 for adjusted shear force of LD and ST muscles, without and with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model...... 67

Figure 3.3 : QTL on BTA 29 adjusted shear force of LD and ST muscles without any calpain 1 genotypes fitted as fixed factors in the model ...... 67

Figure 3.4 : QTL on BTA 25 for adjusted shear force of LD and ST muscles, without and with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model ...... 70

Figure 3.5 : QTL on BTA 29 for adjusted shear force of the LD and ST muscles with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model ...... 70

Figure 3.6 : QTL on BTA 3 for ageing rate for the ST muscle without and with the MSTN F94L genotypes as fixed factors in the QTL model ...... 75

Figure 3.7 : QTL on BTA 11 for ageing rate for the ST muscle without and with the MSTN F94L genotypes as fixed factors in the QTL model...... 76

Figure 3.8 : QTL on BTA 13 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model ...... 77

Figure 4.1 : Location of insertion/deletion in MBNL3 intron 3 ...... 108

Figure 4.2 : Predicted protein structure for CAPN4 exon 2 ...... 110

Figure 4.3 : Predicted protein structure for CAPN5...... 112

Figure 4.4 : Genomic alignment of sequence of CAPN5 exon 7 ...... 112

Figure 4.5 : Genomic alignment of sequence of CAPN5 exon 11 ...... 113

Figure 4.6 : Predicted protein structure for LOXL1 exon 1 ...... 114

Figure 4.7 : Genomic alignment of LOXL1 exon 1 ...... 114

Figure 5.1 : Effects of CAPN1-SNP316 on LD shear force measured at 4 time points (day 1, 5, 12 and 26 post-slaughter) ...... 138

Figure 5.2 : Effects of CAPN1-SNP316 on ST shear force measured at 4 time points (day 1, 5, 12 and 26 post-slaughter) ...... 139

Figure 5.3 : Effect of CAPN5-SNP7 genotypes on LD cooking loss...... 142

Figure 5.4 : Effect of SST-SNP7 genotypes on LD cooking loss ...... 142

Figure 5.5 : Effect of CAPN5-SNP16 genotypes on LD pH ...... 144

Figure 5.6 : Effect of MYO1G-SNP2 genotypes on LD pH ...... 144

Figure 5.7 : Effect of MYO1G-SNP2 genotypes on ST pH ...... 145

Figure 5.8 : Effect of FSTL1-SNP2 genotypes on ST adjusted shear force ...... 146

Figure 5.9 : Effect of SST-SNP2 genotypes on LD ageing rate ...... 147

Figure 5.10 : Effect of MYO1G-SNP2 genotypes on LD ageing rate ...... 148

Figure 5.11 : Effect of FSTL1-SNP2 genotypes on LD ageing rate ...... 148

Figure 5.12 : Effect of LOX-SNP1 genotypes on LD ageing rate ...... 149

Figure 5.13 : Effect of FSTL1-SNP1 genotypes on LD ageing rate ...... 149

Figure 5.14 : Effect of FSTL1-SNP1 genotypes on ST ageing rate ...... 150

Figure 5.15 : Effect of FSTL1-SNP2 genotypes on ST ageing rate ...... 151

Figure 5.16 : Effect of CAPN5-SNP21 genotypes on ST muscle weight ...... 153

Figure 5.17 : Effect of IGF1-SNP1 genotypes on ST muscle weight ...... 153

Figure 5.18 : Effect of CAPN5-SNP21 genotypes on ST muscle weight ...... 154

Figure 5.19 : Effect of IGF1-SNP1 genotypes on ST muscle weight ...... 154

Figure 5.20 : Effect of SNIP1-SNP3 genotypes on LD adjusted shear force ...... 156

Figure 5.21 : Effect of MYO1G-SNP2 genotypes on ST adjusted shear force ...... 157

Figure 5.22 : Effect of FST-SNP7 genotypes on ST adjusted shear force ...... 157

Figure 5.23 : Effect of MYO1G-SNP2 genotypes on LD ageing rate ...... 158

Figure 5.24 : Effect of SNIP1-SNP3 genotypes on LD ageing rate ...... 159

Figure 5.25 : Effect of SST-SNP2 genotypes on LD ageing rate ...... 160

Figure 5.26 : Effect of CAPN5-SNP16 genotypes on LD ageing rate for LD muscle ...... 161

Figure 5.27 : Effect of FSTL1-SNP2 genotypes on LD ageing rate ...... 161

Figure 5.28 : Effect of IGF1R-SNP1 genotypes on ST ageing rate ...... 162

Figure 5.29 : Effect of IGF1-SNP1 genotypes on LD muscle weight ...... 164

Figure 5.30 : Effect of CAPN5-SNP21 genotypes on LD muscle weight ...... 165

Figure 5.31 : Effect of SST-SNP2 genotypes on LD muscle weight ...... 165

Figure 5.32 : Effect of CAPN5-SNP21 genotypes on ST muscle weight ...... 166

Figure 5.33 : Effect of SNIP1-SNP3 genotypes on ST muscle weight ...... 167

Figure 5.34 : Effect of FST-SNP7 genotypes on ST muscle weight ...... 168

Figure 5.35 : Effect of CAPN5-SNP17 genotypes on ST pH ...... 169

Figure 5.36 : Effect of CAPN5-SNP17 genotypes on LD pH ...... 170

Figure 5.37 : Effect of CAPN5-SNP17 genotypes on LD cooking loss ...... 170

Figure 5.38 : Effect of MYO1G-SNP2 genotypes on ST pH ...... 171

xii

Figure 5.39 : Effect of MYO1G-SNP2 genotypes on LD pH ...... 172

Figure 6.1 : Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on adjusted shear force of ST muscle ...... 194

Figure 6.2 : Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on adjusted shear force of LD muscle ...... 195

Figure 6.3 : Interaction of SNIP1-SNP3 and CAPN1-SNP530 genotypes on adjusted shear force of LD muscle...... 196

Figure 6.4 : Interaction of SST-SNP2 and CAPN1-SNP316 genotypes on ageing rate for LD muscle ...... 198

Figure 6.5 : Interaction of LOX-SNP1 and CAPN1-SNP316 genotypes on ageing rate for LD muscle ...... 199

Figure 6.6 : Interaction of CAPN5-SNP21 and MSTN F94L genotypes on LD weight ...... 206

Figure 6.7 : Interaction of LOXL1-SNP1 and MSTN F94L genotypes on LD weight ...... 209

Figure 6.8 : Interaction of CAPN5-SNP21and MSTN F94L genotypes on ST weight ...... 210

Figure 6.9 : Interaction of SST-SNP2 and MSTN F94L genotypes on ST weight ...... 211

Figure 7.1 : Alignment of FST-SNP5 with IL-1...... 231

Figure 7.2 : miRNA predicted from FST intron 3...... 231

Figure 7.3 Myostatin pathway and potential interactions of the candidate gene proteins ...... 233

Figure 7.4 : CBP/p300 complex and interacting proteins...... 235

Figure 7.5 : Signalling pathway of BMPs involving proteasomes ...... 236

Figure 7.6 : Formation of proteasomes with the Smad1/pro-HsN3/Az complex ...... 237

Appendix Figure 18.1 : Effect of FST-SNP7 genotypes on ST day 1 shear force ...... 321

Appendix Figure 18.2 : Effect of FSTL1-SNP1 genotypes on ST day 1 shear force ...... 322

xiii

Appendix Figure 18.3 : Effect of FSTL1-SNP2 genotypes on ST day 1 shear force ...... 322

Appendix Figure 19.1 : Effect of SNIP1-SNP3 genotypes on LD day 26 shear force ...... 323

Appendix Figure 19.2 : Effect of CAPN4-3 base repeat genotypes on ST day 26 shear force ...... 324

Appendix Figure 20.1 : Effect of FSTL1-SNP1 genotypes on ST day 1 shear force ...... 327

Appendix Figure 20.2 : Effect of FSTL1-SNP2 genotypes on ST day 1 shear force ...... 327

Appendix Figure 20.3 : Effect of MYO1G-SNP5 genotypes on ST day 1 shear force ...... 328

Appendix Figure 21.1 : Effect of SNIP1-SNP3 genotypes on LD day 26 shear force ...... 330

Appendix Figure 21.2 : Effect of LOXL1-SNP1 genotypes on LD day 26 shear force ...... 330

Appendix Figure 21.3 : Effect of FST-SNP5 genotypes on ST day 26 shear force ...... 331

Appendix Figure 21.4 : Effect of FST-SNP7 genotypes on ST day 26 shear force ...... 331

Appendix Figure 22.1 : Interaction of SST-SNP2 and CAPN1-SNP316 genotypes on LD day 1 shear force ...... 333

Appendix Figure 22.2 : Interaction of SNIP1-SNP3 and CAPN1-SNP530 genotypes on LD day 1 shear force ...... 334

Appendix Figure 22.3 : Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on ST day 1 shear force ...... 336

Appendix Figure 22.4 : Interaction of SNIP1-SNP3 and CAPN1-SNP530 genotypes on ST day 1 shear force ...... 337

Appendix Figure 23.1 : Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on LD day 26 shear force ...... 341

Appendix Figure 23.2 : Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on ST day 26 shear force ...... 341

Appendix Figure 23.3 : Interaction of LOXL1-SNP1 and CAPN1-SNP316 genotypes on LD day 26 shear force ...... 342

xiv

Index of Tables

Table 2.1 : Significant factors and variance components for shear force factors ...... 39

Table 2.2 : Least squares means for breed and MSTN F94L genotype effects... 40

Table 2.3 : Variances for shear force values of LD and ST muscles for the original and logarithmic measures ...... 44

Table 2.4 : Tests of significance for ageing rate ...... 46

Table 2.5 : Least squares means for breed and MSTN ...... 46

Table 2.6 : Tests of significance of the relationships between shear force or compression and hydroxyproline concentration in the ST muscle ...... 52

Table 3.1 : Summary of adjusted shear force and ageing rate data for LD and ST muscles ...... 65

Table 3.2 : QTL positions for adjusted shear force (across sire families) ...... 66

Table 3.3 : QTL positions for adjusted shear force with CAPN1-SNP316 and CAPN1-SNP530 genotypes in the model ...... 69

Table 3.4 : QTL positions for adjusted shear force with CAPN1-SNP316 genotypes in the model muscles ...... 71

Table 3.5 : Comparison of shear force QTL positions ...... 73

Table 3.6 : QTL positions for ageing rate (across sire families) ...... 74

Table 3.7 : Comparison of QTL for ageing rate with and without the MSTN F94L genotypes as a fixed effect in the QTL model...... 75

Table 3.8 : Comparison of QTL positions for ageing rate with the MSTN F94L and CAPN1 genotypes as fixed factors in the QTL model 76

Table 3.9 : QTL results across sire families for adjusted shear force with the QTL models 1, 2 and 3 with the known DNA variants affecting tenderness .... 80

Table 3.10 : QTL results across sire families for ageing rate with the QTL models 4, 5, 6 and 7 with the known DNA variants affecting tenderness...... 81

Table 3.11: QTL positions for adjusted shear force in each sire family ...... 83

Table 3.12: QTL positions for ageing rate in each sire family ...... 83

Table 3.13: Comparison of QTL positions herein with published QTL 86

Table 3.14: Genes within major QTL adjusted shear force ...... 88

Table 3.15: Genes within major QTL ageing rate ...... 91

Table 4.1 : DNA variants in MYL7 ...... 106

Table 4.2 : DNA variants in MYO1G ...... 107

Table 4.3 : DNA variants in MBNL3 ...... 108

Table 4.4 : MBNL3 in/del genotypes of sire 368 parents ...... 108

Table 4.5 : DNA variants in CAPN4 ...... 109

Table 4.6 : DNA variants in CAPN5 ...... 111

Table 4.7 : DNA variants in LOXL1 ...... 114

Table 5.1 : Selected candidate gene variants ...... 124

Table 5.2 : Additional candidate gene variants ...... 124

Table 5.3 : Traits analysed in association studies ...... 132

Table 5.4 : Basic statistics for all the traits...... 133

Table 5.5 : Least square means of all traits categorised by cohort, breed, sire and MSTN F94L genotype ...... 134

Table 5.6 : Relationships between CAPN1 SNPs and tenderness related traits . 139

Table 5.7 : Tests of significance for candidate gene variants on cooking loss in ST and LD muscles ...... 140

Table 5.8 : Tests of significance of candidate gene variants on pH ...... 143

Table 5.9 : Candidate gene variants associated with adjusted shear force ...... 145

Table 5.10: Tests of significance of candidate gene variants on LD ageing rate 146

Table 5.11 : Candidate gene variants associated with muscle weight...... 152

Table 5.12: Tests of significance of variants on adjusted shear force in LD and ST muscles with the effects of CAPN1 SNP316 and SNP530 in the model ...... 156

Table 5.13: Tests of significance of the candidate variants of on LD ageing rate with the genotypes of MSTN F94L, CAPN1 SNP316 and SNP530 in the model ...... 158

xvi

Table 5.14: Tests of significance of variants on LD muscle weight with the effects of MSTN F94L, CAPN1 SNP316 and SNP530 in the model ...... 164

Table 5.15: Tests of significance of candidate gene variants on ST muscle weight with the effects of MSTN F94L, CAPN1 SNP316 and SNP530 genotypes in the model ...... 166

Table 5.16: Tests of significance of variants on cooking loss and pH with the effects of MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530 in the model ..... 168

Table 5.17: Variation accounted for by candidate gene variants in mixed model ...... 175

Table 5.18: Comparison of SNPs significantly associated with tenderness traits with a size of effect of 1% or greater ...... 178

Table 5.19: Comparison of significant candidate gene variants on tenderness traits with/without the effects of MSTN F94L, CAPN1 SNP316 and SNP530 genotypes in the model ...... 181

Table 6.1 : Significant interactions between genes affecting adjusted shear force ...... 192

Table 6.2 : Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and two variants, CAPN1-SNP316 and SNP530, on adjusted shear force of ST muscle ...... 193

Table 6.3 : Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and two variants (CAPN1-SNP316 and SNP530) on adjusted shear force of LD muscle ...... 194

Table 6.4 : Proportion of the total sum of squares accounted for by the interaction between FST-SNP7 and two variants (CAPN1-SNP316 and SNP530) on adjusted shear force of LD muscle ...... 196

Table 6.5 : Proportion of the total sum of squares accounted for by the interaction between FST-SNP7 and two variants (CAPN1-SNP316 and SNP530) on adjusted shear force of ST muscle ...... 197

Table 6.6 : Significant interactions between major genes affecting ageing rate ...... 197

Table 6.7 : Proportion of the total sum of squares accounted for by the interaction between SST-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for LD muscle ...... 198

Table 6.8 : Proportion of the total sum of squares accounted for by the interaction between LOX-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for LD muscle ...... 199

xvii

Table 6.9 : Proportion of the total sum of squares accounted for by the interaction between CAPN4 -3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle ...... 200

Table 6.10: Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle ...... 200

Table 6.11 : Proportion of the total sum of squares accounted for by the interaction between LOX1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle ...... 201

Table 6.12: Proportion of the total sum of squares accounted for by the interaction between MYO1G-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle ...... 201

Table 6.13: Significant interactions between genes affecting compression ...... 201

Table 6.14: Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on compression ...... 202

Table 6.15: Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on compression ...... 202

Table 6.16: Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on compression ...... 203

Table 6.17: Significant interactions between genes affecting hydroxyproline content ...... 203

Table 6.18: Proportion of the total sum of squares accounted for by the interaction between CAPN5-SNP16 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on hydroxyproline content ...... 204

Table 6.19: Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on hydroxyproline content ...... 204

Table 6.20: Proportion of the total sum of squares accounted for by the interaction between CAPN4 3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on hydroxyproline content ...... 204

Table 6.21: Significant interactions between genes affecting muscle weight ..... 205

Table 6.22: Proportion of the total sum of squares accounted for by the interaction between CAPN5-SNP21 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight ...... 206

xviii

Table 6.23: Proportion of the total sum of squares accounted for by the interaction between IGFR1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight...... 207

Table 6.24: Proportion of the total sum of squares accounted for by the interaction between FST-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight ...... 207

Table 6.25: Proportion of the total sum of squares accounted for by the interaction between FST-SNP7 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight ...... 207

Table 6.26: Proportion of the total sum of squares accounted for by the interaction between MYO1G-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight ...... 208

Table 6.27: Proportion of the sum of total squares accounted for by the interaction between LOXL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight ...... 208

Table 6.28: Proportion of the total sum of squares accounted for by the interaction between CAPN5-SNP21 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight ...... 210

Table 6.29: Proportion of the total sum of squares accounted for by the interaction between SST-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight ...... 211

Table 6.30: Proportion of the sum of squares accounted for by the interaction between FST-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight ...... 212

Table 6.31: Proportion of the sum of squares accounted for by the interaction between CAPN4-3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight ...... 212

Table 6.32: Proportion of the sum of squares accounted for all significant interactions ...... 214

Table 6.33: Comparison of the total of the proportions of the sum of the square of all the significant individual SNPs for the traits with that of the significant interactions ...... 219

Table 7.1 : Number of animals with CAPN1-SNP361 genotypes for each backcross breed ...... 228

Table 7.2 : Number of animals with MSTN F94L genotypes for each backcross breed ...... 229

xix

Table 7.3 : MicroRNA and microRNA binding sites in the 3’UTR of the human CAPN1 gene ...... 232

Appendix Table 18.1 : Tests of significance for candidate gene variants on shear force at day 1 ageing of the ST muscle ...... 319

Appendix Table 20.1 : Tests of significance of candidate gene variants on ST shear force at day 1 ageing ...... 326

Appendix Table 22.1 : Significant interactions between genes affecting wbld1 and wbst1 ...... 332

Appendix Table 22.2 : Proportion of the total sum of squares accounted for by the interaction between SST-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 1 shear force ...... 333

Appendix Table 22.3 : Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 1 shear force ...... 335

Appendix Table 22.4 : Proportion of the total sum of squares accounted for by the interaction between CAPN4 -3 base repeat and 3 variants (MSTN F94L, CAPN1- SNP316 and SNP530) on ST day 1 shear force ...... 335

Appendix Table 22.5 : Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 1 shear force ...... 336

Appendix Table 22.6 : Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 1 shear force ...... 338

Appendix Table 22.7 : Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 1 shear force ...... 338

Appendix Table 23.1 : Significant interactions between genes affecting LD and ST day 26 shear force (wbld26 and wbst26) ...... 339

Appendix Table 23.2 : Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 26 shear force ...... 340

Appendix Table 23.3 : Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 26 shear force ...... 340

Appendix Table 23.4 : Proportion of the total sum of squares accounted for by the interaction between LOXL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 26 shear force ...... 342

Appendix Table 23.5 : Proportion of the total sum of squares accounted for by the interaction between FST-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 26 shear force ...... 343

Appendix Table 23.6 : Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 26 shear force ...... 343

xxi

Index of Models

Chapter 2 Trait Models Model 1: Model with no fixed effects to show the original effects on the traits 50

Model 2: Model with no fixed effects to show the original effects on the traits 50

Chapter 3 QTL Models QTL models 1-7 64

Chapter 5 Association Studies Models Model 3: Model with fixed effects to adjust for cohort, breed and sire 126

Model 4: Model with fixed effects to adjust for cohort, breed, sire and the known tenderness related variants myostatin F94L and calpain 1 (CAPN1-SNP316 and SNP530) 127

Model 5: Model with no fixed effects to show the original effects of variants on adjusted shear force 128

Model 6: Model with fixed effects to adjust for the known tenderness related variants calpain 1 (CAPN1-SNP316 and SNP530) for adjusted shear force 128

Model 7: Model with all fixed effects to estimate the additive and dominance effects 129

Model 8: Model with all fixed effects and the known tenderness related variants myostatin F94L and calpain 1 (CAPN1-SNP316 and SNP530) to estimate the additive and dominance effects 129

Model 9: Model to estimate the additive and dominance effects for adjusted shear force 130

Model 10: Model with calpain 1 effects to estimate the additive and dominance effects for adjusted shear force 131

Model 11: Model with all variants included simultaneously 131

Chapter 6 Gene Interaction Models Model 12: Model to identify effects of interactions between candidate genes on adjusted shear force 188

Model 13: Model 12: Model to identify effects of interactions between candidate genes on other tenderness traits 188

Model 14: Model to identify effects of interactions between candidate genes and major genes on adjusted shear force 189

xxii

List of Abbreviations

adjusted shear force = best linear unbiased prediction which is effectively an average of shear force values (kgF) taken during 4 ageing days (1, 5, 12 and 26 days); derived from a multi-variate mixed model in which key effects were accounted ageing rate = amount of ageing in ln kg per 25 days; calculated as the difference between natural log shear force values after 1 and 26 days ageing bp = base pairs

BTA = cattle chromosome

BLUP = best linear unbiased prediction

CAPN1 = calpain 1; 80-kDa subunit for µ-calpain protease (by convention, italics = gene, no italics = protein) clld = cooking loss (%) measured in LD muscle clst = cooking loss (%) measured in ST muscle cM = centiMorgan; genetic map distance based on recombination rate where 1% = 1 cM dNTPs = deoxyribonucleotide triphosphates

HRM = high resolution melt; genotyping method in/del = insertion/deletion; DNA variants involving nucleotide addition or removal

LD = M. longissimus dorsi muscle

MSA = Meat Standards Australia

Mb = megabases; one million bases

MSTN = myostatin; negative inhibitor of muscle growth (by convention, italics = gene, no italics = protein)

PCR = polymerase chain reaction; method of amplifying specific DNA segments pHld = pH measured in LD muscle averaged across 4 ageing times pHst = pH measured in ST muscle averaged across 4 ageing times

xxiii

QTL = quantitative trait loci; region of genome controlling trait

QTL peak = QTL maximum; location with highest probability of linkage

QTN = quantitative trait nucleotide; DNA variant generating QTL

SE = standard error

SNP = single nucleotide polymorphism; single base change with frequency > 1%

ST = M. semitendinosus muscle

ST_Compression = compression measurements (kgF) of ST muscle

ST_Hydroxyproline = hydroxyproline content of ST muscle (mg/g); measure of collagen

STR = simple tandem repeat; short tandem of less than 10 bases, also known as microsatellites wbld_adjusted = BLUP adjusted shear force (kgF) for LD muscle derived from a multi-variate mixed model in which key effects were accounted wbst_adjusted = BLUP adjusted shear force (kgF) for ST muscle derived from a multi-variate mixed model in which key effects were accounted wbld1 = LD Warner-Bratzler shear force (kgF) at day 1 of ageing wbld26 = LD Warner-Bratzler shear force (kgF) at day 26 of ageing wbst1 = ST Warner-Bratzler shear force (kgF) at day 1 of ageing wbst26 = ST Warner-Bratzler shear force (kgF) at day 26 of ageing

xxiv

DEDICATION

I dedicate this work to my wife, Tzu, and my parents for their unlimited support with love and patience during the whole journey. Completing this research was not an easy job. Not only did I need to become accustom to the way of studying in Australia, but also we had to abandon what we had already had. My wife, Tzu, sacrificed her own job and had to live with me in a peaceful but monotonous place - a place with many lovely people but without suitable jobs for her. This work was done by Tzu and me. I could not get through the journey without her help. Therefore, I would like share this thesis with her.

xxv

ACKNOWLEDGEMENTS

I would like to sincerely thank my supervisors, Associate Professors Cynthia

Bottema and Wayne Pitchford. Both of them not only led me in the right way of doing the research but also were the models of how to be a scientist. During my candidature, Associate Professor Cynthia Bottema offered me infinite help and patience in studying and living. She kept my direction on the right track and forgave all the mistakes I made. Associate Professor Wayne Pitchford carved many of concepts in my heart/brain. I will never forget all those precious words from my supervisors/mentors. Without my supervisors’ help, I could not have completed the whole process. I owe you a great debt of gratitude.

I would like to acknowledge Zbigniew Kruk and David Lines for measuring the tenderness related traits and initial QTL mapping results. Thanks to Ali Esmailizadeh

Koshkoih for providing the QTL mapping results as well. Madan Naik and Rugang

Tian for helping me become productive in the lab. Thanks to Andrew Egarr and his family for teaching me English and inviting us to his house for Christmas every year.

Thanks to Irida Novianti and Nadiatur Akmar Zulkifli for all the joyful time we had shared. Specifically, another gentleman, Dr Graham Webb, whom I would like acknowledge. I learned wisdom and sincerity in science from him. Too many people

I would like to say thank you. Audrey Stratton, Lesley Menzel, Liem Mahalaya and

Jane Copeland accompanied me and gave me much help and friendship. I cannot mention everyone I knew. However, I would like you to know that I deeply appreciate your kindly and warm consideration. Because of you all, Tzu and I had a wonderful and unforgettable studying journey in Australia. Lastly, the J.S. Davies

Bequest funding of the cattle gene mapping project is greatly appreciated. Thanks to

xxvi

Beef CRC for providing excellent opportunities to foster new scientists. Thanks to the University of Adelaide and Australia, you made me feel Australia is just like another home. Thank you very much.

xxvii

ABSTRACT

Tenderness is one of the major meat quality factors that affects the intent of consumers to re-purchase beef. Both genetic and non-genetic factors affect the quantitative trait of tenderness. Among the genetic factors, polymorphisms in key genes, such as the myostatin (MSTN) and calpain 1 (CAPN1), play important roles on tenderness. However, these genes do not explain all the genetic variation associated with tenderness. The aim of this study was to discover additional genes associated with tenderness to help integrate genetic information into beef cattle breeding programmes and meat quality assurance programmes, such as Meat

Standards Australia, and produce high quality tender meat for consumers. Discovery of such genes should also aid in the understanding of mechanisms underlying tenderness.

Backcross QTL mapping progeny based on crosses between two extreme Bos taurus breeds (Limousin and Jersey) were used in the study. There were four new traits created for the QTL mapping and association studies. Two of the traits

(wbld_adjusted and wbst_adjusted) were based on Warner-Bratzler (WB) shear force measurements from the M. longissimus dorsi (LD) and M. semitendinosus (ST) muscles and were derived from a multi-variate mixed model in which the environmental effects, myostatin F94L genotype effect, ageing day effect and the interaction effects were accounted for. The adjusted shear force traits offered a more accurate prediction for average tenderness. The other new trait was the amount of ageing per 25 days (called “ageing rate” herein) for the two muscles, calculated as the difference between natural log shear force values after 1 and 26 days ageing.

xxviii

Quantitative trait loci (QTL) mapping for these traits indicated there were 2 QTL (92 cM on BTA 5 and 52 cM on BTA 29) for adjusted shear force of the LD muscle, 3

QTL (96 cM on BTA 5, 36 cM on BTA 18 and 52 cM on BTA 29) for adjusted shear force of the ST muscle, 2 QTL (40 cM on BTA 4 and 0 cM on BTA 13) for ageing rate of the LD muscle and 2 QTL (48 cM on BTA 1 and 44 cM on BTA 19) for ageing rate of the ST muscle.

Twelve candidate genes were selected for further study based on their physiological functions and the QTL mapping results from herein and elsewhere. Twenty DNA variants in these candidate genes were chosen for the association studies. The analyses were conducted with and without three known tenderness related gene variants (MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530). Variants in the candidate genes were discovered to be significantly associated with traits related to tenderness, most of which were muscle specific effects. Of note, the effects of

CAPN1-SNP316 were muscle specific. The heterozygous genotype (GC) of CAPN1-

SNP316 had the opposite effect on LD and ST muscles in that the G allele was dominant for the LD but recessive for the ST. Another variant of large effect,

MYO1G-SNP2 (myosin 1G), showed an effect on ageing rate of the LD muscle but not the ST muscle.

Importantly, however, the interactions between gene variants frequently explained more of the genetic variation than the individual variants. For example, the interaction between the candidate gene variant SNIP1-SNP3 (Smad nuclear interacting protein 1) and the CAPN1-SNP316 explained more of the variation in the

xxix adjusted shear force of the ST muscle than CAPN1-SNP316 alone (9.5% vs. 5.2%).

The studies also suggest that tenderness is not always affected by the genes that change the muscle weight or collagen content (eg. insulin-like growth factor 1). In fact, the results indicate that the effect of the myostatin gene on tenderness is not caused by the increased muscle mass or collagen changes associated with the myostatin F94L variant. Instead, most of the effect of myostatin on tenderness may be explained by a change in the muscle fibre types which affects calpain activity.

xxx

Chapter 1: Literature Review

1.1. Introduction

Meat eating quality traits include visual quality, flavour and texture. It is common for beef quality studies to assess tenderness, juiciness, flavour and overall satisfaction. However, tenderness is the major sensory factor of meat and the key determinant associated with eating quality (Boleman et al., 1995; Smith et al., 1995;

Issanchou, 1996; Miller et al., 2001; Platter et al., 2003). It is the most important attribute of beef with regards to the intent of consumers to re-purchase (Issanchou,

1996; Alfnes et al., 2008).

Tenderness is a quantitative trait that is affected by genetics, environmental factors and their interactions (Takahashi et al., 1987; Goll, 1991; Koohmaraie, 1994;

Koohmaraie, 1996; Muir et al., 1998; Lee et al., 2000; Sorheim, 2001; Devine et al.,

2002; Nuernberg 2005; Dikeman, 2007). It is not possible to measure tenderness in the live animal, so exploiting genetics to improve tenderness is problematic. However, discovering the genes that affect tenderness will allow producers to improve meat quality through breeding selection programs. Mapping the genomic location of quantitative trait loci (QTL) for specific traits related to tenderness is the first step towards identifying these genes.

Turning muscle into meat can be roughly divided into two processes, rigor mortis and tenderisation (or ageing). Both tenderisation and rigor mortis play important roles in beef tenderness (Smulders et al., 1990; O’Halloran et al., 1997). However, the effect of rigor mortis on tenderness only occurs during the first 24 hours after slaughter. Hence, rigor mortis results in meat toughness directly after slaughter. After

24 hours, tenderisation mainly determines the toughness of meat as meat is typically aged for at least several days prior to purchase by consumers (Taylor et al., 1995;

Koohmaraie and Geesink, 2006). Understanding these processes and their relation to the components of muscle is helpful in discovering important genes that may affect the tenderness of beef.

1.2 Components of muscles

Tenderness is considered to be a function of three muscle-related factors, namely sarcomere length, connective tissue content and the extent of proteolysis of the myofibrillar proteins (Koohmaraie et al., 2002; Hwang et al., 2004). The specific traits related to tenderness are a result of these components. In some instances, the pathways controlling these components are well understood. For example, the calpain and lysosomal proteases have been shown to be the key enzymatic systems that influence tenderisation (Koohmaraie, 1994 and 1996). Moreover, it is known that the effect of connective tissue on beef tenderness is related to the type and quantity of connective tissue and that there are differences in connective tissues between cattle breeds (Norman, 1982). Other factors such as muscle bundle size, pH and muscle type have also been shown to affect tenderness (Cross et al., 1973; Ouali,

1992; Maltin et al., 1997).

The largest portion of meat consumed is skeletal muscle. Skeletal muscle is different from the other two types of muscles, smooth muscle and cardiac muscle. The structure and function of skeletal muscle has been described in detail (Alberts et al.,

1983; Goll et al., 1984). Skeletal muscles are comprised of approximately 75% water,

20% protein, various amounts of lipid and other small components (Lawrie, 1985).

Skeletal muscles contain mainly non-dividing, multi-nucleated muscle fibres that contract to generate force (Alberts et al., 1983).

Skeletal muscle is formed during embryogenesis from myoblasts, the muscle precursor cells (Bryson-Richardson and Currie, 2003). The embryonic myoblasts differentiate into primary muscle fibres and the foetal myoblasts differentiate into secondary muscle fibres (Bryson-Richardson and Currie, 2003; Picard et al., 2002).

The mononucleate myoblasts become post-mitotic and fuse into multinucleate myotubes during the differentiation (Brameld et al., 1998). The myotubes further develop into mature muscle fibres. The total number of muscle fibres is believed to be established by the end of the second trimester of gestation in ruminants (Picard et al., 2002). Hence, postnatal growth of muscle fibres mainly involves increases in fibre size (hypertrophy) rather than numbers of fibres (Rehfeldt et al., 2004).

Muscle stem cells, known as satellite cells, are also derived from myoblasts and are responsible for muscle growth after birth (Relaix and Marcelle, 2009; Cossu and

Biressi, 2005). Satellite cells are located beneath the basal lamina that develops around each fibre (Cossu and Biressi, 2005) and differentiation in the neonatal period involves the fusion of satellite cells with the existing muscle fibres (Brameld et al., 1998). At the end of post-natal growth, satellite cells are static, but they can be activated if the muscle tissue is damaged (Cossu and Biressi, 2005). They resume mitotic activity and repair damaged fibres or fuse with other satellite cells to form new ones (Hess and Rosener, 1970; Cossu and Biressi, 2005).

Thus, post-natal skeletal muscle consists of two muscle cell types, myocytes and satellite cells (Harper 1999). The myocytes are the mature, multinucleated contractile muscle cells. Each myocyte contains chains of sarcomeres that form myofibrils, the contractile units of the cell (Harper 1999). The satellite cells are the small

4 mononuclear stem cells that differentiate and develop into myocytes during growth or repair (as described above, Cossu and Biressi, 2005).

In addition, there are 2 other major cell types associated with skeletal muscle cells, fibroblasts and pre-adipocytes (Harper, 1999). Fibroblasts are the origin of the connective tissues, namely epimysium, perimysium and endomysium (Linder et al.,

1976). The pre-adipocytes within the muscle differentiate into mature adipocytes, which store fat within the muscle (Park et al., 2008).

1.2.1. Muscle fibres

Each myocyte represents a muscle fibre which is composed of myofibrils and surrounded by the endomysium (Goll et al., 1984; Harper, 1999). The muscle fibre is also called as the primary muscle bundle (Rowe, 1981). Many primary muscle bundles are surrounded by the perimysium to form a fascicle. Large numbers of fascicles become a secondary muscle bundle, which is covered by the epimysium

(Figure 1.1.). Approximately 60% of the dry mass of a primary muscle bundle is comprised of myofibrils (Harper, 1999). The myofibrils are a complex of proteins and cylindrical organelles that are responsible for locomotion (Goll et al., 1984;

Robson, 1995). The two main proteins in the myofibril are myosin and actin (Figure

1.1). Myosin forms a thick filament held in position by titin filaments. Actin forms a thin filament held in position by nebulin filaments (Robson et al., 1991). The Z disks consisting of two type of titin (titin and α-actinin) are the lateral boundaries of a sarcomere (Young et al., 1998). The sarcomere is the basic unit of the muscle movement. In each sarcomere, there are also a number of accessory proteins, such as troponin (which prevents the attachment of myosin to actin), desmin (which is located on the two sides of the Z-disk and links the myofibrils) and α-actinin (which

5 forms the Z disk) (Figure 1.2) (Robson, 1995). These proteins form the main structure of muscle, and therefore, have been considered as potentially involved in tenderness (Koohmaraie et al., 2002).

A NOTE: This figure/table/image has been removed to comply with copyright regulations. It is included in the print copy of the thesis held by the University of Adelaide Library.

Figure 1.1 Top down-view of skeletal muscle (Adopted from Campbell and Reece, th 2005 'Biology' 6 ed.)

1.2.2. Muscle fibre types

Muscle fibre type is one of the major determinants of skeletal muscle function. Fibre type can be determined by three major characteristics: contraction, metabolism and immunology. Contraction is based on the myosin heavy chain. Myosin heavy chains form thick filaments within muscle fibres. According to the shortening speed, the muscle fibre types are classified as either the slow-twitch type (Type I) or the fast- twitch type (Type II) (Brooke and Kaiser, 1970; Schiaffino and Reggiani, 1996).

Muscle fibre type, thus, depends upon which form of myosin heavy chain is present.

In terms of metabolism, the aerobic oxidative capacity of these fibre types differs.

Gauthier (1969) categorized the fibre types as red, intermediate and white types by measuring the quantity of succinate dehydrogenase (SDH) present, a principal enzyme in the citric acid cycle located in the inner mitochondrial membrane. The red and intermediate types are considered oxidative and the white type is glycolytic.

Measuring ATPase activity is another widely used method for classifying muscle fibre types (Moody et al., 1980; Stienen et al., 1996; Christensen et al., 2006).

Normally, staining for ATPase in myosin distinguishes the muscle fibres into the two types (Type I and II). However, Brook and Kaiser (1970) modified the ATPase method to further classify Type II fibres into three sub-classes (Type IIA, Type IIB and Type IIC).

Immunological classification methods have been also utilized. The methods are based upon using antibodies to target specific isoforms of the myosin heavy chain to classify the fibre types (Brooke and Kaiser, 1970). There are three types that have been identified using myosin antibodies: HMC-IIA, HMC-IIB and HMC-IIX

(Schiaffino et al., 1989; Gorza, 1990).

Muscles are not composed of a single type of fibre. Gauthier (1969) very early demonstrated that red muscle of semitendinosus muscle consists of 52% red fibres and 40% of intermediate fibres. Generally, however, darker or red muscles contain a greater proportion of slow-twitch fibre type (Type I) due to the binding of oxygen with myoglobin and the presence of many mitochondria (Beecher et al., 1965;

Gauthier, 1969). Muscles that contain a greater proportion of white fibres are the fast-twitch (Type II) due to the lack of myoglobin and the use of anaerobic metabolism to produce energy (Beecher et al., 1965; Gauthier, 1969). Nevertheless, the so-called “white” muscle M. semitendinosus still comprises 82% red fibres

(Gauthier 1969).

1.2.3. Connective tissue

Cross et al., (1973) demonstrated that connective tissue is highly correlated with tenderness. Connective tissue originates from fibroblasts and consists of three major types of proteins: collagen, elastin and reticulin. Collagen is the most abundant protein in the connective tissues of skeletal muscles. Tropocollagen is the subunit of the collagen fibril. The mature collagen needs to be oxidised by the enzyme lysyl oxidase (LOX) to form the polymer collagen fibril (Bailey, 1984).

Collagen is a general term for the proteins that have a characteristic triple helix of three polypeptide chains (Eyre et al., 1992; Sato et al., 2002). Thus far, 26 collagen types have been described in mammals (Kuhn, 1986; Sato et al., 2002). Collagen

Types I (55-75%) and III (25-45%) are the major connective tissue proteins in the epimysium and perimysium. Collagen Type VI is the main component in endomysium (Bateman et al., 1996). The size of collagen ﬁbrils is about 65–67 nm in the perimysium and 47–48 nm in the endomysium (Fang et al., 1999; Rowe,

1974). Other collagen types are important components of the intramuscular connective tissues (Listrat et al., 1999; Listrat et al., 2000).

The endomysial, epimysial and perimysial collagen in the meat causes shrinkage at

52-64℃. It is this shrinkage of collagen that affects the meat and influences the meat texture (Findlay et al., 1986; Kuypers and Kurth 1995). Collagen can explain much of the variation in the toughness of different muscles. Gerrard, et al. (1987) showed that the collagen content decreases with the age of the cattle and varies between different muscles. For example, the collagen content of the semitendinosus muscle is higher than the longissimus dorsi muscle in cattle.

Bendall (1967) found that the content of collagen in beef muscle can vary from 1% to 15% of dry weight compared to 0.6% to 3.7% elastin content. In addition, the variation in the content of collagen dry weight in the perimysium is greater than the variation in the endomysium (Light et al., 1985; Purslow, 1999). The percentage of

Type III collagen can be changed by the type of diet. Listrat et al., (1999) showed that animals fed with hay had a higher proportion of Type III collagen than animals fed with grass silage. The proportion of Type I collagen was unchanged.

Collagen Type III is more heat stable than collagen Type I (Burson and Hunt, 1986).

Consequently, collagen Type III was believed to play a larger role in tenderness as most collagen Type I would be destroyed during cooking. However, there has been controversy over the role of collagen of Type III in toughness (Bailey et al., 1979;

Light et al., 1985; Burson and Hunt, 1986). A difference in the relationship between the collagen Type III content of raw versus cooked meat with tenderness has been found. The correlation between the total collagen in raw meat of cattle and toughness is high and ranges from r = 0.72-0.95 (Dransfield et al., 2003; Torrescano et al.,

2003). However, for cooked meat, the correlation coefficients between total collagen content and tenderness are as low as r = 0.20-0.46, (Ngapo et al., 2002; Dransfield et al., 2003). The results indicate that the collagen thermal stability may not be important for tenderness except in raw meat.

The results of Stolowski et al., (2006) showed that the content of total collagen varies with shear force and the soluble collagens are negatively associated with shear force in different muscles. Collagen solubility has been shown to vary between breeds (Stolowski et al., 2006). However, Riley et al., (2005) found that total collagen is not significantly associated with tenderness but that the insoluble collagen is related. The different conclusions may be a consequence of the fact that different breeds and different muscles were examined in these two studies. Overall, the results suggest it is the total amount of collagen as well as the type of collagen that may influence tenderness or toughness depending upon the breed and muscle types examined.

There is at least one example of cattle genetics that may affect connective tissue and tenderness. McPherron et al., (1997) showed that if the gene for myostatin (MSTN), a negative regulator of muscle development, is mutated in mice such that the myostatin protein is no longer functional, then mice that are homozygous for the knock-out mutation have greatly increased muscle mass. Similarly, cattle that are homozygous for a deletion in the MSTN gene will be “double muscled” and also have greatly increased skeletal muscle mass (Bellinge et al., 2005). The meat from the “double muscle” cattle is considered to be very tender (Albrecht et al., 2006).

Albrecht et al., 2006; and Boccard (1981); reported that “double muscle” cattle have less connective tissue per gram of muscle than normal cattle. Esmailizadeh et al.,

(2008) and Lines et al., (2009) demonstrated a significant effect of an intermediate myostatin variant, MSTN F94L, on tenderness in both the ST and LD muscles. The variant was also associated with decreased collagen content in the ST muscle (Lines et al., 2009). Although postulated (Bellinge, et al., 2005; Warner, et al., 2010), it has not yet been established whether the change in connective tissue, as a result of these myostatin mutations, is the actual cause of the increased tenderness.

Tropocollagen molecules need cross-links to stabilize the structure of connective tissues. The existence of cross-links increases the level of thermal stability in meat

(Bailey, 1984). The side chain of peptidyl lysine residues in the tropocollagen are converted to α-aminoadipic-δ-semialdehyde (AAS) by the enzyme lysyl oxidase

(LOX). AAS and the other lysine residues of the side chain form a series of cross- links called dehydrolysininirleucine (deLNL). Two AAS molecules can also link to each other to form another cross-link called the aldol condensation product (ACP)

(Lucero and Kagan 2006).

In total, five mature cross-link types have been discovered (Ngapo et al., 2002). The cross-link that bridges two collagen molecules is called a divalent cross-link and includes hydroxylysinor leucine (HLNL), dihydroxylysinorleucine (DHLNL) and histidinohydroxymerodesmosine cross-links (HHMD). A cross-link that involves three collagen molecules is called a trivalent cross-link and includes pyridinoline and

Ehrlich chromogen (EC). The bond in a trivalent cross-link is 1.5 times greater than that of a divalent one (Erman and Mark, 1997).

Of the five types of mature cross-links (Ngapo et al., 2002), the amount of pyridinoline and Ehrlich chromogen cross-links is not reduced after meat is cooked at 80℃ for 45 minutes. DHLNL links are also thermostable. However, HLNL and

HHMD links are totally destroyed when meat is cooked at 73℃ for 10 min (Allain et al., 1978; Horgan et al., 1990, 1991). Ngapo et al., (2002) compared the number of cross-links between Belgian Blue normal, heterozygous and homozygous double- muscled cattle. A different number of cross-links were observed between the cattle with different myostatin genotypes. However, the tenderness of the cooked meat was not associated with the concentration of cross-links. The data suggest that although the cross-links stabilise the connective tissue, the number of collagen cross-links were significant at least in the Belgian Blue.

1.3 Measuring tenderness

Many methods have been used to measure tenderness or toughness. The subjective tests for tenderness are based on the chewing and eating by taste panellists. The objective methods utilised for measuring tenderness or toughness are more diverse and include techniques such as myofibrillar fragmentation index (Culler et al., 1978), myofibrillar protein solubility (Claeys et al., 1994), soluble and total collagen content (Culler et al., 1978), UV-fluorescence spectrophotometry (Swatland, 1992), image analysis (Shackelford et al., 1998), tenderness probes (Jeremiah and Phillips,

2000), dual energy X-ray absorptiometry (Kroger et al., 2006), Warner-Bratzler shear force (WBSF) measurements (Boccard et al., 1981) and compression analysis

(Lepetit and Culioli, 1994).

Soluble and total collagen content measurements and UV-fluorescence are used to assess the connective tissue content of meat. In many studies of tenderness, Warner-

Bratzler shear force, which measures the force required to cut through meat (Morgan et al., 1991; Koohmaraie, 1994; Powell et al., 2000; Ruiz de Huidobro et al., 2005), has been used as an indicator of tenderness by imitating biting. Shear force

12 supposedly reflects all effects on tenderness simultaneously (Boccard et al., 1981).

Warner-Bratzler shear force is measured by measuring the resistance (kg F) of shearing muscle cubes with a blade. Although the Warner-Bratzler shear force method has been standardised to give an accurate prediction (Bekhit, 2003), the within-muscle variation in tenderness is not well explained (Gariepy et al., 1990).

Shackelford et al., (1997) found that different parts of a muscle give different values for Warner-Bratzler shear force. This variation is likely to be caused by the different muscle fibre types within a cut of meat.

Although Warner-Bratzler shear force has become the most widely used method to provide an objective prediction of tenderness, shear force does not perfectly represent the subjective sensory measures of tenderness. The correlation between the shear force and the sensory panel scores varies from -0.32 to -0.94 (Caine et al.,

2003). The variation may be caused by the cooking methods, number of ageing days, muscle type, equipment and sensory test panellists (Miller et al., 2001; Caine et al.,

2003). Fortunately, tender meat can still be distinguished from tough meat by simply using a shear force threshold of 4.3~4.07 kg/F (Destefanis et al., 2008; Rodas-

González et al., 2009). In general, consumers and trained testers recognise meat as being tender when the shear force is under 4.3~4.07 kg/F.

In addition to shear force, compression is another objective method used to imitate eating behaviour. This measurement compresses meat at a specific location twice to simulate the “chewiness” of the meat. Compression is correlated with the effect of collagen (content and extent of cross-linking) in meat (Bouton and Harris 1972;

Bailey et al., 1998). Lepetit and Culioli (1994) found that compression explains 36% of the variation in consumer sensory evaluations. 13

Kroger et al., (2006) developed a new measuring system for tenderness using dual energy X-ray absorptiometry (DEXA) to provide a better assessment of tenderness.

The dual energy X-ray absorptiometry method uses the colour of digital images to determine the level of tenderness. However, dual energy X-ray absorptiometry has not produced better results than shear force (Kroger et al., 2006). Thus, there is still room for improvement in objective methods to measure the tenderness of meat that is comparable to the subjective sensory measures of tenderness.

1.4 Tenderisation

After slaughter, beef is normally aged (that is, stored at low temperatures for a period of time). Ageing is a result of the activity of enzymes within muscle and improves the eating quality by improving tenderness. This process of ageing or tenderisation involves the resolution of rigor mortis and protein degradation.

1.4.1 Rigor mortis

Rigor mortis occurs in the muscle within 24 hours post-mortem (Wheeler and

Koohmaraie, 1994; Goll et al., 1995). Rigor mortis is one of the major causes of sarcomere length shortening (Goll et al., 1995). This shortened sarcomere length causes tough meat because of the condensation of the muscle structure (Lawrie,

1992).

Rigor mortis of meat is caused by the stoppage of contraction. Contraction requires several components including ATP, creatine phosphate, glycogen, myofibrils, calcium, and the activities of several enzymes including ATPases, kinases and glycolytic enzymes (Suyama and Konosu, 1987). The energy for contraction is derived from the degradation of glycogen and creatine phosphate by glycolytic enzymes and creatine kinase, respectively. The degradation provides ATP for myosin

(thick filament) to form a charged myosin which has the ability to attach to actin

(thin filament). However, the troponins on the actin can prevent a charged myosin head from binding with myosin-actin site. Calcium released from the sarcoplasmic reticulum can bind with the troponins and remove them from this original site to facilitate the myosin head locking to actin. This forms the actin myosin complex and swivelling of the complex causes filament movement and consequent contraction.

The release of the myosin head binding with actin requires ATP from glycogen and creatine phosphate degradation (Squire and Morris, 1998).

Post-mortem, energy is produced by the degradation of glycogen in muscles (Bendall,

1973; Suyama and Konosu, 1987). This anaerobic metabolism produces lactic acid, which lowers the pH. The process is affected by the levels of glycogen at death, the rate of ATP turnover and muscle metabolism (Valin et al., 1993). When all the ATP has been depleted, the actin myosin complexes cannot separate. Thus, rigor mortis will occur and shortened sarcomeres will remain.

1.4.2 Proteolytic enzymes

The muscle proteolytic systems are the key mechanisms underlying tenderisation during ageing (Ouali, 1990). The three major proteolytic systems that have been widely discussed in terms of their potential roles and contribution to tenderness during the ageing process include the calpain, lysosomal, proteasomal, and caspases pathways.

The calpain system is recognized as the most important post-mortem factor affecting tenderness in beef (Goll, 1991; Koohmaraie, 1994). It has been shown that the greater the concentration of calpain in the muscle, the faster rate of ageing (Dayton et al., 1976). The calpain system consists of three proteins (µ-calpain, m-calpain and

15 calpastatin) and is a Ca2+-dependent process. The µ-calpain and m-calpain consists of an identical 28-kDa subunit and an 80-kDa subunit in which there is only a 55-

60% of amino acid homology between the two proteases (Goll et al., 2003). The 80- kDa subunits for µ-calpain and m-calpain were encoded by two individual genes,

CAPN1 and CAPN2 (Yoshimura et al., 1983; Ohno et al., 1990; Huang and Wang,

2001). The 28-kDa subunit for two calpains is encoded by single gene, called

CAPNS1 or CAPN4 (Ohno et al., 1990). CAPN3 is encoded by the CAPN3 gene and is a tissue specific isoenzyme that binds and affects the stability of tintin in the sarcomere (Ohno et al., 1990; Sorimachi et al., 1995; Huang and Wang, 2001). In addition to the CAPN1 - CAPN4 genes, there are other 8 calpain-like cysteine protease genes (CAPN5-CAPN12) that have been identified (Braun et al., 1999; Dear and Boehm, 1999; Franz et al., 1999; Dear et al., 2000; Carlsson et al., 2005; Davis et al., 2007).

Calpains are located in the sarcomere where the calpain catalysed hydrolysis occurs, mainly fragmenting desmin, nebulin and filamin in the Z-lines, troponin T, tropomyosin, M-line proteins and connectin (King, 1984; Penny et al., 1984) (Figure

1.2). The optimum pH range for calpain activity is 6.5-7.5 (Asghar and Bhatti 1987).

A NOTE: This figure/table/image has been removed to comply with copyright regulations. It is included in the print copy of the thesis held by the University of Adelaide Library.

Figure 1.2 Schematic diagram of the muscle filaments. (From Goll et al., 2008.)

Calpastatin is encoded by a single gene. It is an inhibitor of µ-calpain and m-calpain

(Ouali and Talmant, 1990; Koohmaraie et al., 1991; Bishop et al., 1993; Jiang et al.,

1992, 1994, 1996). The size of calpastatin can vary from 30 to 300-kDa, with the size altered by different promoters and splicing (Goll et al., 2003). There are at least

4 isoforms of calpastatin in most species. The purpose of these isoforms and the mechanism of µ-calpain and m-calpain inhibition by calpastatin are still not well understood.

There is some debate about which factor, calpain, calpastatin or the calpain- calpastatin ratio, is the best indicator of the potential for ageing (Calkins and

Seidman, 1988; Johnson et al., 1990; Ouali and Talmant, 1990; Shackleford et al.,

1991; Ouali, 1992). Koohmaraie (1994, 1996) provided evidence that the concentration of calpain is the most important factor in tenderisation and µ-calpain has been shown to be responsible for most of the post-mortem proteolysis of muscles

(Geesink et al., 2006). On the other hand, the activity of µ-calpain was only 5-10% by day 14 of ageing (Koohmaraie, 1996). Although Stolowski et al., (2006) showed that the calpastatin concentration had no effect on tenderisation, the ratio of calpain to calpastatin is likely to be key as has been demonstrated both in vitro (Geesink and

Koohmariae, 1999) and in vivo (Koohmaraie et al., 1991; McDonagh et al., 1999;

Veiseth et al., 2004). In addition, Casas et al., (2006a) have shown that variants in both the µ-calpain gene (CAPN1) and calpastatin gene (CAST) have significant effects (p<0.05) on tenderness in association studies using single nucleotide polymorphisms (SNP) markers.

The lysosomal proteases include the cathepsins D, B, H and L. The optimum pH range for the lysosomal proteases is usually considered to be around 5.5-6.5

(Koohmaraie, 1989; Lee et al., 1993). However, Asghar and Bhatti (1987) suggested an even wider pH range of 3.5-6.5. Lysosomal enzymes mainly degrade myosin and actin (Gault, 1992). However, lysosomal proteases are also able to attack myosin heavy chain, myosin light chain, actin, troponin T, troponin I, C-protein, titin, nebulin, tropomyosin, α-actinin, cross-links and mucopolysaccharides (Lawrie,

1992). Nevertheless, the lysosomal proteases are not able to degrade the intact myofibrillar proteins as the size of myofibrils is too large to be engulfed into lysosomes (Goll et al., 1989). Hence, the proteolysis of the myofibrillar proteins requires the degradation of the intact myofibrillar structure by the calpain system before lysosomal proteolysis (Goll et al., 1989).

The proteasome is the other major proteolytic system. Proteasomes are a 26S complex that consists of a 20S core particle and a 19S regulatory particle.

Approximately 80-90% of the proteins in a cell are degraded into amino acids by the proteasomes (Goll et al., 1998; Goll et al., 2008). However, proteasomes do not have the ability to degrade the intact myofibrils directly (Koohmaraie, 1992; Solomon and

Goldberg, 1996). The myofibrils must be degraded initially by virtue of other enzyme systems such as calpain and lysosomes. Therefore, although proteasomes are involved in muscle protein turnover, their importance in the process of tenderisation does not appear to be critical.

The caspases are considered to involve in apoptosis, which is thought to cause the myofibril degradation (Kemp and Parr, 2008). The caspases are a family of cysteine aspartate-specific proteases. Fourteen members in the family have been identified

(Earnshaw et al., 1999). In pigs, the change between the activity of the caspase 9 at 0 hours and 32 hours was negatively associated with shear force in the LD muscle

(Kemp et al., 2006). However, in cattle, a relationship between caspase activity of and shear force has not been found and the myofibril fragmentation is not affected by the activity of caspases (Underwood et al., 2008).

1.5 Intrinsic factors affecting tenderness

There are many factors that affect tenderness. These include both intrinsic factors that are part of the biology underlying beef production and extrinsic factors that are part of carcass and meat processing or the animal management (Oddy et al., 2001;

Robinson et al., 2001).

1.5.1 Sarcomere length

In previous studies, sarcomere length as the major cause of toughness has been widely debated (Hosteler, 1972, Smith et al., 1979). This is because if meat is aged properly, the effect of sarcomere length is reduced. Koohmaraie et al., (2002) demonstrated that the relationship between sarcomere length and meat tenderness is modified by ageing. Wheeler and Koohmaraie (1994) showed that in lamb, the M. longissimus dorsi increases in toughness during the first 24 hours after slaughter and then becomes tender during storage. This finding is consistent with other reports

(Goll et al., 1995; Taylor et al., 1995; Koohmaraie and Geesink, 2006; White et al.,

2006a). It is believed that sarcomere shortening during the rigor period is the cause of the increased toughening during the 24 hours after slaughter, and sarcomere length can be regarded as a fundamental source of variation in tenderness during rigor mortis. However, after 24 hours post-mortem, tenderness is affected by other factors.

1.5.2 pH decline and ultimate pH pH decline is a function of the degradation of glycogen to lactic acid. The pH drops from 7.2 in the live animal to 5.4~5.5 in the carcass. Different muscles have different

19 pH values post-mortem (Hertzman et al., 1993). In general, meat with higher pH values will have higher Warner-Bratzler shear force values and meat with lower pH values will have lower Warner-Bratzler shear force values. However, at pH values greater than 7, no relationship between pH and Warner-Bratzler shear force has been observed (Bouton et al., 1982).

Some reports (Yu and Lee, 1986; Purchas, 1990) have indicated that pH values within 5.8-6.3 result in the minimum tenderness. pH values within 5.8-6.3 are neither optimal for the calpains (>6.3) or cathepsin proteolysis activity (<5.8). The results from Purchas and Aungsupakorn (1993) supported this conclusion. Meat with a pH value of 6 was found to be tougher than meat with lower pH values. The findings of

Kanawa and Takahashi (2002) provided further support of the optimal pH theory.

They demonstrated that if the pH is lower than 5.89 and 5.62 at 15℃ or 5℃, respectively, µ-calpain is inactivated.

However, the faster the decline of pH, the more free calcium is released. The high concentration of free calcium increases the activation of proteases (Hwang et al.,

2004). Neath et al., (2007a) demonstrated that the decline of pH to less than 6 occurs within 24 hours post-mortem. Therefore, the pH may be able to directly affect tenderness or toughness in this very early stage of post-mortem. Lawrie (1992) reported that when the pH decreases to less than 6, the activity of ATPase also rapidly decreases and the AMP cannot be recharged to form ATP for the movement of muscle fibres and the release of the actin myosin complex. Thus, rigor mortis cannot be prevented without the sufficient ATP under pH 6. As a consequence, pH may play a role to indirectly trigger the proteolytic systems, but can also influence the activity of ATPases which in turn, directly affects tenderisation post-mortem.

The temperature of meat has an influence on the tenderness (Olsson et al., 1994;

Koohmaraie, 1996; Devine et al., 2002). In particular, a relationship between tenderness and the factors of temperature, pH and free calcium has been demonstrated. Hwang et al., (2004) showed that during normal processing, the temperature of meat is highly correlated with the pH in meat. They found that the pH of meat at 5℃ or 15℃ is higher than meat at 36℃. Higher temperatures cause the pH of meat to decrease rapidly. Furthermore, higher temperatures increase the free calcium concentration in the muscle sarcoplasm because of the inability to retain calcium in the sarcoplasm reticulum and mitochondria (Huff-Lonergan et al., 2000).

Free calcium can directly affect the tenderness by activating the calcium-activated proteases that can degrade the myofibrils (Honikel et al., 1981; Goll et al., 1983).

Therefore, higher post-mortem temperatures result in faster rates of glycolysis which may result in more tender meat as measured by Warner-Bratzler shear force.

Temperature can also directly influence the occurrence of rigor mortis. Storage temperatures under 10℃ or above 20℃ cause the occurrence of rigor mortis. Locker and Hagyard (1963) introduced the muscle shortening concept observed at extreme high or low temperatures. The muscle shortening which results from temperatures below 10℃ is called cold shortening and at temperatures above 20℃, it is called heat shortening (Devine et al., 1999). Shortening is the decrease of the sarcomere length.

Cold shortening occurs due to the rapid decrease of temperature which releases calcium from sarcoplasmic reticulum of muscle. The ionic calcium and the remaining ATP lead to the occurrence of muscle contraction but not relaxation. Heat shortening occurs due to a rapid pH fall which decreases the proteolysis. The depletion of the ATP and the inactivation of proteolysis results in the meat shortening.

The concept of the influence of storage temperature on muscle shortening was expanded by Pearson and Young (1989) and integrated with the decline of pH in meat to form a pH/temperature window model (Figure 1.3) (Thompson, 2002). The pH/temperature window model illustrates the interaction between the two factors of pH and temperature to affect tenderness. The shortening of meat occurs when there are two specific combinations of conditions for these two factors (Thompson, 2002;

Thompson et al., 2006). If the meat temperature is within the range (5℃ ~15℃) and the pH remains above 6, then cold shortening occurs. If the meat temperature is within the high range of 35℃~40℃ and the pH is under 6, then heat shortening occurs. The meat within either of these two windows becomes tough.

Figure 1.3 pH/temperature window model. (From Thompson, 2002.)

However, Hwang et al., (2004) showed that the meat at 36℃ still has lower shear force values than meat at 15℃ and 5℃. This is not in agreement with the concept of the pH/temperature “window” model. An explanation for this discrepancy is that higher temperatures may result in the heat shortening but the higher temperatures also activate the calpain system because a higher concentration of Ca2+ is released.

Therefore, heat shortening can be overcome by other mechanisms. Cold shortening

22 is said to remain a real issue for meat processing. However, although cold shortening does lead to tougher meat, the length of ageing period seems to be the main factor affecting tenderness ultimately, irrespective of any shortening (White et al., 2006b).

1.5.3 Effect of muscle fibre types on ageing

Ouali and Talmant (1990) found that muscle fibre type plays a role in the ageing process. During post-mortem tenderisation, the fast-twitch type fibres were shown to degrade faster than slow-twitch type fibres. However, the calpain content in the slow-twitch type fibres is higher than that in the fast-twitch type fibres. This seems contradictory to the common concept that the greater calpain activity, the greater the tenderness of the meat. There are several possible explanations for this contradiction.

Firstly, Koohmaraie (1996) suggested that the sensitivity of fast-twitch type fibres to calpain causes faster ageing in fast-twitch type fibres than that in slow-twitch type fibres. Secondly, Ouali and Talmant (1990) found that fast-twitch type fibres have a higher calpain/calpastatin ratio than slow-twitch type fibres. The higher calpain/calpastatin ratio implies that the degradation of the muscle structure would be faster because of the inhibitory effect of calpastatin.

Other reports support the potential role of muscle fibre type on tenderness.

Schiaffino and Reggiani (1996) showed that different muscle types have a different composition of myosin heavy chains, and this may result in different degrees of tenderness via difference in proteolytic degradation. Furthermore, in different muscle fibre types, there is variation in ATPase activity which can affect sarcomere shortening as the quantity of ATP differs in the muscle fibres post-mortem (Gorza

1990; Stienen et al., 1996).

Thus, muscle fibre type contributes to the ageing process and therefore, tenderness, albeit indirectly. The proportion of muscle fibre types in the meat varies systematically with age, muscle location, nutrition, exercise, genotype (eg myostatin and callipyge genes) and treatment (eg β-agonists) (Maltin et al., 1987; Picard et al.,

1994; Jurie et al., 1995; Carpenter et al., 1996; Wegner et al., 2000). In addition,

Type II fibres have a larger myofibril diameter than Type I fibres and meat with larger myofibril sizes is more tender (Maltin et al., 1997). Like myostatin mutations in cattle, highly muscled callipyge sheep also have increased numbers of Type II fibres (Koohmaraie et al., 1995). However, the effect on fibre diameter (hypertrophy) is greater than on fibre number (hyperplasia) and there is increased calpastatin activity which is likely associated with the greater toughness. The castration of males often also results in increased proportions of Type II fibres. However, bulls have bigger fibres of both types, which may help explain why meat from bulls is tougher than meat from steers (Dreyer et al., 1977, Young and Bass 1984).

1.5.4 Effect of muscle bundle size on tenderness

Brady (1937) showed the number of muscle fibres per bundle was negatively correlated with shear force (-0.81) and positively correlated with tenderness score

(+0.55). Brady (1937) also considered the muscle bundle size and concluded that the muscle bundle size is positively correlated with shear force (+0.22~+0.53). In the study of Hiner et al., (1953), the results supported Brady’s conclusions that larger fibre diameters resulted in greater shear force. In contrast, Albrecht et al., (2006), reported that the muscle fibre bundle was not associated with shear force. Tuma et al.,

(1962) presented two different outcomes based on the relationships between muscle fibre diameter and shear force that may provide an explanation of these inconsistent

24 observations. In the study by Tuma et al., (1962), samples were collected from different age and muscle types. The correlation between muscle fibre diameter and shear force were significant when the age was taken into account in the analytical model in all muscle types. However, without the age in the model, significant correlations were not found in either the LD or ST muscles at 14 days post-mortem.

At 2 days post-mortem, in LD muscle, a significant correlation was still observed.

The results of Tuma et al., (1962) suggest that the cause of the different findings on the muscle fibre diameter and tenderness is because of the muscle type and number of ageing days.

1.5.5 Intramuscular fat

Intramuscular fat (marbling) has been shown to affect tenderness but explains less than 10% of the variation in sensory tenderness and juiciness (Crouse et al., 1978;

Dikeman, 1987; Tatum et al., 1980). Stolowski et al., (2006) demonstrated that although the percentage of intramuscular fat in meat varied between breeds, there was no significant effect of intramuscular fat on tenderness between muscles.

Nishimura et al., (1999) demonstrated that the relationship between tenderness and intramuscular fat in Japanese Black cattle differs between muscles; it was significant for the M. longissimus dorsi, but not the M. semitendinosus. Thus, the factors that cause differences in tenderness between breeds are not necessarily the same factors that cause differences in tenderness between muscles. While the relationships between intramuscular fat and tenderness are variable and often weak, the relationship between intramuscular fat and consumer acceptance or eating quality is often stronger (Thompson 2004).

1.5.6 Growth rate

Growth rate has been correlated with shear force values. High growth rates results in lower shear force values (Blanchard, 1994; Therkildsen et al., 2002). The effect of growth rate on tenderness is thought to be associated with the rate of protein turnover

(Aberle et al., 1981). Protein turnover consists of two processes, protein synthesis and protein degradation. Protein degradation is mainly controlled by the calpain system. McDonagh et al., (1999) showed that feed restriction in lambs can decrease muscle development by decreasing calpastatin, the calpain inhibitor. However, Sazili et al., (2003) and Therkildsen et al., (2002) demonstrated that the growth rate of cattle did not affect the concentration of calpastatin in beef and the effect of growth rate on tenderness was not due to a change in the calpain system. The difference in the results of McDonagh et al., (1999) with others may be because of the species specificity or the times of measurements. McDonagh et al., (1999) found the highly significant effect of calpastatin on tenderness at day 9 of ageing. Therkildsen et al.,

(2002) only measured the activity of the calpastatin until day 7 of ageing and used different muscle types. Moreover, the feeding strategies were different between the studies. Hence, the effects of growth on tenderness and role of the calpain system are still unclear and need to be further investigated.

1.5.7 Genetic effects

Many of the effects of the composition of muscle and the proteolytic systems on tenderness / toughness are genetic effects (Burrow et al., 2001; Hocquette et al.,

2006; Thompson et al., 2006). However, Robinson et al., (2001) found that the heritability of tenderness measured by shear force in temperate Bos taurus cattle breeds was only 0.06~0.11. Hence, non-heritable factors play a larger role on tenderness in temperate cattle. Post-slaughter factors mainly contribute to these

26 environmental effects. For tropically adapted cattle breeds of Bos indicus descent, the heritability has been reported as much higher (0.38~0.40) (Robinson et al., 2001).

So while improving tenderness using selection is possible in temperate breeds, there are potentially bigger gains to be made in tropically adapted cattle breeds.

The heritability estimates for shear force vary substantially in different studies.

Shackelford et al., (1994) calculated that the heritability for Warner-Bratzler shear force was 0.53 in Bos indicus and Bos taurus steers. However, other researchers found the heritability for shear force after 14 days of ageing was low (0.06 to 0.14) in Brahmans (Riley et al., 2003). The differences may be the result of the breeds used in the studies, the maturity of cattle or the ageing time of the meat. In general, tenderness is low to moderately heritability depending upon the cattle breed

(Shackelford et al., 1994, Robinson et al., 2001; Riley et al., 2003), but is generally higher in Bos indicus breeds. Not many other tenderness traits have heritability estimates although calpastatin activity had a high heritability of 0.65 in Brahman steers (Shackelford et al., 1994),

Notably, a number of DNA variants have been shown to be associated with tenderness related traits, and some with major effects have been incorporated into breeding schemes. These variants are within genes in known tenderness pathways (as discussed above) and include myostatin (MSTN) (Casas et al., 2000), calpain 1

(CAPN1) (Page et al., 2002), calpain 3 (CAPN3) (Barendse et al., 2008), calpastatin

(CAST) (Barendse, 2002; Schenkel et al., 2006) and lysyl oxidase (LOX) (Barendse,

2002; Drinkwater et al., 2006). Of these, CAST and CAPN1 have been well documented to affect tenderness in many breeds under different management systems (Page et al., 2004; White et al., 2005; Drinkwater et al., 2006; Morris et al.,

2006; Casas et al., 2006a; Van Eenennaam et al., 2007, Frylinck et al., 2009).

Multiple variants within MSTN also have been demonstrated to affect tenderness as well although the effects are larger in the ST muscle than the LD muscle (Kobolak and Gocza, 2002; Wheeler et al., 2004; Lines et al., 2009).

1.6 Extrinsic factors affecting tenderness

In order to improve tenderness, both carcasses and animals have been treated to avoid problems related to meat shortening, to progress the ageing process, and to promote animal growth.

1.6.1 Electrical stimulation

The use of electrical stimulation on carcasses has a positive effect on tenderness

(Takahashi et al., 1987; White et al., 2006a). The primary mechanism of electrical stimulation is to increase the rate of post-mortem glycolysis which reduces the probability of cold shortening when carcasses are cooled too quickly (Chrystall and

Devine, 1978; White et al., 2006a). However, Dutson et al., (1980) and Ferguson et al., (2000) found that the electrical stimulation also accelerates the proteolytic process. These observations suggest that the electrical stimulation decreases pH, which can affect tenderness by increasing Ca2+ concentration, thereby, stimulating the calpain system.

1.6.2 Carcass suspension and Tendercut®

The carcass suspension method is used to reduce toughness, based on decreasing the tension caused by rigor mortis. Suspending the carcass by the pelvic bone or sarco- sciatic ligament can increase the length of the sarcomeres and decrease the shear force. Tendercut® is another method that uses the same principle as carcass suspension. By severing the back bone ligaments and tendons, the LD muscle is

28 stretched and more tender meat can be expected from the carcass (Wang et al., 1994;

Sorheim, 2001). Sorheim (2001) reported that carcass suspension or Tendercut® interacts with the chilling rate. The effect of carcass suspension or Tendercut® on shear force can be seen when the chilling rate is fast (that is, when meat is stored at

2℃ such that the core temperature of the meat is decreased to 5℃ after 10 h). If the meat undergoes a medium chilling rate (at 10℃ for the first 7 h and then at 2℃ for the rest of ageing process such that the core temperature of the meat decreases to 9℃ after 10 h), the significant effect of carcass suspension or Tendercut® on shear force disappears. This indicates the effect is through the proportion of shortening that occurs.

1.6.3 Exogenous meat treatments

The level of tenderness in meat can be altered by infusing muscles with calcium chloride or sodium pyrophosphate plus sodium chloride after slaughter. The principle for using these salts is that they increase the concentration of Ca2+, triggering the activation of the calpain system (Koohmaraie et al., 1994; Lee et al., 2000). NaCl is an exogenous chemical reagent that is used to help tenderisation by increasing the susceptibility of the myofibrils to proteolytic enzymes. Basically, these exogenous treatments all elevate the activity of proteolytic enzymes to accelerate tenderisation.

1.6.4 Animal treatment effects

Hormonal growth promotants have been used in some sectors of the beef industry.

For example, β-agonists are used to increase meat yield of meat (Dikeman, 2007).

However, the use of β-agonists can cause the production of tougher meat because although β-agonists increase the activities of calpastatin and m-calpain, they also significantly decrease the activity of µ-calpain (Parr et al., 1992). The use of the

29 electrical stimulation counteracts this effect of β-agonists supposedly by increasing the activity of the calpain system (Wheeler and Koohmaraie, 1997).

There have been contradictory results showing different effects of other hormonal growth promotants on tenderness (Thompson, 2002). Some research indicates that using various hormonal growth promotants causes tough meat (Roeber et al., 2000;

Thompson, 2002). Other studies do not support this observation (Hunter et al., 2001a;

Hunter et al, 2001b). There is no clear explanation for the different conclusions. The implant strategies using different growth promotants may explain some of the discrepancies between growth promotant effects on tenderness. The hormonal growth promotants have included androgenic and estrogenic hormones such as estradiol, trenbolone acetate, zeranol, estradiol benzoate, progesterone, testosterone or a combination of these hormones (Schanbarche, 1984; Roeber et al., 2000). In the study by Roeber et al., (2000), the effects of the different growth promotants on shear force were examined. Most of those growth promotants in the study did not affect shear force or tenderness scores although some did have an effect. The results indicated that under the same conditions of breed and age, the different growth promotants had different effects on shear force and tenderness taste panel scores.

This suggests that it may be the intrinsic nature of some hormones that is affecting tenderness and not increased growth per se, but further research is required to determine why some hormones affect tenderness while most have little or no effect.

1.7 Conclusion

Tenderness is a major factor affecting consumer acceptance of beef. It can be assessed by consumer taste panels or measured as shear force using the Warner-

Bratzler technique. This review has outlined large processing, management and muscle specific effects on tenderness. Within a group of animals treated equally, there are also significant genetic effects on tenderness. Some of these are breed effects, but others represent genetic variation within breeds.

Improving beef tenderness by using breeding programs has been difficult because of the expense and problems of measuring tenderness accurately, in addition to the fact that it is measured on the carcass. Therefore, using DNA marker selection should help in breeding schemes for meat quality traits, such as tenderness (Burrow et al.,

2001; Hocquette et al., 2006). The genes significantly associated with tenderness, thus far, affect muscle fibre characteristics (myostatin) or are involved in proteolysis

(namely, calpain and calpastatin). However, these genes still do not explain a large proportion of the genetic variation in beef tenderness. More genes affecting tenderness are expected to be identified and interactions between the genes should be tested. Discovering novel genes and their relationships may lead to additional biochemical mechanisms underlying tenderness as well as enabling better DNA marker selection for meat quality.

Chapter 2: Resources and trait derivation

It is clear from previous research that there are major genes which affect the tenderness of beef and variants within these genes may be used for selection to improve meat quality (Burrow et al., 2001; Hocquette et al., 2006; Warner et al.,

2010). The hypothesis for the project herein was that there will be other genes in addition to these known genes that affect tenderness, and the aim was to identify other DNA variants as potential genetic markers for selecting tender meat. Further, discovering the genes that harbour these DNA variants should help provide a better understanding of the mechanisms underlying tenderness.

The strategy used herein for identifying the genes and DNA variants involved in tenderness was to map QTL related to tenderness, and then to locate and sequence candidate genes within these QTL or from other sources. The sequences were aligned to identify DNA variants within the genes for genotyping. The DNA variant genotypes were then analysed in association studies to determine if they are related to beef tenderness.

To undertake the QTL mapping and the association studies, the microsatellite (or simple tandem repeat, STR) genotypes and tenderness phenotypes available from the

JS Davies Cattle Gene Mapping Project were utilised (described below; see also

Morris et al., 2009). It has been shown previously that the animals in the JS Davies

Cattle Gene Mapping Project were affected by 2 major genes related to tenderness, myostatin (Esmailizadeh et al., 2008; Lines et al., 2009) and calpain 1, but not calpastatin (Morris et al., 2006). This provided the opportunity to examine the interactions between these major genes and other genes affecting tenderness. The tenderness phenotypes also provided the opportunity to explore the relationships between the various measures of tenderness and to derive two new traits, adjusted shear force and ageing rate, to improve the phenotypes used for the analyses.

2.1 JS Davies Gene Mapping Cattle Project

The experimental herd design, carcass measurements and genotypes of the microsatellite markers and the single nucleotide polymorphisms in the MSTN,

CAPN1, IGF1 and IGFR1 genes were obtained from the JS Davies Cattle Gene

Mapping Project (Morris et al., 2009). All other DNA polymorphisms analysed herein were discovered and genotyped for the JS Davies herd as part of this PhD research project. The available trait data from the JS Davies herd were used to derive new phenotypes herein (described below).

The purpose of the JS Davies Cattle Gene Mapping project was to identify quantitative trait loci (QTL) affecting different traits related to cattle and is one of the largest such mapping projects ever attempted for beef production (Morris et al.,

2009). Half of the animals were raised in the Australia and the other half were raised in New Zealand as a sister herd. The mapping herd was based on crosses between two extreme Bos taurus breeds (Limousin and Jersey). The first cross bulls

(Limousin X Jersey or Jersey X Limousin) were born in 1994 and then backcrossed to purebred Jersey and purebred Limousin cows to generate the mapping progeny.

Limousin, which is a beef cattle breed, has very different phenotypes for growth

(growth rate, weight, body size, retail beef yield) and body composition (degree of muscling, marbling, meat tenderness, fat content and fat colour) from Jersey, which is a dairy cattle breed. Due to the difference in the performance between two extreme breeds, wide genetic variation in traits was observed in the backcross progeny.

In Australia, three of the F1 cross bulls (361, 368 and 398) were used to produce 366 backcross progeny from 1996 to 1998 with an average of 120 progeny per sire (161

Limousin back-cross cattle and 205 Jersey back-cross) (Figure 2.1). Genotypic and

34 phenotypic data from these Australian progeny were utilised herein. Because of the power of the backcross design and the number of informative progeny, it was calculated that QTL explaining only 0.5 standard deviation of any given trait would be detected with a probability of 0.8 for a single family analysis.

Figure 2.1 Experimental design of the JS Davies Cattle Gene Mapping Project.

2.2. Tenderness traits

Traits related to tenderness were measured in the backcross progeny and the data utilized herein included muscle weights of both the M. longissimus dorsi muscle (LD) and M. semitendinosus muscle (ST), ultimate pH of the LD and ST muscles, and cooking loss of the LD and ST muscles. Warner-Bratzler (WB) shear force of the LD and ST muscles after multiple ageing times (meat aged for 1, 5, 12 and 26 days post- slaughter), compression of the ST muscle, and hydroxyproline content of the ST muscle (as an assessment of collagen content) were the direct tenderness related traits measured for this study (Appendix 1). The measurements of all beef tenderness-related traits were performed by Zbigniew Kruk and David Lines for the

JS Davies Cattle Gene Mapping Cattle Project (Esmailizadeh et al., 2008).

2.2.1 Shear force

Warner-Bratzler shear force is the primary objective measurement and is widely used for evaluating beef tenderness (Miller et al., 2001; Caine et al., 2003). However,

Warner-Bratzler shear force is not a perfect tool for reflecting the subjective sensory measures of tenderness (Shackelford et al., 1995; Destefanis et al., 2008; Rodas-

González et al., 2009). Shear force is affected by meat processing, animal handling, muscle type and breed of cattle (Takahashi et al., 1987; Goll, 1991; Koohmaraie,

1994; Koohmaraie, 1996; Muir et al., 1998; Lee et al., 2000; Sorheim, 2001; Devine et al., 2002; Belew et al., 2003; Nuernberg 2005; Stolowski et al., 2006; Dikeman,

2007; Archile-Contreras et al., 2010; Hope-Jones et al., 2010).

Post-slaughter, proteolysis occurs in chilled meat so “ageing” of meat is commonly used to produce a more tender product (Bratcher et al., 2005; Riley et al., 2005;

Stolowski et al., 2006). To date, the shear force measures in this project (and most

36 others) have been analysed separately in univariate analyses (Lines, 2006,

Esmailizadeh et al., 2008). The data include shear force values for 4 ageing times

(days 1, 5, 12 and 26) in 2 muscles (M. longissimus dorsi and M. semitendinosus) for each animal., Therefore, a multi-trait analysis was performed herein to better understand the biology of meat tenderness.

2.2.1.1 Multi-trait analysis of shear force

Shear force was measured for the Davies herd at four time points: 1, 5, 12 and 26 days of ageing post-slaughter. All steaks were stored in boxes at 0-1℃ in the chiller at the abattoir after boning, which occurred the day after slaughter (day 1). On days 1,

5, 12 and 26 after slaughter, boxes were shifted from the chiller to the freezer (-20℃).

The ageing day “treatment” was allocated randomly along the LD and ST muscles

(M. longissimus dorsi and M. semitendinosus, respectively) to prevent ageing day and position being confounded. The differences between ageing days can result from various factors contributing to tenderization or random measurement error.

The linear mixed model fitted to 366 cattle x 8 shear force measurements (=2928) included fixed effects of cohort (combination of year and sex), breed (Jersey or

Limousin dam), sire (3), MSTN F94L genotype (AA, AC, CC), muscle (LD or ST), ageing day (1, 5, 12, 26) and various interactions (see below). Individual animal, individual x muscle and individual x ageing were fitted as random effects. The residual was the three-way interaction individual x muscle x ageing day. The model was fitted using Genstat 8.1 (Lawes Agricultural Trust, 2005). Other factors, such as pH, were not included in the model because they were to be evaluated separately.

Mixed model fixed factors for multi-trait analysis of shear force:

Cohort: sex x year (96H, 96S, 97H, 97S, 98H, 98S)

Breed: XJ (¾ Jersey, ¼ Limousin) and XL (3/4 Limousin, ¼ Jersey)

Sires: three F1 sires (316, 368 and 398)

MSTN: myostatin F94L genotype (AA, AC, CC)

Muscle: LD (M. longissimus dorsi muscle) and ST (M. semitendinosus muscle)

Ageing day: last day in chiller (day 1, day 5, day 12 and day 26)

Cohort.Muscle: interaction between cohort and muscle

Cohort.Ageing day: interaction between cohort and ageing day

Breed.Muscle: interaction between breed and muscle

Breed. Ageing day: interaction between breed and ageing day

Sire.Muscle: interaction between sire and muscle

Sire. Ageing day: interaction between sire and ageing day

MSTN.Muscle: interaction between MSTN and muscle

MSTN.Ageing day: interaction between MSTN and ageing day

Muscle.Ageing day: interaction between muscle and ageing day

Cohort.Muscle.Ageing day: interaction between cohort, muscle and ageing day

Breed.Muscle.Ageing day: interaction between breed, muscle and ageing day

Muscle.Sire.Ageing day: interaction between muscle, sire and ageing day

Muscle.MSTN.Ageing day: interaction between muscle, MSTN and ageing day

Mixed model random factors for multi-trait analysis of shear force:

Individual: 366 JS Davies gene mapping progeny

Individual.Muscle: interaction between individual and muscle (LD or ST; with 366 x 2 = 732 levels)

Individual.Ageing day: interaction between individual and ageing day (day 1, 5, 12 or 26; with 366 x 4 = 1464 levels)

2.2.1.2 Results of multi-trait analysis of shear force

Shear force was first analysed using the full mixed model above. Of the 19 fixed factors, 11 showed significant effects on shear force (Table 2.1). Consequently, 12 fixed factors (the 11 significant factors plus sire) were fitted with a reduced mixed model with the 3 random factors remaining. It was unexpected that no sire effects were observed since there was large amount of data and all other effects were highly significant. Nevertheless, sire was retained in the model to be consistent with other models used herein. Cohort effects were large and demonstrated the importance of the environment (year) and gender (sex).

Table 2.1 Significant factors and variance components for shear force factors.

Fixed factors F-statistic F prob Cohort 50.5 <0.001 Breed 5.5 0.020 Sire 1.2 0.304 MSTN F94L 9.3 <0.001 Muscle 217.7 <0.001 Ageing day 285.0 <0.001 Muscle.Cohort 13.8 <0.001 Ageing day.Cohort 7.8 <0.001 Muscle.Breed 147.3 <0.001 Ageing day.Breed 4.3 0.005 Muscle.MSTN F94L 6.1 0.002 Muscle.Ageing day 14.8 <0.001

SE of Component Random factors Component Individual 0.1282 0.1282 Individual.Muscle 0.2526 0.2526 Individual.Ageing day 0.0019 0.0019 Residual 0.2520 0.2520

There was a significant breed x muscle interaction (Table 2.2). Shear force was higher in the ST than LD for both breeds, but while the Limousin LD was 12% tougher (higher shear force) than the Jersey LD, the opposite was the case for the ST muscle for which the Limousin 8% was lower than Jersey. There was also a significant MSTN F94L genotype x muscle interaction (Table 2.2). In both cases the homozygous MSTN F94L variant (AA) was more tender (less tough) than the wild type (CC). In the ST muscle, the difference was greater (15%) than the LD (5%).

Table 2.2 Least squares means for breed and MSTN F94L genotype effects LD muscle ST muscle Genetic effect ±standard error ±standard error Breed

XJ 3.996b±0.058 4.954c±0.058 XL 4.461a±0.051 4.563a±0.051 MSTN

CC 4.316a±0.067 5.075c±0.067 AC 4.251a±0.045 4.874b±0.045 AA 4.118a±0.098 4.327a±0.098 Breed: XJ (¾ Jersey, ¼ Limousin) and XL (3/4 Limousin, ¼ Jersey) MSTN: myostatin F94L genotype (AA, AC, CC) Superscript: labels as the difference in P<0.05.

There was a significant muscle x ageing day interaction (Table 2.1, Figure 2.2). As expected, there was a greater ageing effect in the LD than ST. Specifically, at day

26, the LD shear force was 20% lower than day 1, whereas the ST was 13% lower.

There was also a significant breed x ageing day interaction (Figure 2.2), where the

Jersey meat aged more (17%) than the Limousin (14%).

Figure 2.2 Differences in ageing between muscles (LD, ST) and breeds (XJ, XL).

The total variance in shear force was 0.635 kgF2 (Table 2.1, Figure 2.3). The standard deviation between individuals was approximately 0.618 (kg shear force).

The repeatability was 0.603. The repeatability was calculated as the sum of the variances (0.383 kgF2) of the random factors (Individual + Individual x Muscle +

Individual x Ageing day) divided by the total variance of 0.635 kgF2. The correlation coefficient for the different ageing times was 0.776 (=√0.603). The accuracy (that is, the correlation between the estimated value and the true value) of the shear force value based on 4 measures and a repeatability of 0.603 was 0.93.

Individual

Individual.Muscle

Individual.Ageing day

Residual

Figure 2.3 Components of variance in shear force.

While there was significant individual variance (Figure 2.3), the individual x muscle interaction component was large. This suggests that tenderness of the LD and ST are separate traits and, therefore, were treated as such in the remainder of this project.

The very small variance for individual x ageing day suggests that there is little re- ranking across days and so averaging shear force across days was appropriate. Thus, predictions of individual values for the separate muscles (“adjusted shear force”) were used as the primary traits of interest throughout the project. Adjusted shear force is a best linear unbiased prediction and could be considered effectively as an average of the shear force measurements. This value is a highly accurate (0.93) measure for each animal which should increase the power of detection of genetic effects. The correlation between ST adjusted shear force and LD adjusted shear force was only 0.44, so they are not closely related traits (Figure 2.4).

8 y = 0.2536x + 3.686 7 R² = 0.1981 6 5 4 3 2 1

Adjustedshear force LD for msucle 0 0 2 4 6 8 10 Adjusted shear force for ST msucle

Figure 2.4 Relationship between ST (wbst_adjusted) and LD adjusted shear force (wbld _adjusted).

2.2.2 Ageing rate

The role of ageing time on shear force has been widely discussed (Shanks et al.,

2002, Gruber et al., 2006). Longer ageing times result in lower shear force values. 42

The change in shear force over time or ageing rate in specific periods of tenderization is controlled by proteolysis (Calkins and Seideman 1988). Calkins and

Seideman (1988) investigated the effect of the cathepsins, calcium-dependent proteases, on ageing. The activities of the cathepsins B and H were significant and explained the 35% and 58% of the variation in the shear force change between day 1 and day 14 of ageing and between day 3 and day 6 of ageing, respectively.

Furthermore, the effect of calpain 1 on tenderness is also largely ascribed to differences in ageing rate in many studies (Dayton et al., 1976; Koohmaraie 1994;

Koohmaraie 1996).

The studies also suggest the possibility that the genes that affect shear force at specific times may be different from the genes that controlling ageing rate. However, the discovery of genetic markers specifically for ageing rate has seldom been investigated. Delineating the mechanism of ageing and the function of the genes related to ageing should benefit the understanding of tenderness. Therefore, derivation of the new trait, called “ageing rate” herein, was derived with the aim of exploring tenderness from another point of view.

2.2.2.1 Analysis of ageing rate

Two measures of tenderness from the Davies herd (shear force at day 1 and day 26) were used for developing another new trait, called “ageing rate” herein. The correlation between shear force at day 1 and day 26 were 0.66 and 0.71 for the ST and LD muscles, respectively. The analysis of shear force suggested there was only a small amount of individual variation in shear force across ageing days (0.0019, Table

2.1, Figure 2.3). That is, the individual effects were relatively consistent across different ageing days. However, because the multiple measures of shear force herein

43 were “averaged” across days, it was possible to estimate ageing rate in addition to average tenderness. Furthermore, given that shear force of the LD and ST were not strongly correlated, it was possible that ageing rate in the different muscles is also not strongly correlated due to the biology of the muscles, not just measurement error.

The residual variance of shear force was quite large (Table 2.1, Figure 2.3) and this included the individual x muscle x ageing rate variation. Thus, another trait representing the ageing rate was derived as the difference of between day 1 and day

26 for this project. Consequently, “ageing rate” in this instance is amount of ageing that occurred in natural log kilogram per 25 days.

The log-transformed shear force values based on the Euler's Constant e=2.7182 were used for the calculation of the differences between shear force values at day 1 of ageing and shear force value at day 26 of ageing. There was a difference in variance between day 1 and 26, whereas the log-transformed shear force values had minimal difference in the two variances of day 1 and day 26. Therefore, the log transformed shear force values were used to increase the accuracy of the estimated ageing rate

(Table 2.3). A large difference between the log of the shear force values indicated a greater proportion of ageing (that is, a greater decrease from day 1 to day 26). The log values effectively represent a percentage decrease in shear force.

Table 2.3 Variances for shear force values of LD and ST muscles for the original and logarithmic measures. wbld1 wbld26 wbst1 wbst26 LD ST ageing ageing rate rate Mean 4.875 3.959 5.304 4.669 Variance 1.738 0.875 0.783 0.626 Mean on log scale 1.526 1.361 1.649 1.526 0.206 0.129 Variance on log scale 0.061 0.057 0.163 0.172 0.034 0.018 wbld1 = log-transformed Warner-Bratzler (WB) shear force measured at day 1 of ageing the M. longissimus dorsi (LD) wbld26 = log-transformed Warner-Bratzler (WB) shear force measured at day 26 of ageing the M. longissimus dorsi (LD) wbst1 = log-transformed Warner-Bratzler (WB) shear force measured at day 1 of ageing the M. semitendinosus (ST) wbst26 = log-transformed Warner-Bratzler (WB) shear force measured at day 26 of ageing the M. semitendinosus (ST)

The factors (cohort, breed, sire, MSTN F94L genotype) were examined for their effects on ageing rate. All factors were fitted as fixed factors into a fixed model (see below). The analyses were performed by using the statistics package Genstat 8.1

(Lawes Agricultural Trust, 2005).

Fixed model fixed factors for ageing rate: Cohort: sex x year (96H, 96S, 97H, 97S, 98H, 98S)

Breed: XJ (¾ Jersey, ¼ Limousin) and XL (3/4 Limousin, ¼ Jersey)

Sires: three F1 sires (316, 368 and 398)

MSTN: myostatin F94L genotype (AA, AC, CC)

2.2.2.2 Results from analysis of ageing rate

The results showed that the ageing rate was not affected by breed, sire, and MSTN

F94L genotype with one exception, the breed effect on ageing rate for the LD muscle

(Table 2.4). Like adjusted shear force, a sire effect was not found for ageing rate.

The ageing rate for the LD muscle from the backcross Jersey (0.23) was significantly higher than that for the backcross Limousin (0.17) (Table 2.5, Figure 2.2). However, the variation of the ageing rate for LD muscle explained by breed was relatively low

(2%). Cohort was the only fixed factor with a highly significant effect on both muscles. Cohort was responsible for 18% and 6% of the variation in ageing rate for

LD and ST muscles, respectively. Thus, the largest portion of the variation in ageing rate is likely to be caused by factors other than genetic factors.

Table 2.4 Tests of significance for ageing rate.

Ageing rate for LD Ageing rate for ST Cohort <.001 <.001 Breed 0.003 0.891 Sire 0.177 0.971 MSTN F94L 0.753 0.348

Table 2.5 Least squares means for breed and MSTN F94L genotype. Ageing rate for LD Ageing rate for ST ±standard error ±standard error Breed

backcross Jersey 0.2336a±0.065 0.1227c±0.044 backcross Limousin 0.1693b±0.071 0.1308c±0.049 MSTN F94L

CC 0.1942e±0.017 0.1414f±0.014 AC 0.2063e±0.012 0.1172f±0.010 AA 0.2185e±0.026 0.1275f±0.020 Superscripts of a and b: significant difference (P<0.05) for breed effect on ageing rate on LD. Superscripts of c: significant difference (P<0.05) for breed effect on ageing rate on ST. Superscripts of e: significant difference (P<0.05) for MSTN F94L effect on ageing rate on LD. Superscripts of f : significant difference (P<0.05) for MSTN F94L effect on ageing rate on ST.

The variability of ageing rate in the LD and ST muscles of the 366 mapping progeny was examined (Table 2.3). The ageing rate for the LD muscle was significantly higher than that of the ST muscle (P<0.001) and the variance in LD ageing rate was also 85% greater than that of the ST muscle. Thus, the LD muscle had a greater decrease in ageing rate and a wider distribution.

For some samples, the day 26 shear force values were greater than day 1 so that 36

LD samples and 46 ST samples actually had negative ageing rates. There were also 6

LD samples and 14 ST samples that showed no decrease in shear force after 26 days of ageing. It is unlikely that shear force truly increased with ageing. There are two possible explanations. A possible explanation is that the differences between shear force values at day 1 and day 26 were not large and were within experimental error.

Although the LD muscle had a relatively high ageing rate (Table 2.3), examination of the relationships between shear force at day 26 and ageing rate in the LD and ST muscles showed that ageing rate of the LD muscle only explained 4% of the total variation in shear force at day 26 of ageing. This is less than the 12% for the ST muscle. Therefore, the ageing rate was more critical for the ST muscle than the LD muscle. On the other hand, the variance in ageing rate for the LD muscle is almost 3 times that of the ST muscle. This means that in the ST muscle, the ageing rates for all samples are relatively consistent although the ageing rate itself for the ST muscle was lower than that of the LD muscle. Due to the consistent effect of ageing rate on the ST muscle, the percentage of variation in the ST shear force at day 26 explained by the ageing rate is higher than the percentage of LD shear force at day 26 explained by the ageing rate.

As expected, the relationship between the adjusted shear force value and ageing rate in the ST muscle was not significant (F=0.337) (Figure 2.5), demonstrating they are independent traits. While this could be simply because ageing day was in the model for adjusted shear force, it would not necessarily remove the correlation between the traits. The relationship between the adjusted shear force and ageing rate in the LD muscle was significant (P=0.005) (Figure 2.6), but the R2 was only 1.9% (r= 0.137).

Given that the correlation coefficient was so low, ageing rate and adjusted shear force were interpreted as being independent measures of separate biological processes in the LD muscle and therefore, should aid gene discovery if mapped as separate traits. A relationship between ageing rate for LD muscle and ST muscle was observed (P=0.002) (Figure 2.7), but the r2 was only 2.6% (r= 0.164). Given that the correlation coefficient was so low, ageing rates between muscles were also regard as separate traits.

Fitted and observed relationship with 95% confidence limits

0.4

0.2

-0.0 Ageing_rate_for_ST_muscle

-0.2

-0.4 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0

Adjusted_shear_force_for_ST Figure 2.5 Scatter plot for adjusted shear force and ageing rate in ST muscle. The red line indicates the line of best fit. The blue lines indicated the 95% confidence limits.

Fitted and observed relationship with 95% confidence limits

0.8

0.6

0.4

0.2 Ageing_rate_for_LD_muscle 0.0

-0.2

3 4 5 6 7 8 9

Adjusted_shear_force_for_LD Figure 2.6 Scatter plot for adjusted shear force and ageing rate in the LD muscle. The red line indicates the line of best fit. The blue lines indicated the 95% confidence limits.

0.8

0.6 y = 0.2237x + 0.1779 R² = 0.0269 0.4

0.2

-0.6 Ageing rate LD for muscle -0.4 -0.2 0 0.2 0.4 0.6 -0.2

-0.4 Ageing rate for ST muscle

Figure 2.7 Relationship between ageing rates of ST and LD.

2.3 Relationships between tenderness traits

In order to determine which traits are highly correlated and hence, which traits may have the same associated SNPs, the relationships between a range of traits that may be related to tenderness were examined. The analysis included the following traits for both the LD and ST muscles except where indicated: adjusted shear force, ageing rate, shear force at day 1 of ageing, shear force at day 26 of ageing, cooking loss, pH, compression (ST only), hydroxyproline (ST only) and muscle weight.

Two models (models 1 and 2) were developed to identify the relationships between traits. Model 1 was utilized to describe the relationships between traits related to tenderness (except adjusted shear force) with cohort, breed and sire as fixed factors.

Model 2 was used to describe the relationship between traits and adjusted shear force without any fixed factors as these were already accounted for in the trait derivation of adjusted shear force.

Model 1: Model with no fixed effects to show the original effects on the traits

The model was used to show that the relationship between two given traits (except adjusted shear force).

Yijkl= μ + αi + βj + γk + bXl + Rijkl

Yijkl the response variable (for any trait)

μ the overall mean

th αi the effect of i cohort (six levels)

th βj the effect of j breed (Jersey or Limousin)

th γk the effect of k sire (three sires)

bXl b is the regression coefficient. Xl is any trait

Rijkl the residual

Model 2: Model with no fixed effects to show the original effects on the traits

The model was used to show that the relationship between a given trait and adjusted shear force.

Yi= μ + bXi + Ri

Yi the response variable (for any trait)

μ the overall mean

bXi b is the regression coefficient. Xi is any trait

Ri the residual

Using these models, most of the relationships between the traits were not significant or did not explain much of the variation. However, in terms of understanding the biology of tenderness and finding genes controlling tenderness, there were some important findings regarding which traits were related and which traits were not related (detailed below).

2.3.1 Relationship between compression in ST muscle and tenderness

Ageing rate and adjusted shear force as derived herein appear to be independent traits for both the LD and ST muscles. Compression was not measured for the LD muscle, so the relationship between ageing rate and compression in the LD muscle could not be analysed herein. However, using model 1, a potential positive relationship between ageing rate and compression in the ST muscle was found

(P=0.057). However, the amount of variation in compression explained by ageing rate was only 1.3%. This small amount of variation in compression that was accounted for by ageing rate indicates that these are basically independent traits even though their relationship was significant. Therefore, compression and ageing rate were both analysed in the association study. On the other hand, shear force (at day 1 and 26 of ageing and adjusted shear force) were responsible for 10.0%, 4.7% and

8.0% of the variation of the compression, respectively. Therefore, in the LD muscle, shear force and compression were closely related.

2.3.2 Relationship between hydroxyproline content in ST muscle and tenderness

The amount of hydroxyproline in the muscle is directly related to the amount of collagen present (Hill, 1966; Colgravea, et al., 2008). Hydroxyproline content was measured in the ST muscle as it has been reported that collagen has an effect on the tenderness of this muscle (Riley, et al., 2005). However, the relationship between shear force or compression and hydroxyproline concentration in the ST muscle was not significant using a model with cohort, breed and sire fitted as fixed factors (Table

2.6). The lack of a relationship between collagen and compression was unexpected.

Table 2.6 Tests of significance of the relationships between shear force or compression and hydroxyproline concentration in the ST muscle.

Trait Significance

Compression 0.221

Adjusted shear force for ST 0.954

Ageing rate for ST 0.821

2.3.3 Relationship between cooking loss and tenderness

The regression analysis with the model 1 showed that cooking loss was associated with shear force at ageing day 1 for both the LD (P<0.001) and ST muscles

(P<0.001) and ageing day 26 for the ST muscle (P=0.015). However, cooking loss only explained a small amount of variation in shear force (1.2% for ST muscle at day

1 of ageing and 0.8% for the other days). On the other hand, the relationship between compression and cooking loss was significant in the ST muscle (P<0.001). Cooking loss explained 4% of the variation in compression (regression coefficient = 0.0474).

Hence, although cooking loss is associated with tenderness, the effect of the water- holding capacity of the meat was only moderate.

2.3.4 Relationship between pH and tenderness

The relationships between the pH of the meat and shear force or compression were identified by linear regression with cohort, breed and sire fitted as fixed factors

(model 1). In the ST muscle, there was no effect of pH on shear force (at ageing day

1, day 26 and adjusted shear force) or compression. However, the relationship between pH of meat and shear force (at ageing day 1 and day 26) in the LD muscle was highly significant (P<0.001) and positively correlated (model 1 from above).

The pH was also associated with adjusted shear force of the LD muscle (P=0.002)

(model 2 from above). The pH explained 2.9%, 2.0% and 2.4% of the variance in

52 shear force at day 1, day 26 and adjusted shear force, respectively. As expected, higher pH values resulted in higher shear force values in the LD muscle and clearly explained at least some of the variation in tenderness observed.

2.3.5 Relationship between muscle weights and tenderness

Muscle weight is regarded as representative of the effects of hyperplasia and/or hypertrophy. The MSTN F94L variant in the population herein results in increased tenderness and increased hyperplasia, but not increased hypertrophy (Lines et al.,

2009). Therefore, the association between muscle weight and tenderness was also considered. At day 1 of ageing, the LD muscle weight was not associated with shear force but the ST muscle weight was significantly associated and explained 3.1% of the variation. The relationship between tenderness and muscle weight (LD and ST) was significant at day 26 of ageing. The correlation between muscle weight and shear force was negative (r=-0.433) in the ST muscle and positive (r=0.071) in the

LD muscle. The muscle weights (LD and ST) explained 1.0% and 3.2% of the variation for the LD and ST muscles at day 26 of ageing, respectively. The relationship between compression and muscle weight for the ST was not significant.

One of the proposed mechanisms by which myostatin can affect tenderness is that the myostatin mutations increase the muscle fibre number and therefore, the volume relative to collagen, thereby decreasing shear force (Albrecht et al., 2006; Lines et al.

2009). Hence, identifying which genes can affect tenderness by changing the muscle weight may be important. Because the adjusted shear force value calculations included the myostatin genotype, the effects of muscle weight on tenderness could be determined by examining the relationship between adjusted shear force and muscle weight. The regression analyses (model 2) showed that LD and ST muscle weights

53 were not associated with adjusted shear force although the ST muscle weight was close to significance (P = 0.091). However, the variation of adjusted shear force accounted for by ST muscle weight was only 0.5%. Thus, muscle weight does not appear to have any effect on the shear force of the LD muscle or the ST muscle once the myostatin genotype was considered.

2.4 Discussion

The two newly derived traits, adjusted shear force and ageing rate, were shown to have independent roles in tenderness. These traits should allow a more accurate depiction of the genetic effects on tenderness as many of the common other effects have been accounted for in their derivation. This is especially the case for adjusted shear force.

2.4.1 Adjusted shear force

The effects of the breed, muscle and ageing time on adjusted shear force were observed. These findings are in accord with the report of Stolowski et al., (2006).

Breed and muscle type have been shown to affect a variety of the characteristics of the muscles. For example, the proteolytic systems (such as calpain/calpastatin), soluble collagen content, pH, and perimysium thickness are affected by muscle type and breed (Brooks and Savell, 2004; Stolowski et al., 2006; Lefaucheur, 2010).

Furthermore, changes in muscle characteristics, such as muscle bundle size, muscle types and pH, have been reported to be associated with tenderness in some studies

(Cross et al., 1973; Ouali, 1992; Maltin et al., 1997).

Herein, the interaction between breed and muscle showed that the meat from the

Limousin backcross cattle had a different adjusted shear force between the LD and

ST muscles, but this was not true for the Jersey backcross cattle. The ST muscle

54 content is different from other muscles, including the LD, and the genetic background of the Jersey backcross cattle is different from the Limousin backcross cattle. Interestingly, the interaction between the MSTN F94L genotype and muscle only showed a MSTN genotype effect on the shear force of the ST muscle. This obviously implies that this genetic effect is muscle specific. This phenomenon explains why the results of QTL mapping and association studies vary in the literature as specific muscle types are investigated usually without the consideration of the related major genes.

Moreover, if factors affecting tenderness are not taken into account, the genetic markers discovered affecting tenderness may have different effects in different muscles or in different breeds. Consequently, the genetic markers may not be used effectively for selection of tenderness across breeds and muscles or applied to investigate the biological mechanisms underlying tenderness. All previous analyses have used each LD ageing day as separate traits (Riley et al., 2003; Casas et al.,

2006; Alexander et al., 2007; Gutie´rrez-Gil et al., 2007). The adjusted shear force trait derived herein should provide the opportunity to identify significant genes whose influence would extend across muscles, breeds and ageing time. Given that the adjusted shear force for each individual in this data set is effectively an average across ageing times, it was also valuable to develop a trait that quantifies the change in shear force as proteolysis occurs, namely ageing rate.

2.4.2 Ageing rate

The different shear forces at a variety of ageing times in different muscles revealed that the ageing rate is dependent on muscle structure. Furthermore, the ageing process in muscles obviously occurs at different periods of tenderization (Gruber et

55 al., 2006; Stolowski et al., 2006). Herein the ageing rate was taken as the difference in log shear force between the first and last day available (that is, the proportion of ageing that occurred over 25 days) to maximise the measure relative to calculating over smaller time periods. However, ageing rate is not a trait normally used to study the genetics of tenderness. Consequently, genetic markers for ageing rate have not been identified even though the effect of the proteolytic systems (such as the cathepsins and calpain) has been suggested to affect ageing rate (Calkins and

Seideman, 1988; Dayton et al., 1976; Koohmaraie 1994). Thus, it is worth mapping

QTL for ageing rate to further discovery of genetic markers for tenderness.

2.4.3 Related tenderness traits

From the results, shear force was partially associated with the compression in the ST muscle and explained a significant proportion of the variation (up to 10% at day 1 of ageing). Hence, the factors affecting shear force should be, in part, the same factors affecting the compression.

The analysis showed that the cooking loss in the experimental samples affected compression and shear force. Cooking loss is largely water loss and therefore, believed to be linked to water-holding capacity (Aaslyng et al., 2003; Straadt et al.,

2006). High water loss can lead to dry meat and toughness. Cooking loss explained

4% of the variance of the residual of compression and approximately 1% of the variance of the residual of shear force when cohort, breed and sire were taken into account as three fixed factors. It appears that cooking loss, in fact, did not play a large role on either tenderness or compression herein. However, the effects of cooking loss on tenderness may have been underestimated because cooking loss did not vary greatly (mean = 26.25 Kg, SD = 1.81 Kg). While cooking loss was not

56 critical to either shear force or compression herein, it may still be an indirect indicator of eating quality. Cooking loss will be lower when intramuscular fat (IMF) levels are higher (Ueda et al., 2007). Intramuscular fat and juiciness improve the eating quality of beef (Thompson, 2002).

The pH influences tenderness by affecting the activities of the proteolytic enzymes and/or the water-holding capacity (Yu and Lee, 1986; Asghar and Bhatti 1987;

Purchas, 1990). Based on the analyses herein, pH was only associated with tenderness of the LD muscle, and only accounted for 2% of variation in shear force at days 1 and 26 and in adjusted shear force. There was no relationship between pH and compression. The difference of pH between LD and ST muscles was less than

0.07. A relationship between pH and tenderness was expected. However, the range of the pH in LD and ST muscles was between pH 5.5 and 5.8 in the experimental samples. The ultimate pH within this range will not affect proteolysis (Yu and Lee,

1986; Purchas, 1990). Hence, the variation in pH of the meat was probably not large enough to show any pH effects on tenderness herein.

Although the models used to examine the effect of collagen were not the same as the model used previously (Lines, 2006), the lack of the effect of collagen on tenderness of the ST muscle was consistent. The collagen content did not show an effect on tenderness or compression. The correlation between hydroxyproline content and ST day 26 shear force was only 0.27. The correlation between hydroxyproline content and compression was only 0.063. The type of collagen can be affected by the diet

(Listrat et al., 1999) and the collagen content varies greatly from muscle to muscle.

Therefore, the reason that collagen content was not associated with tenderness or compression could be a large diet effect on the muscles analysed. Alternatively, the

57 insoluble collagen content and not total collagen may be affecting tenderness (Riley et al., 2005). Hence, although a collagen effect on tenderness and compression has been found in other studies (Cross et al., 1973; Findlay et al., 1986; Kuypers and

Kurth 1995), confounding factors herein may have decreased the amount of variation in collagen content to the point that no resulting variation in tenderness or compression was observed.

Muscle weight was considered as one of the potential factors which may affect tenderness because the increased muscle weight in animals of the same age could result in a change of collagen content or the size and number of the muscle fibres

(Albrecht, et al., 2006; Lines, et al. 2009; Renand, et al., 2001; Allais, et. al, 2009).

Thus, increased muscling has been hypothesised to explain why double muscled cattle produce very tender meat (Swatland and Kieffer, 1974; Wegner et al., 2000;

Bellinge et al., 2005). The results showed that the muscle weight was only slightly associated with shear force at day 1 of ageing for the ST muscle and at day 26 of ageing for both the LD and ST muscles. There was no relationship between adjusted shear force and muscle weight. Therefore, the theory that the tender meat will result from increasing the muscle weight was not strongly supported.

2.5 Conclusion

The adjusted shear force trait was developed as a more accurate objective trait.

Ageing rate was derived as it appears to be independent measure of overall shear force. These two new traits provide better accuracy and different phenotypes for finding of genes associated with tenderness using QTL mapping and association studies. The identified relationships between these tenderness related traits should also provide a better understanding of the mechanisms underlying tenderness.

Chapter 3: QTL Mapping

3.1 Introduction

The purpose of quantitative trait loci (QTL) mapping is to identify regions of chromosomes potentially containing genes related to the traits of interest. For tenderness, previous QTL mapping results for tenderness with and without myostatin

F94L genotypes in the analytical model (Lines, 2006) were used to find candidate genes for the study herein. In addition, two new traits (adjusted shear force and ageing rate) were mapped using linkage analysis and compared with these previous mapping results (Lines, 2006).

Many QTL mapping methods have been developed during the last 4 decades. Soller et al., (1976) used regression to identify markers located near QTL and presented the advantages of a linear regression model that takes marker information into consideration. However, Lander and Botstein (1989) pointed out that utilizing linear regression would under-estimate the phenotypic effects of QTL. In addition, they emphasized that a putative QTL may not be identified when a QTL with only a small effect is closely linked with the markers or when a QTL with large effect is only loosely linked with the adjacent markers. Therefore, they suggested using interval mapping with pairs of markers, which can enhance the accuracy of identifying QTL.

However, there can be false positive results with the interval mapping approach, where there are no real QTL in the detected region. The false positive results can be caused by the existence of other QTL that are close to the detected region. Moreover, interval mapping can only detect a pair of flanking markers at a time, making the analysis inefficient (Zeng, 1993; Jansen, 1994). Zeng (1993) presented a combined method that consists of an interval mapping and multiple regression approach, the so-called composite interval mapping. This method eliminates the interference from other QTL and retains both the advantages of interval mapping and the regression

60 approach. Fan and Xiong (2002) demonstrated that methods based on regression are more credible than linkage disequilibrium methods for QTL mapping and are adequate for multiple marker analysis. Therefore, analytical methods tend to be based on regression for QTL mapping (Anderson et al., 2006). However, when there are a large number of markers taken into consideration at the same time, the power of the statistical test decreases (Zhao et al., 2005). For a microsatellite marker study such as the one herein, the number of markers would not be as large as when using

SNP markers. Much higher numbers of SNP markers are required than microsatellite markers to obtain the same resolution in linkage mapping because microsatellite markers are more polymorphic. The average heterozygosity for SNP markers is 0.36, while the average heterozygosity for microsatellite markers is 0.6 (Powell et al.,

1996; Wong et al., 2004). Therefore, the problem of statistical power is less serious when microsatellites are used, and individual markers rather than marker haplotypes can be used, thereby increasing accuracy.

Consequently, the Haley-Knott regression method (Haley and Knott, 1992) was used for the QTL mapping in this study, which utilizes marker genotypes for the model rather than using marker haplotypes. The Haley-Knott regression method is based on a simple regression approach. The possibility of having inherited a paternal QTL given the marker genotypes is fitted as a dependent variable into the regression model. If linked markers are significant when regressed on a trait, they provide the genomic position of a potential QTL for that trait.

3.2 Methods for quantitative trait loci (QTL) mapping

The two newly derived traits of adjusted shear force and ageing rate were used herein for the quantitative trait loci linkage analysis to improve the accuracy of

61 mapping tenderness as a trait. Preliminary QTL mapping results for shear force and compression were available from the JS Davies Gene Mapping Cattle Project (Lines,

2006) for comparison. Mapping QTL in this study was based on the Haley-Knott regression method, an interval mapping approach. QTL mapping experiments were performed by utilizing QTL Express software (http://qtl.cap.ed.ac.uk/). Finding the position of the QTL was achieved by examining the probability of having a QTL between two microsatellite markers. If any position showed a significant probability of having a QTL in the regression analysis, then there may be a QTL near the position of the marker.

The microsatellite genotypes previously obtained for 366 backcross progeny from the 3 sire families in the Davies herd (161 Limousin back-cross cattle and 205 Jersey back-cross) were analysed. The microsatellite markers were selected from their location in the bovine genome (except the X and Y chromosomes) based on the Meat

Animal Research Centre map (http://www.marc.usda.gov/genome/cattle/cattle.html).

There were 285 microsatellite markers across the autosomes used in the QTL mapping, but only those markers heterozygous in the sires were genotyped in the corresponding progeny. On average, 150 microsatellite markers were genotyped for each progeny with an average density of 1 informative marker per 20 cM, thereby providing good coverage of the genome. The markers were genotyped using gel electrophoresis of individual microsatellite PCR products by AgResearch.

The two new traits (adjusted shear force value and ageing rate) and two muscle types

(LD and ST muscles) were used as dependent variables in the analysis. As in the other QTL studies, shear force values at specific ageing days had been used in mapping QTL for tenderness previously in the Davies herd. In order to provide a

62 better resolution of shear force QTL, however, adjusted shear force was mapped herein. Because different genes may affect the rate of tenderization versus the average shear force, ageing rate (herein, the difference between the log-transformed shear forces at ageing day 1 and day 26) was also mapped to find QTL related to tenderisation. The aim of mapping these two new traits was to determine if the final level of tenderness might be due to a combination of genes affecting tenderness during different times in the ageing process.

For each trait, it was necessary to fit different fixed effects into the models because it was previously established that there are 2 major genes affecting tenderness segregating in the herd, calpain 1 and myostatin (Morris et al., 2006; Esmailizadeh et al., 2008; Lines et al., 2009). For adjusted shear force, three models were performed for QTL discovery (QTL models 1-3). The purpose of using three models was to find QTL for adjusted shear force with and without the effects of calpain 1 as calpain 1 is known to affect shear force. By taking calpain 1 into account within the models, the QTL responsible for the remaining variation in tenderness will be more significant. Comparing the models with and without calpain 1 should reveal new

QTL and uncover the epistatic interactions between calpain 1 and other QTL. Three models were required as there are 2 different single nucleotide polymorphisms

(SNPs) within calpain 1 that are known to affect shear force, SNP316 and SNP530.

A fourth model with the myostatin F94L genotype was not required for adjusted shear force because the MSTN genotype was already incorporated.

Thus, the first 3 QTL models (QTL models 1-3 below) were used for adjusted shear force with two other known tenderness associated DNA variants, CAPN1-SNP316 and SNP530, taken into account. Comparing model 1 with models 2 and 3

63 determined the amount of the variation accounted for by CAPN1-SNP361 and

CAPN1-SNP530 separately and revealed other significant QTL that explained some of the remaining genetic variation.

Y= bx + e (QTL model 1)

Yi = μ + αi + bx + ei (QTL model 2)

Yijl = μ + αi + βj + bx + eijl (QTL model 3)

μ overall mean,

αi effect of the ith calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC) βj effect of the jth calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA) bx the allele substitution effect of the putative QTL, X is the probability of including a QTL within the markers. e(i, ijl) residual effect

Four QTL models (QTL models 4-7) were required for ageing rate to take the MSTN

F94L genotype into account as well as the CAPN1 genotypes (that is, CAPN1-

SNP316 and SNP530). Comparing model 4 with models 5, 6 and 7 determined the amount of the variation accounted for by MSTN F94L, CAPN1-SNP361 and CAPN1-

SNP530 separately and revealed other significant QTL that explained some of the remaining genetic variation.

Yijk = μ + αi + βj + γk+ bx + eijk (QTL model 4)

Yijkl = μ + αi + βj + γk + λl + bx + eijkl (QTL model 5)

Yijklm = μ + αi + βj + γk + λl + δm + bx + eijklm (QTL model 6)

Yijklmn = μ + αi + βj + γk + λl + δm + θn + bx + eijklmn (QTL model 7) μ overall mean, th αi effect of the i cohort (six levels) th βj effect of the j breed (Jersey or Limousin) th γk effect of k sire (three sires) th λl effect of the l myostatin F94L genotype (AA, CA, CC), th δm effect of the m calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC) th θn effect of the n calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA) bx the allele substitution effect of the putative QTL, X is the probability of including a QTL within the markers. e(ij, ijk, ijklm, ijklmn) residual effect

3.3 Results QTL mapping was conducted for the two new traits (adjusted shear force and ageing rate) derived based on the Warner-Bratzler shear force data from day 1,5, 12 and 26 of ageing post-slaughter for both the LD and ST muscles (section 2.2 and section 2.3, respectively). The variation in both adjusted shear force and ageing rate were greater in the LD muscle than ST muscle (Table 3.1).

Table 3.1 Summary of adjusted shear force and ageing rate data for the LD and ST muscles.

wbld_adjusted wbst_adjusted LD ageing rate ST ageing rate Number of 366 366 352 351 observations Mean 4.228 4.758 0.2059 0.1287 Standard 0.676 0.385 0.1831 0.1384 deviation Variance 0.456 0.148 0.0335 0.0182

3.3.1. Quantitative trait loci for adjusted shear force

QTL mapping for adjusted shear force were implemented with three models (QTL models 1, 2 and 3 as described above). In the QTL model 1, cohort, breed and sire were fitted as fixed factors. In the QTL model 2, the effect of SNP316 of the calpain

1 gene was included as a factor with 3 levels. In the QTL model 3, the effect of the 2

SNPs in the calpain 1 gene (SNP316 and SNP530) that are known to affect tenderness was included.

By comparing the results from these different models, it was possible to determine which QTL are affected by these DNA variants. Moreover, once the genetic effects of these variants have been accounted for, new QTL were often uncovered which explained some of the remaining genetic variation. Therefore, the different models were used for QTL mapping and the results contrasted.

For the QTL mapping across the 3 sire families, F values ≥ 4 were regarded as above the threshold for statistically inferring the potential position of a QTL (Churchill and

Doerge, 1994). If multiple testing was not accounted for, this would be equivalent to a P value (F3, 366=4) of 0.008. Adjusted shear force values were calculated from the shear force measures on both muscles (LD and ST). From the analysis using QTL model 1, two significant QTL segregated for adjusted shear force of the LD muscle and three significant QTL segregated for adjusted shear force value of the ST muscle

(Table 3.2) (Appendix 2).

Table 3.2 QTL positions for adjusted shear force (across sire families).

Trait BTA Peak (cM) F value Adjusted shear force value of LD muscle 5 92 4.00 Adjusted shear force value of ST muscle 5 96 4.72 Adjusted shear force value of ST muscle 18 36 4.33 Adjusted shear force value of LD muscle 29 52 6.54 Adjusted shear force value of ST muscle 29 52 6.89

The QTL at 92 and 96 cM on BTA5 for LD and ST, respectively, showed significant association with adjusted shear force values (Figure 3.1A). Given the small distance between 92 and 96 cM, the QTL for LD and ST muscle are likely to be the same.

A) Without CAPN1 genotypes. B) With CAPN1 genotypes. Figure 3.1 QTL on BTA 5 for adjusted shear force of LD and ST muscles, without and with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model. (Blue line: LD muscle; Green line: ST muscle)

On BTA 18, there was a QTL at 36 cM for the ST muscle only (Figure 3.2A). As the

QTL was specific for ST muscle, this implies the underlying genes may be related to a difference in muscle types.

A) Without CAPN1 genotypes. B) With CAPN1 genotypes. Figure 3.2 QTL on BTA 18 for adjusted shear force of LD and ST muscles, without and with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model. (Blue line: LD muscle; Green line: ST muscle)

On BTA 29, QTL were identified at 52 cM for both LD and ST and again are most likely to be the same gene(s) (Figure 3.3). Interestingly, there was another significant

QTL peak around 30 cM, suggesting that there may be two QTL on BTA 29.

A) LD muscle. B) ST muscle. Figure 3.3 QTL on BTA 29 adjusted shear force of LD and ST muscles without any calpain 1 genotypes fitted as fixed factors in the model.

In addition to the five main QTL for the adjusted shear force identified across the sire families, there were QTL discovered segregating in individual families

(Appendices 9 and 10). These QTL did not have large effects and may not be easily detected when examined in association studies. However, they are still worth further investigation as their size of effect may be larger in other populations.

3.3.1.1 Effect of CAPN1-SNP316 and SNP530 genotypes on adjusted shear force QTL

Two SNPs with the calpain 1 gene, CAPN1-SNP316 and CAPN1-SNP530, have been well documented to affect tenderness (Page et al., 2002; Casas et al., 2006a).

The CAPN1-SNP316 genotype had highly significant effects (P<0.001) on adjusted shear force of both the LD and ST muscles in the samples herein as well (Appendix

18). The CAPN1-SNP530 genotype had significant effects (P<0.05) on adjusted shear force of the LD muscle. Therefore, the genotypes of the CAPN1-SNP361 and

SNP530 were fitted as fixed factors in the QTL model (QTL model 3, section 3.2).

The analysis was used to determine if the calpain 1 gene was harbouring the quantitative trait nucleotide (QTN) for any of the QTL or may be interacting with other QTL and thus, affect their significance.

Using this QTL model, there were four significant QTL (Table 3.3) (Appendix 4).

The QTL discovered at 92~96 cM on BTA 5 for the LD and ST muscles were again unaffected by the CAPN1 genotypes (Figure 3.1A versus Figure 3.1B). Since the adjusted shear force values accounted for MSTN F94L genotype, these QTL are not likely to be related to the CAPN1 or MSTN genes. Interestingly, the QTL on BTA 5 segregated differently in the sire families. The QTL had a significant effect of 0.39 kgF on the LD muscle and 0.25 kgF on the ST muscle in the family of sire 368. In the sire 361 family, only an effect of 0.19 kgF for the ST muscle was observed.

Table 3.3 QTL positions for adjusted shear force with CAPN1-SNP316 and CAPN1- SNP530 genotypes in the model.

Trait BTA Peak (cM) F value Adjusted shear force value of LD muscle 5 92 4.05 Adjusted shear force value of ST muscle 5 96 5.01 Adjusted shear force value of ST muscle 18 40 4.24 Adjusted shear force value of ST muscle 25 0 3.87 Adjusted shear force value of LD muscle 29 28 4.15

The QTL found at 40 cM on BTA 18 was also not influenced by the CAPN1 genotypes (Figure 3.2A versus Figure 3.2B). Again the QTL was only present for the

ST muscle across sire families. This suggests that the QTL at 40 cM is based on specific characteristics of the ST muscle. This QTL for the ST muscle had an increase of 0.4976 kgF in the sire 361 family. This QTL also was not influenced by the CAPN1 genotypes.

A new putative QTL on BTA 25 at 0 cM for the ST muscle appeared when the

CAPN1 genotypes were fitted into the QTL model (Figure 3.4A versus Figure 3.4B).

The CAPN1 locus is not likely to be affecting or interacting with the QTL on BTA

25, in this instance, as the significance of the QTL would decrease rather than increase. The most plausible explanation of the new QTL is that once the additional variance is accounted for by including the CAPN1 genotypes in the model, then any

QTL of moderate size are more likely to be observed.

A) Without CAPN1 genotypes. B) With CAPN1 genotypes. Figure 3.4 QTL on BTA 25 for adjusted shear force of LD and ST muscles without and with CAPN1-SNP316 and SNP530 genotypes fitted as fixed factors in the QTL model. (Blue line: LD muscle; Green line: ST muscle)

Only the 2 adjusted shear force QTL on BTA 29 were affected by fitting the CAPN1 genotypes in the QTL model. The QTL found at 52 cM on BTA 29 disappeared for both the LD and ST muscles (Figure 3.5A). This indicated that the QTL at 52 cM are most likely due to the calpain 1 (CAPN1) gene. The CAPN1 gene is located at 45

Mb and 63 cM on BTA 29 (Ensembl: http://www.ensembl.org/Bos_taurus/

Gene/Summary?g=ENSBTAG00000010230). Hence, there is only a short distance between CAPN1 gene and the inferred QTL maximum position at 52 cM.

A) With CAPN1-SNP316 and SNP530 genotypes. B) With CAPN1–SNP316 genotypes. Figure 3.5 QTL on BTA 29 for adjusted shear force of the LD and ST muscles with CAPN1-SNP316 and SNP530 genotypes or with CAPN-SNP316 alone fitted as fixed factors in the QTL model. (Blue line: LD muscle; Green line: ST muscle)

3.3.1.2 Effect of CAPN1-SNP316 genotype on adjusted shear force QTL

The CAPN1-SNP316 genotype has been shown to affect tenderness in the Jersey-

Limousin backcross population rather than the CAPN1-SNP530 (Page et al., 2002).

Therefore, the genotype of the CAPN1-SNP316 was fitted as a fixed factor in the

QTL model by itself (QTL model 2, section 3.2) to determine if the effects of

CAPN1-SNP316 would parallel those previously observed in which the SNP316 accounted for the variance rather than SNP530 (Page et al., 2002).

When the effect of the CAPN1-SNP316 genotype was taken into account in the QTL mapping alone, there were still three QTL for the ST muscle and two QTL for the

LD muscle (Table 3.4) (Appendix 3). The QTL for the ST and LD muscles on BTA 5 were not affected by the CAPN1-SNP316. The QTL at 36 cM for the ST muscle on

BTA 18 only shifted marginally to 40 cM and the significance of the QTL at 0 cM on

BTA 25 for the ST muscle only increased slightly from an F value of 3.87 to 4.01.

Table 3.4 QTL positions for adjusted shear force with CAPN1-SNP316 genotypes in the model.

Trait BTA Peak (cM) F value Adjusted shear force value of LD muscle 5 92 4.22 Adjusted shear force value of ST muscle 5 96 4.98 Adjusted shear force value of ST muscle 18 40 4.43 Adjusted shear force value of ST muscle 25 0 4.01 Adjusted shear force value of LD muscle 29 28 3.91

However, when the CAPN1-SNP316 genotype was fitted into the QTL model, the peak at 52 cM on BTA 29 decreased for both the LD and ST muscles with F-values of 3.42 and 1.81, respectively, nearly identical to fitting both the CAPN1-SNP316 and SNP530 simultaneously (Figure 3.5A versus Figure 3.5B). The decrease was again larger for the ST muscle. In addition, the second QTL peak on BTA 29 at 30

71 cM decreased. However, the F-value was still close to 4.0 for the LD muscle.

Moreover, in the family of sire 361, there was now a significant effect of 0.43 kgF for this QTL at 28 cM for the LD muscle. Therefore, the possibility remains that there is another QTL at least for the LD muscle on BTA 29 near 30 cM.

3.3.1.3 Comparison of the QTL for adjusted shear force

The previous QTL mapping results for ST shear force at specific ageing days with myostatin in the model (Lines, 2006) were compared with the QTL herein as the same data set from the Davies cattle gene mapping project was used in both instances (Table 3.5). The two QTL on BTA 5 and 29 overlapped with previously identified QTL. The QTL on BTA 5 in the previous study was at 85 cM for tenderness as measured by Warner-Bratzler shear force of the ST muscle at day 1

(WBST1). This QTL position is close to the QTL for adjusted shear force value (at

92~96 cM) and is most likely to be the same QTL. The other QTL identified at 60 cM on BTA 29 in previous study was 8 cM away from the QTL found in this study and again is likely to be the same QTL.

However, the QTL for the ST muscle on BTA 18 and the QTL for the LD muscle at

28 cM on BTA 29 were not detected in the previous analysis in which only ST data were analysed (Lines, 2006). In addition, the QTL for ST shear force at different ageing days on BTA 4 (38 cM) and BTA 16 (85 cM) found in the previous analysis with MSTN genotypes in the model (Lines, 2006) were not detected for adjusted shear force.

The new QTL discovered herein reflect the importance of maximizing the accuracy of phenotypic values. Though unlikely, they may also reflect real phenotypic differences between adjusted shear force and shear force values from specific days.

The QTL for the adjusted shear force, theoretically, would be based on effects throughout the ageing period and not those only influencing shear force at specific ageing days. For example, the QTL on BTA 16 for day 1 (WBST1) found in the previous study may only have an effect at day 1 of ageing and not at the other ageing days.

Table 3.5 Comparison of shear force QTL positions*

Trait BTA Peak (cM) F value

WBST1 (Lines, 2006) 5 85 4.75

Adjusted shear force value of LD muscle 5 92 4.00

Adjusted shear force value of ST muscle 5 96 4.72

Adjusted shear force value of ST muscle 18 36 4.33

WBST12 (Lines, 2006) 29 60 5.81

Adjusted shear force value of LD muscle 29 52 6.54

Adjusted shear force value of ST muscle 29 52 6.89 *Comparison between previously mapped shear force QTL (Lines, 2006) and QTL mapped herein using same population. WBST1 = Warner-Bratzler shear force of ST on day 1; WBST12 = Warner- Bratzler shear force of ST on day 12

3.3.2 Quantitative trait loci for ageing rate

Since ageing rate as defined herein was shown to affect tenderness independent of adjusted shear force (Chapter 2), QTL for ageing rate were mapped using four QTL models (QTL models 4, 5, 6 and 7, section 3.2). In QTL model 4, there were only three fixed factors (cohort, breed and sire) fitted into the QTL model. In QTL model

5, the effect of MTSN F94L genotype was added as a fixed factor. With QTL model

6 and QTL model 7, the additional effects of the calpain 1 genotypes (SNP 316 and

SNP530) were included and so these are equivalent to QTL models 2 and 3 used for adjusted shear force. Model 5 was necessary because unlike adjusted shear force, the effect of the other major DNA variant known to affect tenderness in the 73 population herein, MSTN F94L, was not taken into account when calculating ageing rate.

The QTL mapping was conducted across the three sire families. Again, all F-values of 4 or greater were regarded as above the threshold for statistically inferring QTL.

Ageing rate was calculated from the measurements of LD and ST muscles at days 1 and 26 (section 2.2.2). There were two significant QTL segregating for the ageing rate of the LD muscle and five significant QTL segregating for ageing rate of the ST muscle (Table 3.6) (Appendix 5, Appendix 10B).

Table 3.6 QTL positions for ageing rate (across sire families).

Traits BTA Peak (cM) F value Ageing rate for ST muscle 1 48 4.50 Ageing rate for ST muscle 3 76 4.10 Ageing rate for LD muscle 4 40 5.06 Ageing rate for ST muscle 7 116 4.50 Ageing rate for ST muscle 11 80 4.18 Ageing rate for LD muscle 13 0 4.39 Ageing rate for ST muscle 20 56 4.03

There were no QTL for both ST and LD muscles on the same chromosome. This implies that the ageing rate of different muscles is affected by different genes as might be expected. In other words, the ST and LD muscles may have different biological pathways underlying ageing rate. The fact that none of the QTL were the same as the QTL for adjusted shear force in each muscle also suggested the genetic independence of these two measures of tenderness.

3.3.2.1 Effect of MSTN F94L genotype on ageing rate QTL

The MSTN F94L genotype has been shown to have a large effect on tenderness in

74 this population (Lines et al., 2009). Hence, the MSTN F94L genotype was fitted into the QTL model (QTL model 5, section 3.2). After including the MSTN F94L genotype in the QTL model, the number of QTL technically reduced from 7 to 5

(Table 3.7) (Appendix 6).

Table 3.7 Comparison of QTL for ageing rate with and without the MSTN F94L genotypes as a fixed effect in the QTL model. Trait BTA Peak F value F value (cM) w/o MSTNa w/MSTNb Ageing rate for ST muscle 1 48 4.50 4.00 Ageing rate for ST muscle 3 76 4.10 3.34 Ageing rate for LD muscle 4 40 5.06 4.84 Ageing rate for ST muscle 7 116 4.50 4.18 Ageing rate for ST muscle 11 80 4.18 3.42 Ageing rate for LD muscle 13 0 4.39 4.63 Ageing rate for ST muscle 20 56 4.03 3.98 aModel 4: without MSTN F94L genotype; bModel 5: with MSTN F94L genotype

The 2 QTL located on BTA 3 and 11 for the ST muscle were no longer significant due to the effect of MSTN F94L genotype. However, the QTL peaks were still present (Figure 3.6 and Figure 3.7). As the myostatin gene is located on BTA 2, the most plausible explanation of these observations is that there is epistasis between the myostatin gene and the QTL.

A) Without MSTN F94L genotypes. B) With MSTN F94L genotypes. Figure 3.6 QTL on BTA 3 for ageing rate for the ST muscle without and with the MSTN F94L genotypes as fixed factors in the QTL model.

A) Without MSTN F94L genotypes. B) With MSTN F94L genotypes. Figure 3.7 QTL on BTA 11 for ageing rate for the ST muscle without and with the MSTN F94L genotypes as fixed factors in the QTL model.

3.3.2.2 Effect of MSTN F94L, CAPN1-SNP316 and SNP530 genotypes on ageing rate QTL

The CAPN1-SNP315 and CAPN1-SNP530 genotypes were added to the QTL model

(QTL model 7, section 3.2) so that any calpain 1 haplotype effects could be accounted for. After including the MSTN F94L genotype and both CAPN1 genotypes, the number of QTL was four (Table 3.8) (Appendix 8, Appendix 10B). There were two QTL found for the LD muscle (BTA 4 and 13) and 2 QTL for the ST muscle

(BTA 1 and 19).

Table 3.8 Comparison of QTL positions for ageing rate with the MSTN F94L and CAPN1 genotypes as fixed factors in the QTL model. Traits BTA Peak F value F value F value (cM) w/MSTNa w/ MSTN+CAPN w/MSTN+ SNP316+SNP530b CAPN316c ST ageing rate 1 48 4.00 3.92 4.55 LD ageing rate 4 40 4.84 4.82 4.95 ST ageing rate 7 116 4.18 3.17 2.98 LD ageing rate 13 0 4.63 6.03 5.40 ST ageing rate 19 48 3.25 4.74 4.64 ST ageing rate 20 56 3.98 3.59 3.07 aModel 5: with MSTN F94L genotype; bModel 7 = MSTN F94L + CAPN-SNP316 + CAPN SNP530; cModel 6: with MSTN F94L + CAPN1-SNP316 genotypes.

With the calpain 1 genotypes in the QTL model, the F value for the QTL at 48 cM on BTA 1 was only slightly less significant, decreasing from 4.00 to 3.92. Although the F value did not quite reach the significance threshold, the position of the QTL may be still regarded as being important given the size of effect. The F value for the

QTL at 40 cM on BTA 4 was not affected by fitting any factors in the QTL model.

Hence, these QTL would appear to be independent of the MSTN and CAPN1 genes.

The QTL at BTA 7 and BTA 20 were no longer significant with the calpain 1 genotypes fitted into the QTL model. It could be the CAPN1 genotypes explained some of the variance of the QTL on BTA 7 and BTA 20 by chance or the more likely explanation is that there is an interaction between these QTL and calpain 1.

The F value for the QTL at 0 cM on BTA 13 for ageing rate for the LD muscle increased after fitting all the fixed factors into the QTL model from 4.39 to 6.03

(Figure 3.13). Furthermore, within all three sire families, the QTL had a significant effect on the ageing rate (Appendix 10A).

A) Without MSTN F94L + CAPN1 genotypes. B) With MSTN F94L + CAPN1 genotypes. Figure 3.8 QTL on BTA 13 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model. (Blue line: ST muscle; Green line: LD muscle)

Moreover, a marginal QTL on BTA 19 became statistically significant. The F value

77 for the QTL on BTA 19 increased from 3.25 to 4.74. This is most likely to be caused by the change in the variance of the residual explained by the CAPN1-SNP316 genotype, allowing additional QTL to be uncovered.

3.3.2.3 Effect of MSTN F94L genotype and CAPN1-SNP316 genotype on ageing rate QTL

Again as the CAPN1-SNP316 genotype appears to have the major effect in the

Jersey-Limousin backcross rather than the CAPN1-SNP530 (Page et al., 2002), it was of interest to determine if the CAPN1-SNP316 alone could explain the effects of calpain 1 on the ageing rate QTL or whether it was a CAPN1 haplotype effect.

Therefore, the CAPN1-SNP316 genotype was also included in QTL model by itself

(QTL model 6, section 3.2) with the MSTN F94L genotype. After including the

MSTN F94L genotype and CAPN1-SNP316 genotype in the model, the same 4 QTL were still significant (Table 3.8) (Appendix 7). The QTL on BTA 7 and BTA 20 once again disappeared and the QTL on BTA 4 and BTA 19 were unchanged. The only discernable differences were a slight decrease in the F value for the QTL on BTA 13 increased from 6.03 to 5.40 and a slight increase in the F value for the QTL on BTA

1 from 3.92 to 4.55.

3.4 Discussion

As expected, the QTL for adjusted shear force and for ageing rate were different. For adjusted shear force, the QTL with the largest effects were on BTA 5, 18 and 29.

There was also one QTL on BTA 25 with moderate size effects on LD adjusted shear force. For ageing rate, the QTL with the largest effects were on BTA 1, 4, 13 and 19.

The QTL of moderate size effects on ageing rate were on BTA 3, 7, 11 and 20. The differences suggest that the genes affecting ageing rate are independent of the genes affecting shear force. However, this is not to say separate biological mechanisms are

78 involved as there was also an ageing effect on adjusted shear force.

Using adjusted shear force as a trait removed the difference between ageing times and other interactions between some of the fixed factors (section 2.2.1.1).

Theoretically, using adjusted shear force should reduce the background noise in the data for QTL mapping. Hence, new QTL were discovered. Therefore, it was not surprising that there was only three QTL that overlapped with the previous QTL mapping results in the same 3 sire families (Lines, 2006). For adjusted shear force of the LD muscle, the QTL at 92 cM on BTA 5 is adjacent to the QTL at 85 cM for

WBST1. For ageing rate in the LD muscle, the QTL at 40 cM on BTA 4 is adjacent to the QTL at 38 cM for shear force at day 26 found in an earlier analysis by Lines

(2006). Notably, the QTL for shear force on BTA2 detected by Lines (2006) was not observed herein because the myostatin F94L variant (which is the quantitative trait nucleotide explaining this QTL) was accounted for in the adjusted shear force trait derivation.

There was also overlap between the QTL on BTA 29 observed herein and a QTL for shear force at day 12 observed by Lines (2006). The analysis herein indicated that the CAPN1-SNP316 variant is the likely quantitative trait nucleotide (QTN) for these

QTL. The QTL mapping results were nearly identical when the two QTL models were used that included the CAPN1-SNP316 genotypes with and without the

CAPN1-SNP530 genotypes (model 2 versus model 3 and model 6 versus model 7, section 3.2). Importantly, the adjusted shear force QTL on BTA 29 at 52 cM were no longer significant when the CAPN1-SNP316 genotypes were included in the model even without the CAPN1-SNP530 genotypes.

It is possible that the CAPN1-SNP316 is not the causative variant for the BTA 29

QTL, but is in linkage disequilibrium with the true QTN. However, this SNP has been shown to effect tenderness in many different cattle breeds (Page et al., 2002;

Casas et al., 2006; Curi et al., 2009; Curi et al., 2010) so the linkage disequilibrium would need to be very tight. Moreover, the SNPs may be functional as both SNP316 and SNP530 cause amino acid substitutions in the CAPN1 protein (glycine316 to alanine and isoleucine530 to valine, respectively; Page et al., 2002), albeit these are relatively conservative amino acid changes.

The QTL discovered herein should be more informative than the QTL found by

Lines (2006) in which shear force was analysed at specific ageing time points and the effect of CAPN1 was not considered. Indeed, including the CAPN1 variants in the models removed the QTL on BTA 29 at ~50 cM for adjusted shear force and revealed a putative second QLT on BTA29 at 15-30 cM (Table 3.9).

Table 3.9 QTL results across sire families for adjusted shear force with the QTL model 1, 2 and 3 with the known DNA variants affecting tenderness.

Trait QTL Model* BTA Peak (cM) F value Adjusted shear force LD 1 5 92cM 4 Adjusted shear force LD 2 5 92cM 4.22 Adjusted shear force LD 3 5 92cM 4.05 Adjusted shear force ST 1 5 96cM 4.72 Adjusted shear force ST 2 5 96cM 4.98 Adjusted shear force ST 3 5 96cM 5.01 Adjusted shear force LD 1 18 36cM 4.33 Adjusted shear force LD 2 18 40cM 4.43 Adjusted shear force LD 3 18 40cM 4.24 Adjusted shear force LD 1 25 0cM 3.14 Adjusted shear force LD 2 25 0cM 4.01 Adjusted shear force LD 3 25 0cM 3.87 Adjusted shear force LD 1 29 52cM 6.54 Adjusted shear force LD 2 29 28cM 3.91 Adjusted shear force LD 3 29 28cM 4.15 Adjusted shear force ST 1 29 52cM 6.89 Adjusted shear force ST 2 29 16cM 2.86 Adjusted shear force ST 3 29 16cM 3.08 *Model 1: no SNPs, Model 2: CAPN SNP316, Model 3: CAPN SNP316 + CAPN SNP530.

On the other hand, for ageing rate, including the CAPN1 variants in the QLT models decreased the significance of the ageing rate QLT on BTA 7 and BTA 11, implying epistatic interactions between the calpain 1 gene and these QTL (Table 3.10). Again the results were identical with or without the CAPN1-SNP530 in the model.

Table 3.10 QTL results across sire families for ageing rate with the QTL model 4, 5, 6 and 7 with the known DNA variants affecting tenderness. Trait QTL Model* BTA Peak (cM) F value Ageing rate for ST muscle 4 1 48 4.50 Ageing rate for ST muscle 5 1 48 4.00 Ageing rate for ST muscle 6 1 48 4.55 Ageing rate for ST muscle 7 1 48 3.92 Ageing rate for ST muscle 4 3 76 4.10 Ageing rate for ST muscle 5 3 76 3.34 Ageing rate for ST muscle 6 3 76 3.34 Ageing rate for ST muscle 7 3 76 3.35 Ageing rate for LD muscle 4 4 40 5.06 Ageing rate for LD muscle 5 4 40 4.84 Ageing rate for LD muscle 6 4 40 4.95 Ageing rate for LD muscle 7 4 40 4.95 Ageing rate for ST muscle 4 7 116 4.50 Ageing rate for ST muscle 5 7 116 4.18 Ageing rate for ST muscle 6 7 120 2.98 Ageing rate for ST muscle 7 7 120 3.17 Ageing rate for ST muscle 4 11 80 4.18 Ageing rate for ST muscle 5 11 80 3.42 Ageing rate for ST muscle 6 11 80 2.6 Ageing rate for ST muscle 7 11 28 2.43 Ageing rate for LD muscle 4 13 0 4.63 Ageing rate for LD muscle 5 13 0 4.63 Ageing rate for LD muscle 6 13 0 5.40 Ageing rate for LD muscle 7 13 0 5.40 Ageing rate for ST muscle 4 19 44 3.25 Ageing rate for ST muscle 5 19 44 4.64 Ageing rate for ST muscle 6 19 48 4.64 Ageing rate for ST muscle 7 19 44 4.64 Ageing rate for ST muscle 4 20 56 4.03 Ageing rate for ST muscle 5 20 56 3.98 Ageing rate for ST muscle 6 20 56 3.07 Ageing rate for ST muscle 7 20 56 3.59 *Model 4: no SNPs, Model 5: MSTN F94L, Model 6: MSTN F94L + CAPN SNP316, Model 7: MSTN F94L + CAPN SNP316 + CAPN SNP530.

Two of the QTL for adjusted shear force were muscle specific (BTA 18 and BTA 25) and two adjusted shear force QTL were not muscle specific (BTA 5 and BTA 29).

However, the QTL for ageing rate were all muscle specific in that there were no

QTL for both ST and LD ageing rate on the same chromosome (Table 3.10). This would suggest that ageing rate (as defined herein) is controlled by different mechanisms in these muscles. Differences in muscle fibre content of the LD and ST muscles and/or proteolytic differences may be the cause.

While ageing rate herein describes the decline in shear force and improvement in tenderness between day 1 and day 26, the QTL for shear force on BTA29 suggests that enzymes in the calpain system are affecting tenderness even after minimal ageing (day 1). In fact, CAPN1 affected tenderness at each of the measured days

(days 1, 5, 12 and 26; data not presented). However, the early effect of CAPN1 at day 1 of ageing did not result in different rates of ageing in the subsequent days of 5,

12 and 26. Therefore, CAPN1 was not associated with ageing rate but was associated with adjusted shear force. While it is not obvious why this has occurred, it appears that some ageing has occurred prior to the day 1 measurements. Some evidence for prior ageing is the low shear force values at day 1 (Figure 2.2), although there is clearly further ageing after day 1. It is possible the heavy and fat carcasses cooled slower than expected or some ageing occurred after the samples were initially frozen

(see also section 5.3.1).

It is anticipated that other QTL affecting tenderness may exist within the population herein but have not been detected because there are limitations associated with the different types of tenderness measurements (Chapter 1.3). Moreover, although the cattle linkage mapping population herein is one of the largest in terms of informative

82 meioses, the design still suffers from the lack of power to detect QTL of very small effects as is the case in all livestock experiments.

Verification of QTL that have been discovered, however, is important if a large scale effort is to be made to identify the quantitative trait nucleotide (QTN) underlying the

QTL. Most of QTL discovered herein were only found in one sire family (Tables

3.11-3.12) and therefore, were not verified in the experiment. This is not unexpected as the alleles explaining the QTL are unlikely to be segregating in all sire families.

Because the number of sires used in cattle QTL mapping is limited, more families are required for QTL verification or results from other studies should be compared.

Table 3.11 QTL positions for adjusted shear force in each sire family. QTL from QTL from QTL model 1 models 2 + 3 BTA 5 (LD) Sire 2 (92 cM) Sire 2 (92 cM)

BTA 5 (ST) Sires 1+ 2 (96 cM) Sires 1+ 2 (96 cM)

BTA 18 (LD) Sire 1 (36 cM) Sire 1 (40 cM)

BTA 25 (ST) Sire 2 (0 cM) Sire 2 (0 cM)

BTA 29 (LD) Sire 1, Sire 2 (52 cM) Sire 1 (28 cM)

BTA 29 (ST) Sire 2 (52 cM) Sire 2 (16 cM)

Table 3.12 QTL positions for ageing rate in each sire family. QTL from QTL from QTL from QTL from BTA model 4 model 5 model 6 model 7 Sires 1+3 (48 Sires 1+3 (48 Sires 1+3 (48 Sires 1+3 (48 BTA 1 (ST) cM) cM) cM) cM) BTA 1 (LD) Sire 1 (64 cM) Sire 1 (48 cM) Sire 1 (52 cM) Sire 1 (52 cM) BTA 2 (ST) Sire 2 (0 cM) BTA 2 (LD) Sire 2 (56 cM) Sire 2 (56 cM) Sire 2 (104 cM) Sire 2 (104 cM) BTA 3 (ST) Sire 2 (79 cM) Sire 2 (76 cM) Sire 2 (76 cM) Sire 2 (76 cM) Sires 2+3 (44 Sires 2 +3 (44 BTA 3 (LD) Sire 3 (84 cM) Sire 3 (84 cM) cM) cM)

Table 3.12 QTL positions for ageing rate in each sire family (continued).

BTA 4 (LD) Sire 1 (40 cM) Sire 1 (40 cM) Sire 1 (40 cM) Sire 1 (40 cM) BTA 5 (LD) Sire 2 (8 cM) BTA 6 (ST) Sire 2 (72 cM) Sire 2 (72 cM) Sire 2 (72 cM) Sire 2 (76 cM) BTA 6 (LD) Sire 1 (0 cM) Sire 1 (0 cM) Sire 3 (116 cM) Sire 3 (116 cM) Sires 1+3 (116 Sires 1+3 (116 BTA 7 (ST) Sire 1 (120 cM) Sire 1 (120 cM) cM) cM) BTA 9 (ST) Sire 3 (52 cM) Sire 3 (52 cM) Sire 3 (48 cM) Sire 3 (48 cM) BTA 9 (LD) Sire 2 (28 cM) Sire 2 (28 cM) Sire 2 (28 cM) BTA 10 (LD) Sire 3 (48 cM) Sire 1 (0 cM) Sire 1 (0 cM) Sire 1 (0 cM) BTA 11 (ST) Sire 3 (80 cM) Sire 3 (80 cM) Sire 3 (80 cM) Sire 2 (28 cM) BTA 12 (LD) Sire 3 (12 cM) Sire 3 (12 cM) Sire 3 (16 cM) Sire 3 (16 cM) Sires 1+ 3 (0 Sires 1+2+3 (0 BTA 13 (LD) Sires 1+3 (0 cM) Sires 1+3 (0 cM) cM) cM) BTA 14 (ST) Sire 2 (40 cM) Sire 1 (68 cM) Sire 1 (68 cM) BTA 15 (ST) Sire 1 (0 cM) Sire 1 (0 cM) Sire 1 (0 cM) Sire 1 (0 cM) BTA 16 (LD) Sire 2 (0 cM) Sire 2 (0 cM) Sire 2 (4 cM) Sire 2 (4 cM) BTA 18 (ST) Sire 1 (80 cM) Sire 1 (80 cM) Sire 1 (80 cM) BTA 19 (ST) Sire 1 (44 cM) Sire 1 (44 cM) Sire 1 (48 cM) Sire 1 (44 cM) BTA 19 (LD) Sire 3 (28 cM) Sire 3 (28 cM) BTA 20 (ST) Sire 3 (56 cM) Sire 3 (56 cM) Sire 3 (56 cM) Sire 3 (56 cM) BTA 21 (LD) Sire 2 (0 cM) Sire 2 (4 cM) Sire 2 (4 cM) Sire 2 (4 cM) BTA 22 (ST) Sire 1 (56 cM) Sire 1 (56 cM) Sire 1 (56 cM) Sire 1 (52 cM) BTA 23 (ST) Sire 3 (68 cM) Sire 1 (8 cM) BTA 24 (LD) Sire 2 (0 cM) Sire 2 (0 cM) Sire 2 (0 cM) Sire 2 (0 cM) BTA 27 (ST) Sire 1 (0 cM) Sire 1 (0 cM) Sire 1 (0 cM) Sire 1 (0 cM) BTA 27 (LD) Sire 2 (32 cM) Sire 2 (32 cM) Sire 2 (28 cM) BTA 28 (ST) Sire 2 (20 cM) Sire 2 (20 cM) Sire 2 (24 cM) Sire 2 (20 cM) BTA 29 (ST) Sire 3 (4 cM) Sire 3 (4 cM) Sire 2 (60 cM) Sire 2 (60 cM)

There are some other studies that have reported QTL for tenderness in cattle.

Gutie´rrez-Gil et al., (2007) found three QTL (BTA 10 (79 cM), BTA 25 (32 cM) and BTA 29 (59 cM)) for tenderness as measured by sensory panels. Casas et al.,

(2000, 2001 and 2003) found 5 QTL for shear force (BTA 4 (20 cM), BTA 5 (65

84 cM), BTA 9 (cM unknown), BTA 20 (72 cM) and BTA 29 (54 cM)). Alexander et al., (2007) showed that QTL on BTA 5 (0 cM) and BTA 10 (95 cM) are associated with sensory panel tests and shear force, respectively. Keele et al., (1999) reported that there is a QTL on BTA 15 (28 cM) for shear force. Drinkwater et al., (2006) found 2 QTL on BTA 7 affecting shear force at 35 cM and 85 cM. Davis et al., (2007) found 2 QTL for compression and tenderness index on BTA 7 (68 cM) and one QTL for compression on BTA 10 (56 cM). Combining LD data from the same animals herein with data from the sister herd in New Zealand, Esmailizadeh et al., (2011) reported 3 QTL on BTA 3 (102 cM), BTA 12 (82 cM) and BTA 18 (72 cM) for shear force at different ageing days. In general, few of these identified QTL positions in the different studies have overlapped. The exceptions are the QTL on BTA 7, BTA 10 and BTA 29 where the QTL peaks have been reported within ~10 cM.

Comparing these QTL results with the QTL discovered herein revealed that the positions of the QTL for adjusted shear force and ageing rate do not clearly correspond to the previously discovered QTL (Table 3.13). The exception is the QTL on BTA 29 shown here to reflect the QTN within calpain 1. However, as the QTL mapping intervals are broad for cattle, it is quite possible that the QTL on BTA 7

(Drinkwater et al., 2006) and BTA 20 (Casas et al., 2003) overlap with those herein as these QTL peaks are within 20 cM. The best approach to refine the mapping intervals is to have more informative meioses (Mackay, 2001), but this is quite difficult given the time and expense of generating large numbers of progeny in livestock projects. So meta-analysis of published results may be the most viable option for the confirmation of QTL (Khathar et al., 2004; Silva et al.,

2011).Theoretically, the results can be from either other QTL linkage analyses or genome-wide association studies.

Table 3.13 Comparison of QTL positions herein with published QTL.

BTA Peak Trait Published tenderness QTL Location 1 48 cM ST ageing rate 3 76 cM ST ageing rate Esmailizadeh et al., (2011) 102 cM 4 40 cM LD ageing rate Casas et al., (2001) 4 cM LD + ST adjusted Alexander et al., (2007); 0 cM; 65 cM 5 92-96 cM shear force Casas et al., (2000) Drinkwater et al., (2006); 35 and 85 cM; 7 116 cM ST ageing rate Davis et al., (2007) 68 cM 11 80 cM ST ageing rate 13 0 cM LD ageing rate ST adjusted shear 72 cM 18 36 cM Esmailizadeh et al., (2011) force 19 44 cM ST ageing rate 20 56 cM ST ageing rate Casas et al., (2003) 72 cM LD adjusted shear 32 cM 25 0 cM Gutie´rrez-Gil et al., (2007); force Casas et al., (2000, 2003); 54 cM, 59 cM, LD + ST adjusted 29 52 cM Gutie´rrez-Gil et al., (2007); 60 cM shear force Esmailizadeh et al., (2011)

The QTL have led to discovery of the quantitative trait nucleotides (QTNs) in several genes that are consistently associated with tenderness traits in beef. These genes include calpastatin on BTA 7 (Chung et al., 2001; Barendse, 2002), lysyl oxidase on BTA 7 (Barendse, 2002; Drinkwater et al., 2006), calpain 3 on BTA 10

(Barendse et al., 2008) and calpain 1 on BTA 29 (Page et al., 2002). In addition, other candidate genes have been recently associated with beef tenderness traits, including the myogenic factor 5 (MYF5) gene on BTA 5 (Ujan et al., 2011a), the myogenin (MYOG) gene on BTA 16 (Ujan et al., 2011b), the ankyrin 1 gene (ANK1) on BTA27 (Aslan et al., 2010), calpain small subunit 1 (CAPNS1) gene on BTA18

(Iwanowska et al., 2011), the homogentisate 1, 2 dioxygenase (HGD) gene on BTA 1

(Zhou et al., 2010), and the peroxisome proliferator-activated receptor gamma

(PPARγ) gene on BTA 22 (Fan et al., 2011). However, none of the other genes are in

86 known QTL with the exception of MYF5, which is located in a tenderness QTL on

BTA5.

Fine mapping with more markers and more animals should provide a more precise position of the QTL for tenderness (Deng et al., 2000; Meuwissen and Goddard,

2000). However, more markers and more animals will not lead to verification of the

QTL if the same trait is not being mapped in the different populations. The inconsistencies between the QTL discovered for tenderness between these studies are most likely to be a result of different muscle types, ageing times, and types of tenderness measurements and experimental conditions. The results herein suggest that the same phenotypic measurements in the same muscle must be mapped if the

QTL are to be verified.

However, even with the same phenotypic measurements and the same muscle, the mapping results herein also showed different QTL segregating in each sire family

(Table 3.11 and 3.12) (Appendices 9 and 10). This reflects 2 other important factors, namely whether a particular QTL is segregating in a given sire family and which models are used in the analyses.

Another major difference between the studies is the use of different cattle breeds.

QTL will not be verified if the DNA variant causing the QTL does not exist or is in low frequency in a particular cattle breed (eg. the low frequency of the calpastatin

QTN variant in Bos taurus versus Bos indicus) or does not cause a large effect in the breed. Ideally, for selection purposes, the QTL will be independent of cattle breed.

However, it is difficult to determine at this point whether the QTL for tenderness have not been confirmed because of breed differences, trait measurements or an insufficient number of studies. It is of interest to note that many of QTL discovered

87 in the Jersey-Limousin progeny herein have been confirmed in a sister herd in New

Zealand (Morris et al., 2009). Thus, by using the same breeds and the same trait measurements, it is possible to confirm QTL even if the animals are raised and slaughtered under quite different conditions.

3.4.1 Potential candidate genes for adjusted shear force

Based on the QTL mapping results, candidate genes within the QTL were selected for further study herein (Chapter 4), based on their functions which may directly or indirectly affect tenderness However, it should be pointed out that many other potential candidate genes exist within these QTL regions. Although it was beyond the scope of this study to investigate all of these genes, they may be considered in the future if either the tenderness QTL are verified in additional linkage studies and/or the role of the genes in tenderness is confirmed in biological experiments.

Towards this end, several genes within the QTL were highlighted.

The QTL mapping herein revealed three major chromosomal regions for adjusted shear force on BTA 5, 18 and 29. The putative QTL on BTA 29 at 28 cM was independent of calpain 1 and therefore, was included in the candidate gene search.

Several particularly interesting genes within these QTL were noted (Table 3.14).

Table 3.14 Genes within major QTL adjusted shear force.

Gene name Position BTA Function cell division kinase 4 60Mb 5 Muscle cell differentiation cell division kinase 2 60Mb 5 Muscle cell differentiation MIR-interacting 61Mb 5 Stabilize myosin light chain saposin-like protein proteasome beta 10 34Mb 18 Proteasome subunit cysteine-rich protein 3 26Mb 29 Muscle cell differentiation

On chromosome 5, the QTL peak was located around 92 cM or 54 Mb. Of the 238 genes in the region (36 Mb to 62 Mb at the end of the chromosome), 3 genes were intriguing, 2 cell division kinases (CDK4 and CDK2) and a muscle structural protein

(CNPY2). The cell division kinase 4 or cyclin-dependent kinase 4 gene (CDK4) is located at 60 Mb on chromosome 5 and is a member of serine/ threonine protein kinase family. The serine/ threonine protein kinase family is an important enzyme system in signalling transduction. The protein is involved in cell proliferation and differentiation by inhibiting the cell cycle G1 to S phase transition. It also interacts with the muscle determination factor myogenin to control myogenesis (Zhang et al.,

1999) and myogenin has been associated with tenderness in Chinese cattle (Ujan et al., 2011b).

The cell division kinase 2 or cyclin-dependent kinase 2 gene (CDK2) is located at 60

Mb on chromosome 5. The functions of CDK2 are similar to the CDK4 gene because they are both in the serine/ threonine protein kinase family and regulate cell proliferation through the same signalling pathway. Myostatin regulation of hyperplasia or hypertrophy requires CDK2 mediation (Thomas, et al., 2000).

Because of the effect of myostatin on tenderness (Casas et al., 2000; Wheeler et al.,

2001; Lines, et al., 2009), CDK2 is another potentially important candidate gene, which may be related to the effects of myostatin on tenderness.

The MIR-interacting saposin-like protein (CNPY2), also called canopy 2 homolog

(TMEM4), is located at 61 Mb on chromosome 5. CNPY2 has the ability to stabilize the myosin light chain (MYL9) (Watanable, et al., 2007), which forms the motor protein in muscle. Given the ability of CNPY2 to maintain the stability of muscle structure, it may be a candidate for tenderness.

Another gene worthy of note was found for the QTL on BTA18, PSMB10. The QTL on BTA 18 had 306 genes in the region and the proteasome beta 10 subunit gene

(PSMB10) is located at 34 Mb on BTA 18. The PSMB10 gene encodes a subunit of proteasomes which has ATP-dependent proteolytic activity. The proteasome cleaves peptides from amino acid chains. However, the structure of muscle must be disassembled before proteasomes can act. Nevertheless, as the entire proteolytic system plays a critical role in the tenderization process, the proteasome subunit may warrant further investigation.

On chromosome 29, a new putative QTL was discovered at 20~28Mb. Amongst the genes in this region from 0 Mb to 40 Mb, there was another interesting gene identified, CSRP3. Cysteine-rich protein 3 (CSRP3) is located at 26 Mb and belongs to the CSRP family of LIM domain proteins. It is critical for the cell growth and cellular differentiation. Furthermore, CSRP3 is the positive regulator of myogenesis.

Lehnert et al., 2006 presented evidence that CSRP3 is involved in the regulation of myogenic differentiation of muscles in beef cattle.

3.4.2 Potential candidate genes for ageing rate

There were 3 major QTL for ageing rate on BTA 4, 13 and 19. Again several genes were noteworthy within these QTL. There were 153 genes in the region on BTA 4

(4.6 Mb ~ 32 Mb), the 34 genes in the region of BTA 13 (0 Mb ~ 4 Mb) and the 728 genes in the region on BTA 19 (9 Mb ~ 40 Mb).

For the QTL at 40 cM (14 Mb) on BTA 4, two genes were of interest, SAMD9 and

SHFM (Table 3.15). SAMD9, located at 10 Mb on BTA 5, encodes a protein of the

SMAD system. SAMD9 is one of the five receptor regulated SMADs which are

90 critical to the R-SMADs intracellular pathway. BMPs (bone morphogenetic proteins) and GDF2 (growth differentiation factor 2) are mediated by the SMAD9. Thus,

SMAD9 has a function on cell proliferation (Li, et al., 2007a).

Table 3.15 Genes within major QTL for ageing rate.

Gene name Position BTA Function SAMD9 10Mb 4 Transduction pathway Split hand/foot malformation 14Mb 4 Proteolysis type 1 protein MYH1_BOVIN 30Mb 19 Myosin heavy chain, skeletal muscle, adult 1 MYH2_BOVIN 30Mb 19 Myosin heavy chain, skeletal muscle, adult 2 myosin, heavy chain 3, skeletal muscle, MYH3 30Mb 19 embryonic

The split hand/foot malformation type 1 protein gene (SHFM1) at 14 Mb on BTA 4 produces a subunit of the 26S proteasome. Although the proteasome does not have the ability to digest intact myofibrils directly, proteolytic activity may still be important to tenderness as discussed earlier.

The QTL on BTA 13 for ageing rate was well supported statistically and had a large effect. However, in examining to 20 cM around the putative QTL peak, there were only 3 genes. Unfortunately, none of these genes (phospholipase C, beta 1 (PLCB1) at 0.8 Mb, LOC513094 at 2.4 Mb, and synaptosomal associated protein 25 at 3.3 Mb) were good candidates for ageing rate although PLCB1 is involved in myogenesis albeit indirectly. Phospholipase C, beta 1 is an enzyme that affects cell differentiation and myogenesis by catalyzing the formation of the second messengers, inositol 1,4,5-trisphosphate (IP3) and the 1,2-diacylglycerol (DAG) from the phosphatidylinositol 4,5-bisphosphate (PIP2). LOC513094 is a hypothetical predicted protein of unknown function. Synaptosomal associated protein 25 is part of

91 a complex responsible for membrane fusion. One of these genes may affect tenderness by an unknown mechanism or there may be other genes in the region which have not yet been identified, as the bovine genome is still not well annotated.

The genes in the QTL at 44 cM (approximately 22 Mb) on BTA 19 were also not particularly remarkable except for 3 of the genes that encode heavy chain myosins

(MYH1, MYH2 and MYH5). Myosins are major structural components of myofibrils and part of the apparatus driving muscle contraction. As structural proteins, there could potentially be link to tenderness.

3.4.3 Conclusion

Currently, the number of known DNA variants in bovine genes is very limited

(Matukumalli et al., 2009). Consequently, when embarking on gene association studies in cattle, one must select the candidate gene with care as the genes must be first sequenced to find potential variants for genotyping. The genes identified above for adjusted shear force and ageing rate were not obviously directly related to tenderness. However, once reliable DNA variants are available across the bovine genome, these candidate genes may still be worth further investigation as the QTL are of sizeable effect and significance (Appendices 9-10). Moreover, when the bovine genome is better annotated, it would be also worth re-scanning the QTL regions for additional candidates. However, using all the QTL results from the study herein and the study of Lines (2006), a set of stronger candidate genes were selected

(Chapter 4) for sequencing in order to identify appropriate variants for association studies (Chapter 5).

Chapter 4: Candidate Genes

4.1 Introduction

To determine if a specific gene affects a particular trait, it is common to perform association studies in which DNA variants within the gene are statistically analysed in a population with phenotypic variation in the trait of interest. Usually, the genetic effect of the gene within the population is determined by one of two types of statistical models. The first type is a general linear model which is used to evaluate the effect of a gene by fitting all factors as fixed factors/covariates in the model (Curi et al., 2009). The other type is a mixed model using the restricted maximum likelihood method (REML), in which the factors are fixed or random in the model

(Riley et al., 2003; Casas et al., 2006a; Esmailizadeh et al., 2008).

In either approach, those factors considered to affect the trait of interest are incorporated into the model as well as the genotypes of the DNA variants within the gene. As an example, for tenderness, factors known to affect meat toughness, such as breed, diet, farm of origin, age and sex, may be included (Hiner et al., 1953; Reagan et al., 1976; Shackelford et al., 1994; Vestergaard et al., 2000; Robinson et al., 2001;

Riley et al., 2003). Therefore, when determining the size of effect of a given gene, any factor affecting the trait of interest is usually taken into account dependent on the conditions of the study. By taking these factors into account, any DNA variants that directly affect a trait or are closely linked to the causative variant should explain more of the remaining variation in the trait.

Using such linear regression methods, a number of variants in several genes (MSTN,

CAPN1, CAPN3, CAST and LOX) have been associated and confirmed to affect tenderness (Barendse, 2002; Juszczuk-Kubia et al., 2004; Casas et al., 2006a;

Barendse et al., 2008; Lines et al., 2009; Frylinck et al., 2009). However, a large

94 proportion of the genetic variation in tenderness remains unaccounted for, and hence, the search for additional genes controlling tenderness herein.

In order to estimate the effect of the variants of the candidate genes, the candidate genes must be first sequenced in order to locate potential variants for genotyping and association analysis. This is necessary as the bovine genome sequence is currently not of sufficient depth to provide DNA variants within the candidate genes. In this study, 6 candidate genes (namely, MYL7, MYO1G, MBNL3, CAPN4, CAPN5, and

LOXL1) were chosen to be sequenced for variants based on their biological functions and the QTL mapping results herein and elsewhere (Lines, 2006; Esmailizadeh,

2006).

4.2 Materials and Methods

QTL results for the two newly derived traits (adjusted shear force and ageing rate) and for other tenderness traits including shear force at specific ageing days, hydroxyproline content and compression (available from Lines, 2006 and

Esmailizadeh, 2006) were used for candidate gene identification. In order to find all the genes within the 40 cM region of the putative QTL peaks, the linkage map units

(cM) were converted into physical map units (base pairs). This was done by identifying two markers at each end of the 40 cM region or relatively close to the two extreme ends (http://www.animalgenome.org/cgi-bin/QTLdb/BT/viewmap). The positions of these two markers in base pairs were then located in the bovine genome

(http://www.animalgenome.org/cgi-bin/QTLdb/BT/viewmap) and all genes within

20 cM flanking either side of the QTL maximum peak position were listed using the

Ensembl search engine (http://www.ensembl.org/Bos_taurus/Info/Index). The function of the genes was determined from Gene Cards (http://www.genecards.org/)

95 and the literature, and then candidate genes were selected based on their potential involvement with tenderness.

The coding regions, untranslated regions, promoters, and flanking regions of all candidate genes were sequenced using the genomic DNA of the 3 F1 sires in order to find DNA variants within the candidate genes. Primers for PCR were designed for each gene region using the BTAU 3 bovine genome sequence available from the

Ensembl database (http://www.ensembl.org/index.html). The amplified regions were sequenced for each sire, and the sequencing data were aligned and compared using

Sequencher software to locate any variants. The DNA variants were verified by sequencing the grandparents of the progeny to ensure Mendelian inheritance and to establish haplotypes. The single nucleotide polymorphisms (SNPs) chosen for genotyping from each gene were confirmed variants that were not in linkage disequilibrium with each other. The variants were genotyped for the association studies using gel electrophoresis, the Illumina system and high resolution melt

(HRM).

4.2.1 Optimization of PCR

After selecting the candidate genes, the coding regions of all candidate genes were sequenced using the genomic DNA of the 3 F1 sires in order to find DNA variants within the candidate genes. The DNA variants were verified by sequencing the grandparents of the progeny to ensure Mendelian inheritance and to establish haplotypes. Primers for PCR were designed using the software Primer 3

(http://frodo.wi.mit.edu/primer3/). The size of the PCR products was limited to less than 850 base pairs to meet the performance of the DNA sequencer. Primer dimers can decrease or inhibit the performance of PCR. Therefore, primer dimers and

96 hairpin structures needed to be avoided and their potential presence was checked using the software IDT SciTool Oligoanalyser 3.1

(http://www.idtdna.com/analyzer/Applications/OligoAnalyzer/Default.aspx).

Normally, each reaction was in a volume of 25 μl and contained 13.5 μl H2O, 50 ng

DNA template, 2.5 μl dNTPs (1.25 mM), 2.5 μl MgCl2 (25 mM), 2.5 μl 10x buffer,

2.5 μM forward primer (Geneworks), 2.5 μM reverse primer (Geneworks) and 0.5 U

Amplitaq Gold Polymerase (Roche). If the PCR results showed that there were extra bands or the band was too faint, the concentration of MgCl2 was adjusted from 1~4 mM to eliminate the extra bands or enhance the PCR product. If the PCR was not efficient because of a high GC content, betaine 5 μl (5M) (Sigma) was added.

FastStart Taq DNA polymerase (2.5 μl GC-RICH solution 5x, 2.5 μl PCR Buffer 10x

+ MgCl2 accompanied) (Roche) was also used to overcome the high GC content.

Four PCR programs (three touchdown programs and one gradient program) were utilized in order to find out the most suitable PCR conditions. Touchdown programs were set at 3 different ranges of annealing temperatures (listed below). If the three touchdown programs did not produce an acceptable PCR product or the change in the magnesium concentration did not improve the outcome of PCR, then gradient

PCR was used to determine the best annealing temperature.

Touchdown Program 1:

Cycle 1: 95 ºC for 10 minutes (1 repeat)

Cycle 2: 95 ºC for 1 minute

70 ºC for 1 minute

72 ºC for 1 minute (1 repeat)

Cycle 3: 95 ºC for 1 minute

68 ºC ~58 ºC for 1 minute (touchdown)

72 ºC for 1 minute (39 repeats)

Cycle 4: 72 ºC for 10 minutes (1 repeat)

4 ºC for 4 minutes

Touchdown Program 2:

Cycle 1: 95 ºC for 10 minutes (1 repeat)

Cycle 2: 95 ºC for 1 minute

60 ºC for 1 minute

72 ºC for 1 minute (1 repeat)

Cycle 3: 95 ºC for 1 minute

59 ºC ~50 ºC for 1 minute (touchdown)

72 ºC for 1 minute (39 repeats)

Cycle 4: 72 ºC for 10 minutes (1 repeat)

4 ºC for 4 minutes

Touchdown Program 3:

Cycle 1: 95 ºC for 10 minutes (1 repeat)

Cycle 2: 95 ºC for 1 minute

55 ºC for 1 minute

72 ºC for 1 minute (1 repeat)

Cycle 3: 95 ºC for 1 minute

54 ºC ~45 ºC for 1 minute (touchdown)

72 ºC for 1 minute (39 repeats)

Cycle 4: 72 ºC for 10 minutes (1 repeat)

4 ºC hold

Gradient program:

Cycle 1: 95 ºC for 10 minutes (1 repeat)

Cycle 2: 95 ºC for 1 minute

70 ºC ~50 ºC for 1 minute (gradient)

72 ºC for 1 minute (40 repeats)

Cycle 3: 1 72 ºC for 10 minutes (1 repeat)

4 ºC hold

The reactions were loaded into 96 well Thermoquick PCR plates Type 2 (Greiner

Bio-one) or 96 well full skirted PCR microplates (Axygen). PCR was performed using one of three thermocyclers, Hybaid OmniGen, Corbett Palm-Cycler or Peltier

Thermal Cycler DNAEngine (Bio-Rad). When using the Hybaid OmniGen system, each reaction was covered with a drop of mineral oil to prevent the mixture from evaporating.

4.2.2 Electrophoresis

Electrophoresis were performed with 2.0% agarose (Applichem) gels in 1X TAE buffer (Appendix 11) for 40-45 minutes at 110 V or 3.0% agarose (NuSieve 3:1

Agarose) gel for 75 minutes in 1X TAE buffer at 90 V (for higher resolution). 5 μl of

PCR product and 1 μl of loading dye (Blue/Orange 6x, Promega) were used for the electrophoresis gels. The DNA marker ladder was composed of 1 μl of loading dye

(Blue/Orange 6x, Promega) and 1 μl of pGEM® DNA Markers.

4.2.3 Staining

Gels were stained in 0.5 μg/ml ethidium bromide for 10-15 minutes and then a Gel

Documentation 1000 system (Biorad) was used to visualise the PCR products under

UV illumination.

4.2.4 Purification for PCR products

After a single PCR product was identified from the stained gel image, PCR products were purified for sequencing. Purification kits (Ultra Clean PCR Clean-up Kit,

MoBio, Appendix 12) were used to remove contaminants such as the dNTPs, magnesium and buffer. After the purification, 2 μl of post-purified PCR products was checked by gel electrophoresis for concentration and purity.

4.2.5 Sequencing

Post-purified PCR products were dried at 60 ºC for 60~80 minutes and then measured for concentration using a spectrophotometer ND-100 (NanoDrop). 50~100 ng of post-purified PCR products were used for sequencing. The sequencing reactions included 2 μl Big Dye Terminator (Applied Biosystems), 1 μl glycogen

(Roche), 1 μl (5 μM) primer (Geneworks) and H2O added to a volume of 10 μl in total., The sequencing was performed on either Corbett Palm-Cycler or Peltier

Thermal Cycler DNAEngine (Bio-Rad) using the following program:

Cycle 1 95ºC, 5 seconds (1 repeat)

Cycle 2: 96 ºC, 30 seconds (24 repeats)

50ºC, 15 seconds

60ºC, 4 minutes

Cycle 3 4 ºC hold

The sequencing products were collected in new tubes and precipitated using by 75% isopropanol (Appendix 13). Sequencing reaction products were separated by capillary electrophoresis using an ABI 3730 analyzer at the Institute of Medical and

Veterinary Science (IMVS), Adelaide, South Australia. The software Sequencher 4.9

Demo (http://www.genecodes.com/) was used to find the variants in the sequences.

100

4.2.6 Genotyping

Variants found in the candidate genes were selected for genotyping based on the haplotypes, avoiding any variants in linkage disequilibrium. All potentially functional variants were genotyped. Genotyping was performed for single nucleotide polymorphisms (SNPs) using high resolution melt (HRM) (Reed and Wittwer,

2004), or the Illumina system located at DPI, Atwood, Victoria. Standard gel electrophoresis (using NuSieve 3:1 agarose at 90 volts for 40 minutes) was used for genotyping the insertion/deletions (in/dels). Gels were stained in 0.5 μg/ml ethidium bromide for 10-15 minutes and then the Gel Documentation 1000 system (Biorad) were used to visualise the PCR products under UV illumination.

A Rotor-Gene 6000 (Corbett) machine was used to perform the HRM genotyping.

The Rotor-Gene 6000 was used to perform PCR, melting and fluorescence detection in a single run. The volume of HRM reactions was 20 μl in total. The proportion of each ingredient and the conditions were dependent on the optimized PCR conditions for the specific primer set (Appendix 14). HRM reactions included 0.8 μl EvaGreen dye (Biotium Inc) for each reaction. EvaGreen dye is a saturated florescent dye used for identifying genotypes based on the different release rates of the EvaGreen dye from the double stranded PCR products. PCR products were amplified on either

Corbett Palm-Cycler, Peltier Thermal Cycler, DNAEngine (Bio-Rad) or the Rotor-

Gene 6000. The PCR on Rotor-Gene 6000 used the following program:

Hold 95ºC, 10 minutes

Cycling (40 repeats) 95ºC, 50 seconds

Annealing, 50 seconds (dependent on the optimized

PCR conditions)

72ºC, 50 seconds

101

Hold 2 72ºC, 10 minutes

Hold 3 25ºC, 10 minutes

HRM Melting temperature range (normally 70ºC-97ºC)

The HRM was performed by melting the products at 95 ºC for 10 minutes, then decreasing to 70 ºC and slowly increasing the temperature to 97 ºC (0.1 ºC/step). The genotype assignments were acquired using the Rotor-Gene 6000 Series Software, version 1.7.

4.2.7 Prediction of protein structure

Two protein structure prediction web-based programs (Predict Protein and PS2) were used to predict the protein structure of potentially functional amino acid variants

(http://www.predictprotein.org/and http://ps2.life.nctu.edu.tw).

4.2.8 Genomic sequence alignments

The alignments of the genomic sequences were carried out by using the function of genomic alignments in Ensembl (http://www.ensembl.org/index.html).

4.3 Results

The candidate genes were selected based on two criteria. One was their role in physiological pathways believed to be potentially involved in tenderisation. The other was the genomic location of the gene relative to the position of a putative QTL.

The QTL for tenderness were chosen from herein and from previous studies (Lines,

2006; Esmailizadeh, 2006). In addition, supporting QTL information from the sister herd in New Zealand was provided from Dr. C. Morris and N. Cullen (AgResearch) although the direct comparisons were not possible as the tenderness measurements differed and there were no data for the ST muscle. Lastly, genotypes were already

102 available from gene variants located in many of the QTL (see Chapter 5). Therefore, the candidate genes chosen for sequencing were from QTL that had good statistical support but no gene variants near the QTL peak.

In total, six candidate genes (MYL7, MYO1G, MBNL3, CAPN4, CAPN5, and LOXL1) were selected for sequencing based on their function and location (Appendix 15). All the variants were identified in these genes by sequencing the coding regions, untranslated regions and immediate flanking regions of the genes in the three Davies gene mapping sires. DNA variants were confirmed by sequencing the parents of the sires. DNA variants were then selected for genotyping in the progeny based on the grandparental haplotypes.

Of the six candidate genes, 1 gene encodes a muscle structural protein (MYL7), 1 gene encodes a protein involved in lysosomes (MYO1G), 1 gene encodes a regulator of muscle differentiation (MBNL3), 2 genes code for calpain system subunits

(CAPN4 and CAPN5) and 1 gene code for a protein involved in collagen cross- linking (LOXL1). Thus, the genes represent a variety of pathways potentially affecting tenderisation.

Myosin regulatory light chain 7 (MYL7 or MLC2a) is a member of the myosin family of motor proteins. The structure of the myosin itself consists of two heavy chains and four light chains. Two heavy chains form the head and tail of the main structure of myosin. Four light chains are responsible for maintaining the structure of myosin. The MYL7 gene encodes the protein that forms one of the myosin regulatory light chains, and is important for the stability of the muscle structure (Hailstones et al., 1992). The MYL7 gene is located at 79 cM on chromosome 4.

103

Myosin IG (MYO1G) is a member of the myosin 1 family. Class 1 myosins are motor proteins with the ability to bind to actin filaments and cellular membranes

(Krendel and Mooseker, 2005; Redowicz, 2007). These myosins serve as divalent crosslinkers and have many functions including lysosomal vesicle transport, adhesion, endocytosis and exocytosis (Adams and Pollard, 1986; Krendel and

Mooseker, 2005). The proteolytic enzymes, cathepsins, are lysosomal. The MYO1G gene (79 cM) was on cattle chromosome 4.

The muscleblind-like 3 (MBNL3) gene is located at 116 cM on cattle chromosome 17 near a QTL for ST compression (Lines, 2006). MBNL3 inhibits terminal muscle differentiation. The MBNL3 gene has been associated with myotonic dystrophies

(DM) caused by different polymorphic patterns in the repeat regions of the myotonic dystrophy type 1 gene, DM1 (DMPK) and the myotonic dystrophy type 2 gene, DM2

(CNBP/ZNF9) (Carango et al., 1993; Krahe et al., 1995; Miller et al., 2000;

Kanadia et al., 2003). Myotonic dystrophy (DM) is a dominant, inheritable multisystem disease. Defects in the DM1 gene cause muscle weakness and myotonia

(Carango et al., 1993; Krahe et al., 1995; Mulders et al., 2010). Defects in the DM2 gene cause muscle weakness, wasting, myotonia, and cataracts plus cerebral, endocrine and cardiac abnormalities (Finsterer et al., 2002). The product of the

MBNL3 gene regulates the expression of these DM1 and DM2 genes by co-localizing with the repeat regions of DM1 and DM2 genes. Defects in the MBNL3 gene can be lethal in severe conditions as reported in human and mice studies. Less severe defects in the MBNL3 gene result in myotonic dystrophy (Carango et al., 1993;

Miller et al., 2000 and Kanadia et al., 2003). Therefore, variants of the cattle MBNL3 gene may affect muscle structure and hence, tenderness.

104

The calpain, small subunit 1 (CAPN4 or CPNS1) gene is located at 41.9 cM on cattle chromosome 18 within the QTL for adjusted shear force. Calpain is a proteolytic enzyme complex known to be a major factor contributing to the tenderisation. This calcium-dependent cysteine protease is comprised of two large subunits (µ-calpain and m-calpain) (Goll et al., 2003). The protein encoded by CAPN4 gene is a small component of both of the two large subunits (Arthur et al., 2000). DNA variants in

CAPN1 have been associated with effects on tenderness in cattle and chickens (Page et al., 2002; Zhang et al., 2008). Therefore, CAPN4 can be reasonably postulated as a candidate gene for tenderness.

The LOC536988 gene corresponds to the CAPN5 gene. Calpain 5 is one the large subunits of calpain without the calmodulin-like domain IV. Its characteristics are different from other known calpains related to tenderness (Matena et al., 1998).

However, given the importance of the calpain system, calpain 5 was regarded as a potential candidate gene. Thus, the gene LOC536988 located at 49 cM on cattle chromosome 15 was selected as one of the candidates as it underlies an LD shear force QTL (Esmailizadeh, 2006).

Lysyl oxidase-like 1 (LOXL1) is a member of the lysyl oxidase family. LOXL1 is believed to have the same function as LOX in stabilizing the structure of elastin by forming collagen cross-links (Liu et al., 2004; Thomassin et al., 2005). The LOXL1 gene is located at 33 cM on cattle chromosome 21 near an ST QTL (Lines, 2006) and therefore, was also chosen as a candidate gene.

4.4 DNA Variants

Prior to identifying the relationship between these candidate genes and traits related to tenderness in association studies, single nucleotide polymorphisms (SNP) and

105

insertion/deletions were discovered by sequencing the genomic DNA from the three

Davies gene mapping sires. The variants that cause non-synonymous amino acid

substitutions or may affect splicing sites, transcription factor binding sites or mRNA

stability were genotyped in the progeny for association studies. If variants were only

found in the non-coding regions, the DNA variants used for genotyping in the

progeny were selected based on the grandparental haplotypes to avoid markers in

linkage disequilibrium. There were 48 variants discovered in the six candidate genes

(Appendix 16).

4.4.1 MYL7 variants

The MYL7 gene contains 6 exons. There were only 2 SNPs discovered in intron 2

(Table 4.1). The first SNP1 (C/G) was located 46 bases away from the exon 2 and

the second SNP2 (G/T) was 71 bases from the exon 2 (Table 4.1.). Since only two

SNPs were discovered in the MYL7 gene, both of these SNPs were genotyped in the

progeny.

Table 4.1 DNA variants in MYL7 Sire genotypes Name DNA variant Location 361 368 398 Potential effect MYL7 SNP1 C/G Intron 2 C/C G/C G/G None MYL7 SNP2 G/T Intron 2 G/G G/G G/T None

4.4.2 MYO1G variants

The MYO1G gene includes 22 exons. Three SNPs were identified in the promoter

region. Another three SNPs were discovered in three different introns (introns 1, 7

and 13). One silent SNP (based on the BTAU 3 data in Ensembl) was found in exon

7 (Table 4.2). Since there were no variants likely to cause significant changes in the

function of the peptide, MYO1G SNP2 and MYO1G SNP3 were selected for

106 genotyping. However, according to the most recent database assembly of the bovine genome (BTAU 4), the SNP in exon 7 is a potentially functional SNP. The substitution of an adenine with guanine leads to the replacement of arginine255 to glutamine (Figure 4.1). The R groups for arginine and glutamine are in different classes and have different polarities which could affect the hydrophilicity of the protein and change the function of MYO1G. Therefore, the substitution is considered non-conservative and this SNP was also genotyped.

Table 4.2 DNA variants in MYO1G. Sire genotypes

Name DNA variant Location 361 368 398 Potential effect MYO1G SNP1 C/T 5’ flanking C/T T/T T/T Promoter MYO1G SNP2 C/T 5’ flanking C/T C/C C/C Promoter MYO1G SNP3 G/A 5’ flanking G/G G/A G/G Promoter MYO1G SNP4 C/T Intron 1 C/T C/C C/C MYO1G SNP5 G/A Exon 7 G/A A/A A/A R255> Q255 MYO1G SNP6 C/T Intron 7 C/T C/C C/C MYO1G SNP7 G/T Intron 13 G/G G/G G/G

4.4.3 MBNL3 variants

The MBNL3 gene includes nine exons. There was only an insertion/deletion discovered in intron 3 (Table 4.3). In intron 3, there is a series of thymidine residues close to exon 4, and the number of thymidine residues in the three sires was different.

Sire 368 was a homozygous carrier for 15T, while the other 2 sires were homozygous for 16T (Figure 4.1). However, genotyping the grandparents revealed an unusual phenomenon in that one of the parents of sire 368 was homozygous for

17T and the other was homozygous for 16T (Table 4.4). Since this phenomenon did not fit the pedigree and as the polymorphic variant is comprised of a series of thymidines, it is most likely that the discrepancy in the sequencing results was 107 caused by a DNA polymerase error during PCR. Given that the sires were homozygous, the difficulties of accurately genotyping this variant, and the question of non-Mendelian inheritance, this insertion/deletion could not genotyped in the progeny. Consequently, MBNL3 could not be analysed as part of the association study. This is unfortunate as the variation in the thymidine repeat may affect the splicing of MBNL3 mRNA given its location.

Table 4.3 DNA variants in MBNL3. Sire genotypes Name DNA variant Location 361 368 398 Potential effect MBNL3 in/del T Intron 3 16T/16T 15T/15T 16T/16T Splicing

Table 4.4 MBNL3 in/del genotypes of sire 368 parents. Parents of 368 75 266 Genotype 17T/17T 16T/16T

AAAGAATATGTACTGTTGCTCTTACATTCACTGGCAAAATAAATAAAAACT GATAACCTTCCAAGCGTCAATGTGATTTTTTTTTTTTTTTTAAGGTTTGCC GTGAATTTCAGCGAGGTAATTGTACCCGTGGTGAGAATGATTGCCGC TATGCTCACCCTACTGATGTGTCCATGATTGAGACAAGTGATAATACT

Figure 4.1 Location of insertion/deletion in MBNL3 intron 3. Insertion/deletion is highlighted in yellow with the exon 4 in bold.

4.4.4 CAPN4 variants

The CAPN4 gene consists of eleven exons. In total, there were seven variants identified (Table 4.5). Two insertion/deletions were discovered. One of the insertion/deletions was in the 5’UTR which may affect the regulation of transcription.

The other insertion/deletion was in exon 2. This variant is polymorphic due to a variation in the number of GCC repeats encoding glycine in exon 2. In the

108 population used in this project, there were 3 different variants of repeats identified:

(GCC)8, (GCC)10, and (GCC)12. There were no individuals homozygous with two sets of twelve GCC repeats found. This insertion/deletion did not change the polypeptide sequence except the number of glycines. The series of glycine residues forms a loop structure (Figure 4.2). The change in the number of glycines may modify the tertiary structure of the CAPN4 protein and hence, could influence the activity of CAPN4. Therefore, the two insertion/deletions of CAPN4 were chosen for genotyping in the progeny.

Another five SNPs were identified in exon 2 and one SNP was found in exon 11.

However, all were silent variants and unlikely to cause any effect on the function of

CAPN4. Therefore, these SNPs were not genotyped in the progeny.

Table 4.5 DNA variants in CAPN4.

Sire genotypes Name DNA variant Location 361 398 Potential effect CAPN4 in/del AC 5’ UTR - AC/-AC + AC/+AC Regulation CAPN4 SNP1 A/C Exon 2 AA AA Silent CAPN4 SNP2 A/C Exon 2 AC AC Silent CAPN4 SNP3 A/C Exon 2 CC CC Silent CAPN4 SNP4 T/C Exon 2 TT TC Silent CAPN4 3 base GCC Exon 2 10/10 GCC 8/10 GCC Glycine loop repeat CAPN4 SNP5 T/C Exon 2 TC CC Silent CAPN4 SNP6 T/C Exon 11 TC CC Silent

109

Figure 4.2 Predicted protein structure for CAPN4 exon 2 (from Predict Protein). Series of glycines in CAPN4 exon 2 is highlighted in the orange dotted box. (H = alpha-helix, E = beta-sheet, L = loop.)

4.4.5 CAPN5 variants

There are 19 exons in the CAPN5 gene (Table 4.6). Two SNPs were identified in the

5’UTR. Fourteen SNPs were discovered in seven exons (exons 2, 4, 5, 7, 10, 11, and

12). Moreover, there were another seven SNPs in five introns (introns 1, 4, 6, 5, 7, 9 and 11). Interestingly, there were two potentially functional SNPs identified in exons

7 and 11, which cause amino acid substitutions.

The first potentially functional SNP (T/C) in exon 7 changes arginine338 to cysteine.

The polarities for arginine and cysteine are different, and this may affect the function of CAPN5. The prediction of the 3D protein structure for this section of CAPN5

(from 2 to 522 amino acids) clearly showed differences in the folding structures, the number of α loops and the length of the β sheets with this amino acid change (Figure

4.3). The protein prediction suggests that this SNP in exon 7 could affect the activity of CAPN5 due to a large conformational change in the protein tertiary structure.

110

Table 4.6 DNA variants in CAPN5.

Sire genotypes Potential Name DNA variant Location 361 368 398 effect CAPN5 SNP1 G/A 5’ UTR AA AG GG Regulation CAPN5 SNP2 G/A 5’ UTR AA AG GG Regulation CAPN5 SNP3 T/C Intron 1 TC CT TT CAPN5 SNP4 T/C Exon 2 TC CC CC Silent CAPN5 SNP5 T/C Exon 2 TC TC TT Silent CAPN5 SNP6 T/C Exon 4 TC TC TC Silent CAPN5 SNP7 T/C Exon 4 TC TC TC Silent CAPN5 SNP8 T/C Intron 4 TC TC TC CAPN5 SNP9 C/G Exon 5 CG CG CG Silent CAPN5 SNP10 C/T Exon 5 CT CT CT Silent CAPN5 SNP11 G/T Exon 5 GT GT GT Silent CAPN5 SNP12 C/G Intron 5 CG CG CG CAPN5 SNP13 T/A Intron 5 TA TA TA CAPN5 SNP14 A/G Intron 5 AG AG AG CAPN5 SNP15 A/G Intron 6 AG AG AG CAPN5 SNP16 T/C Exon 7 TC CC CC R358 > C358 CAPN5 SNP17 G/A Exon 7 GA GA GA Silent CAPN5 SNP18 T/C Intron 7 CC TC TC CAPN5 SNP19 T/C Intron 9 TC TC TC CAPN5 SNP20 T/C Exon 10 TC TC TC Silent CAPN5 SNP21 G/A Exon 11 GA GA AA R556 > H556 CAPN5 SNP22 G/T Intron 11 GT GT GG CAPN5 SNP23 T/C Exon 12 TC TC TT Silent CAPN5 SNP24 A/T 3’UTR AA AA AT Silent CAPN5 SNP25 G/C 3’UTR GG GG GC Silent

111

A B

Figure 4.3 Predicted protein structure for CAPN5. (A) T allele and (B) C allele for the SNP in exon 7 of CAPN5 (predicted by (PS)2 ). Arrows highlight differences.

Furthermore, aligning the sequences of CAPN5 for twelve species demonstrated that the region including the non-synonymous SNP in exon 7 is highly conserved (Figure

4.4). Consequently, the non-synonymous SNP in exon 7 was chosen for genotyping in the progeny and association studies.

Figure 4.4 Genomic alignment of sequence of CAPN5 exon 7 (from Ensembl). Box indicates location of the potentially functional SNP in exon 8 for 12 species.

The other important potentially functional SNP (G/A) in exon 11 changes the amino acid arginine556 to histidine. The polarities for arginine and histidine are the same.

This implies the amino acid change may not affect the function of CAPN5.

Unfortunately, the prediction of the 3D protein structure for this section of CAPN5

112 was only possible from 2 to 522 amino acids using the currently available software and did not include the position of the amino acid (arginine556 to histidine). However, the genomic alignment with twelve species for CAPN5 showed that the sequence is highly conserved (Figure 4.5). Hence, this SNP was also chosen for genotyping.

Figure 4.5 Genomic alignment of sequence of CAPN5 exon 11 (from Ensembl). Box indicates location of the potentially functional SNP in exon 12 for 12 species.

4.4.6 LOXL1 variants

The LOXL1 gene has 7 exons. Three SNPs were identified (Table 4.7). One non- synonymous SNP (A/G) in exon 1 was discovered. This non-synonymous SNP changes serine209 to glycine. The amino acid change may cause an effect on protein function because a hydrophilic amino acid has been substituted with a hydrophobic amino acid, resulting in large switch in polarity. Nevertheless, based on the Predict

Protein results, the amino acid change is not likely to change the secondary loop structure (Figure 4.6). However, the sequence for the region containing this SNP is conserved (Figure 4.7). Consequently, the SNP in exon 1 was still the best candidate

SNP for representing the effect of LOXL1 gene in association studies and was genotyped.

113

Table 4.7 DNA variants in LOXL1.

Sire genotypes Name DNA variant Location 361 368 398 Potential effect LOXL1 SNP1 A/G Exon 1 GA GA GG S209>G209 LOXL1 SNP2 C/T Intron 1 CC CC CT LOXL1 SNP3 G/A Intron 2 GC GC GA

Figure 4.6 Predicted protein structure for LOXL1 exon 1 (from Predict Protein). The presence of the serine (left) or glycine (right) residue in exon 1 does not alter the loop structures. (Predict Protein data base; http://www.predictprotein.org/) H = alpha-helix, E = beta-sheet, L = loop.)

Figure 4.7 Genomic alignment of LOXL1 exon 1 (from Ensembl). Box indicates the location of the potentially functional SNP in exon 1 for 12 species.

4.5. Discussion

Forty eight variants were identified by sequencing six candidate genes (Appendix

16). Of these 48 variants, ten variants were selected for genotyping in order to investigate the relationship between these genes and tenderness traits (Chapter 5).

Apart from the silent SNPs and the SNPs in introns, two insertions/deletions were

114 discovered in the 5’UTR and exon 2 of CAPN4, and 4 non-synonymous amino acid substitutions were identified including CAPN5 exon 7 (arginine338 to cysteine),

CAPN5 exon 11 (arginine556 to histidine), MYO1G exon 7 (arginine255 to glutamine) and in LOXL1 exon 1 (serine209 to glycine). Little is known about the structure of these proteins so the potential effects of these changes on protein activity could not be predicted. Therefore, all the variants that might affect protein function were specifically chosen for genotyping.

By comparing variants discovered in this study and variants in the most recent

Ensembl database update (http://www.ensembl.org/index.html) and NCBI

(http://www.ncbi.nlm.nih.gov/SNP/index.html), two non-synonymous SNP (CAPN5-

SNP21 and LOXL1-SNP1) and one other SNP (MYL7-SNP2) have been previously reported as rs41774805, rs109700555 and rs109541226. The remaining variants have not been previously identified, including the non-synonymous SNPs in MYO1G

(SNP5) and CAPN5 (SNP16) and the insertion/deletions in CAPN4 which are likely to affect protein function. This suggests that despite recent efforts such as the development of the cattle HAPMAP (Childers et al., 2010), deep sequencing of many animals in more breeds is required to identify DNA variants within bovine genes.

The other non-synonymous SNPs reported in these genes previously were not found in the sires herein (namely, the 5 non-synonymous SNPs in CAPN5 (rs41773712, alanine > valine rs137806597, threonine > proline, rs137205316, tyrosine > serine, rs41774801, serine > asparagine and rs137336128, asparagine > serine) and the 2 non-synonymous SNPs in LOXL1 (rs133149184, valine > glycine and rs135516198, tyrosine > serine). These SNPs may be present in the Davies mapping progeny as the

115 variants could have been inherited from the dams. However, these SNPs were not genotyped herein because linkage analysis using the sire information was to be conducted as well as the association analysis using the dam information.

Apart from some exons in the MYO1G gene (exons 4, 5, 11, 12 and 16), all exons of the candidate genes were sequenced. The exons in the MYO1G gene that could not be sequenced had a very high G/C content in the corresponding regions. Since the

GC percentage was above 70%, amplification was not successfully performed even under a variety of different PCR conditions. Regions of high G/C content are more likely to have a higher frequency of polymorphism (Cooper et al., 2011), so it is quite possible that there are amino acid substitutions within these exons that could not be identified.

In discovering the candidate gene variants, there were approximately 33,700 base pairs (bp) sequenced in the 6 candidate genes. The SNPs occurred with an average spacing of 1 variant per around 800 bp. This is similar to the 1 SNP per 1000 bp estimated for human populations by Taillon-Miller et al., (1998) and Marth et al.,

(1999). Miller and Kwok (2001) stated that around 17% of the SNPs in humans had no detectable variation in any of the three major populations in the USA. However, herein, all the variants were discovered by sequencing 3 F1 sires from divergent genetic backgrounds. Therefore, finding 1 SNP per 800 bp is well within expectation.

All the selected candidate genes covered different aspects related to tenderness. For example, tenderness is thought to be affected by both the proteolytic systems and collagen content. Therefore, variants in the CAPN4 and CAPN5 genes plus LOXL1 were chosen for genotyping. MYL7 and MYO1G are not as likely to affect tenderness directly. However, as muscle structure and protein turnover may affect

116 tenderness indirectly, DNA variants within these genes were also selected for association studies. Such candidate genes may help draw a more complete picture of the mechanisms underlying tenderness.

In addition to the non-synonymous SNPs and insertion/deletions discovered in this study, many synonymous SNPs were also found. Although synonymous SNPs are silent and do not change the amino acid composition of a protein, they are just as likely to be associated with a trait as non-synonymous SNPs (Chen et al., 2010).

Chen et al., (2010) evaluated the number of different types and positions of SNPs associated with disease by using the published human genome-wide association studies data. They observed that there is no difference between the percentage of the non-synonymous SNPs and the synonymous SNPs being associated with diseases.

Furthermore, the percentage of the associated-disease SNPs in the 5’UTR was the same as the non-synonymous SNPs. The percentage of the associated-disease SNPs in 3’UTR showed an approximate 1/3 lower value than SNPs in 5’UTR. The findings indicate that SNPs in most locations within the genes share a similar percentage of association with diseases with the exception of the SNPs in the 3’UTR.

However, the authors (Chen et al., 2010) suggested that SNPs in the 3’UTR are equally important and that the current bias is because there has been insufficient investigation of 3’UTR SNPs, leading to the lower percentage of disease being associated with SNPs in 3’UTR.

There has been renewed interest in 3’UTR SNPs since the discovery of the microRNA binding sites in 3’UTRs of many genes. These sites allow microRNAs to bind the mRNA and inhibit translation, thereby decreasing gene expression (Guo et al., 2010). The prediction algorithms for the microRNA binding sites in 3’UTRs

117 have been subsequently developed (John et al., 2004; Lall et al., 2006; Kertesz et al.,

2007; Hammell et al., 2008). None of the 3’UTR SNPs found herein are located in microRNA target sites. Unfortunately, the identification of microRNA sequences and their target sites is incomplete for all species. For instance, the bovine microRNA database has less than 700 entries thus far. This suggests that the current prediction equations are not sufficient and that all 3’UTR SNPs should be re-evaluated as new software is developed to ensure that potentially important SNPs are not overlooked.

Hence, all the non-coding variants discovered in this study may still be theoretically linked with the traits of interest.

118

Chapter 5: Association Studies

119

5.1 Introduction

Association studies are designed to identify the relationships between traits and variables or factors, in this case, DNA variants in the bovine genome. Association analyses for DNA variants are performed mainly using linear regression methods

(Casas et al., 2006a; Schenkel et al., 2006; Miquel et al., 2009; Curi et al., 2010).

Commonly, fixed factors such as age, sex, sire and group/herd are fitted into the analytical model in order to exclude the influence of these environmental effects

(Casas et al., 2006a; Schenkel et al., 2006; Serrano et al., 2007; Bhuiyan et al., 2009;

Curi et al., 2009). The DNA variants are then tested to determine if they can explain any of the remaining residual effects. Once a DNA variant genotype is known to be associated with the trait, it can be also added into the model to uncover the effect of other DNA variants (Casas et al., 2006a).

In order to discover the genes associated with tenderness and related traits herein, association studies were performed for the candidate gene variants using general linear regression. The association models included the effects of two major genes known to affect meat quality, myostatin (Kobolak and Gocza, 2002; Wheeler et al.,

2004; Lines et al., 2009; Allais et al., 2010) and calpain 1 (Page et al., 2004; White et al., 2005; Drinkwater et al., 2006; Morris et al., 2006; Casas et al., 2006a; Van

Eenennaam et al., 2007, Frylinck et al., 2009; Cafe et al., 2010; Johnston and Graser,

2010; Allais et al., 2011). Specific variants in these genes (namely MSTN F94L and

CAPN1-SNP316 and SNP530) were analysed (as described below). The genotypes of these SNPs were included to ensure that any effect attributed to a candidate gene variant was not related to these previously identified major genes.

120

Deletion and nonsense mutations in the bovine myostatin (MSTN) gene lead to the double muscling in cattle with improved meat tenderness in breeds such as the

Belgian Blue and Piedmontese (Bellinge et al., 2005). Another variant of myostatin, the F94L, is commonly found in the Limousin breed and moderately increases muscle mass in cattle (Sellick et al., 2007). This SNP variant is also associated with improved beef tenderness in the population herein (Esmailizadeh, et al., 2008; Lines, et al., 2009). Therefore, in association studies for tenderness using Limousin backcrosses, this SNP should be accounted for.

The calpain system is a critical factor affecting tenderness through the proteolysis of the muscle during ageing. The calpain 1 gene encodes is the large subunit of µ- calpain and polymorphisms in the CAPN1 gene may affect tenderization by altering the activity of calpain (Koohmaraie and Geesink, 2006; Neath et al., 2007b). Greater calpain activity will increase fragmentation, which would then decrease shear force and improve tenderness. CAPN1-SNP316 and CAPN1-SNP530 are the variants of µ- calpain suspected to affect tenderness by changing the degradation rate of the muscle structure (Page et al., 2002). These variants of calpain 1 have been shown to affect tenderness in the cattle used herein (Morris et al., 2006) as well as in the New

Zealand Jersey-Limousin sister herd (Page et al., 2002). Therefore, the SNPs were also examined in the association studies herein.

There are other known variants in the calpastatin (CAST), calpain 3 (CAPN3) and lysyl oxidase (LOX) genes verified as affecting beef tenderness (Chung et al., 2001;

Barendse, 2002; Barendse et al., 2008; Drinkwater et al., 2006). The variants in

CAST and CAPN3 are in low frequency within the Jersey-Limousin backcross

121 utilised herein and none of the sires were heterozygous for these SNPs. Therefore, these SNPs were not genotyped and analysed in the progeny.

However, the patented SNP in the LOX gene was genotyped in the Jersey-Limousin progeny herein subsequent to sequencing the region in the sires and discovering that

2 of 3 sires were heterozygous for this variant. Lysyl oxidase is encoded by a gene at

28 cM on cattle chromosome 7, and plays an important role in forming cross-links which stabilize the structure of collagen and elastin in the extracellular matrix. Lysyl oxidase is a copper dependent amine oxidase that oxidizes peptidyl lysine and hydroxylysine side chains into α-aminoadipic-δ-semialdehyde (AAS). AAS combines with other un-oxidized lysine side chains or AAS itself to form the collagen cross-links (Lucero and Kagan 2006). A polymorphic fragment of LOX gene has been patented as a DNA marker for compression (Barendse, 2002). Since compression is one of the traits related tenderness, the LOX patented SNP was chosen for genotyping even though no QTL for tenderness were located near the gene in the Jersey-Limousin backcross mapping progeny.

In addition to the LOX variant and the variants chosen from sequencing the tenderness candidate genes (Chapter 4) (Table 5.1), there were nine previously discovered variants from six other genes selected for the association studies with tenderness (follistatin precursor, follistatin-related protein 1 precursor, smad nuclear interacting protein 1, insulin-like growth factor I, insulin-like growth factor

1 receptor precursor and growth hormone release-inhibiting factor) (Table 5.2).

These variants had been discovered during the sequencing of candidate genes for net feed efficiency and then genotyped (Naik, 2008). The variants were included in the association studies herein based on their biological functions which may be related to

122 tenderness (as described below) and their location within significant tenderness QTL

(Table 5.1). Therefore, 20 variants in total were used in the association analyses to establish the effects of the genes on tenderness.

The follistatin precursor gene (FST) on chromosome 20 encodes a binding protein in the TGF-β superfamily. The activity of TGF-β superfamily members, such as myostatin, bone morphogenetic proteins, growth differentiation factor 11, and activin, are neutralized or modified by the binding of follistatin (Funkenstein et al., 2009;

Sidis et al., 2006). The TGF-β superfamily is responsible for muscle growth and development and lower shear forces are proposed to be affected by muscle hyperplasia. For example, myostatin is an inhibitor to muscle growth and has known effects on tenderness in cattle (Bellinge et al., 2005). Follistatin (FST) is an antagonist of myostatin and can induce muscle growth by inhibiting myostatin.

Similarly, the follistatin-related protein 1 precursor gene (follistatin-like 1, FSTL1) on chromosome 1 was investigated because its function also mediates the muscle growth (Ouchi et al., 2008).

Table 5.1 Selected candidate gene variants (from Chapter 4).

Variants Genotype Location Sequence MYL7 SNP1 C/G Intron 2 TCACACCAGC/GGACCCGGCTCAGAGCT MYO1G SNP2 C/T 5’ CTGTCCTC/TACTACTAGA MYO1G SNP3 G/A 5’ CAAAG/AGCCTGGGTATCAGCAT MYO1G SNP5 G/A Exon 7 AGAGCCACCACC A/GGGCAGTGACG LOXL1 SNP1 A/G Exon 1 CCGCAGCGCGA/GGCGGCGGCCT CAPN4 insertion/deletion AC 5’ UTR ACGCGCGAACGAGGCGGTGAG CAPN4 3 base repeat GCC Exon 2 GGCGGCGGCGGAACCGCCATGCGCATC CAPN5 SNP16 T/C Exon 7 CCTTCGAGGACTTGTGCC/TGCTACTTC CAPN5 SNP17 G/A Exon 7 TCCACAAGACG/ATGGGAGGAGGCCC CAPN5 SNP21 G/A Exon 11 GGACAAAGTCCG/ACTCGGCCGTGC

123

Table 5.2 Additional candidate gene variants from other sources (Naik, 2006).

Variant Location Sequence FST-SNP5 91th base in intron 3 ATAGCCTAATAATAATAA/GTATCAAATA FST-SNP7 503th base in intron 4 AGTTTTGAGTATA/GTGTGTGTGTGTTTCAG FSTL1-SNP1 3’UTR CTACGACT/CGCCAAATCACCTGTA FSTL1-SNP2 37th base of intron 10 AGCCCTGTCACCACCA/GGGCCCGTCCTCC SNIP-SNP3 77th base in intron 4 GGCACCAC/AGGTGCTCTCTGTAATGC 289th base before IGF-SNP1 CGCCCATCCTCC/TACGAATATTCC 5’UTR IGF-SNP2 56998th base in intron 4 TCTTCGCGAGATC/TGGAGTTATGA IGF1R-SNP1 269th base in intron 11 CTTAATAACTTAATA/CTTTCTGAA 258th base in exon 2 SST-SNP2 TGATTCCTCTTC/TTCCAAACCCCTT (3'UTR) LOX-SNP1 124th base in intron 5 AGAACTACTTT/CTAAACCAAC

The Smad nuclear interacting protein 1 (SNIP1) is a mediator of the SMAD system, which is the signal transduction pathway of the TGF-β superfamily. SNIP1 interacts with Smad 1, 2, 4 and CBP/p300 (Kim et al., 2000). CBP/p300 is the co-activator for transcription regulation of the SMAD system in the nucleus. SNIP1 is a repressor of CBP/p300 and is a candidate gene as it is in the same pathway as myostatin.

Moreover, a SNP (SNIP-SNP3, C/T) at the 77th base in intron 4 on chromosome 3 has been shown to interact with myostatin to affect muscle weight (Novianti, 2009).

This indicates that there may be epistasis between the SNIP1 and myostatin genes.

Hence, this SNIP1-SNP3 was included in the association analysis.

Insulin-like growth factor I (IGF1) has been shown to play an important role in various aspects of tissue growth and development, including muscle (Bunter et al.,

2005; Davis and Simmen, 2006). Due to the effect of IGF1 on the hypertrophy of muscle cells, muscle fibre diameter may be affected directly by IGF1 (Musaro et al.,

2001). Since increasing muscle fibre diameter may decrease tenderness (Herring et al., 1965), it was postulated that IGF1 may also affect tenderness by increasing the size of the muscle fibres (Paul and Rosenthal, 2002). Moreover, the activity of m-

124 calpain is increased by IGF1 treatment (McDonagh et al., 1999). IGF1 may be able to affect tenderness by mediating the activity of m-calpain (Leloup et al., 2007).

Therefore, the IGF1 and insulin-like growth factor 1 receptor precursor (IGF1R) genes on chromosomes 5 and 21, respectively, were selected for further investigation.

The growth hormone release-inhibiting factor (SST) gene on chromosome 1 encodes somatostatin, which is an inhibitor of growth hormone (Roh et al., 1997). The growth hormone pathway is also related to tissue growth, including muscles. Purchas et al., (2002) showed that the different growth rates in cattle result in significant differences in beef tenderness. In general, cattle with faster growth rates provide more tender meat although it is possible that this is due to reaching market specifications at a younger age (Therkildsen et al., 2002). Given the relationship between growth and tenderness, the growth hormone release-inhibiting factor gene was also used in the association study.

In the study herein, two muscles (LD and ST) were analysed in the association studies. Other previous such studies have focused only on the longissimus muscle

(Page et al., 2002; Casas et al., 2006a; Therkildsen et al., 2008; Curi et al., 2010). It has been demonstrated in Chapter 2 that shear force in the two muscles is a different trait. Thus, by analysing both muscles, it is possible to delineate muscle specific pathways that may affect tenderness and determine their relative importance.

In addition to expanding the association study to include two muscles, by using ageing rate (as developed in Chapter 2) in this study, the relationship between DNA variants and ageing was explored for the first time. The other traits analysed herein included the direct tenderness measures of adjusted shear force (Chapter 3), shear force at specific ageing days 1 and 26, and compression as well as the tenderness-

125 related traits of cooking loss, meat pH, hydroxyproline content, and the meat weights of the LD and ST muscles. If a gene is associated with any of these 4 additional traits and tenderness, it may help explain the mechanisms behind the effects on tenderness.

Lastly, multiple candidate genes were also taken into account in a mixed model simultaneously in order to acquire a better estimate for each candidate variant. All individual candidate gene effects discovered should help develop a more comprehensive marker selection system.

5.2 Methods

Two main types of the models were used to examine the individual gene effects, fixed and mixed. In the fixed models, 20 candidate variants and 12 traits in total were used in the association studies. General linear models with and without three fixed factors accounting for major gene effects (MSTN F94L, CAPN1-SNP316 and

CAPN1-SNP530 genotypes) were used in Genstat to identify relationships between the variants and the traits (models 3~6). The additive and dominance effects were estimated by using model 7 and model 10 for all traits and adjusted shear force, respectively. In the mixed model (model 11), all the variants were fitted as random factors to infer the maximum effect for each variant. All the genotyping of the candidate gene variants was performed using high resolution melt (HRM) or gel electrophoresis. The additional variants (Table 5.1) were genotyped using the

Illumina system by DPI Victoria as part of the Beef CRC net feed efficiency project and the data provided for the analyses herein.

Model 3: Model with fixed effects to adjust for cohort, breed and sire

The model was used to show that the variant effect on the traits after the fixed effects have been taken into account for all traits except adjusted shear force.

126

Yijkl = μ + αi + βj + γk + λl + Rijkl

Yijkl the response variable

μ the overall mean

th αi the effect of i cohort (six levels)

th βj the effect of j breed (Jersey or Limousin)

th γk the effect of k sire (three sires)

th λl the effect the l variant

Rijkl the residual effect

Model 4: Model with fixed effects to adjust for cohort, breed, sire and the known tenderness related variants myostatin F94L and calpain 1 (CAPN1- SNP316 and SNP530)

The model was used to show that the variant effect on the traits while the known tenderness associated polymorphisms, MSTN F94L, CAPN1-SNP316 and CAPN1-

SNP530, were taken into account. By comparing model 4 with model 2 or model 3, the variation accounted for by MSTN F94L, CAPN1-SNP361 and CAPN1-SNP530 is removed and other significant polymorphisms may be revealed that explain the remaining genetic variation.

Yijklmno = μ + αi + βj + γk + λl +δm + θn + τo+ Rijklmno

Yijklmno the response variable

μ the overall mean

th αi the effect of i cohort (six levels)

th βj the effect of j breed (Jersey or Limousin)

th γk the effect of k sire (three sires)

th λl the effect the l myostatin F94L genotype (AA, CA, CC)

th δm effect of the m calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC )

127

th θn effect of the n calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA)

th τo effect of the o variant

Rijklmno the residual effect

Model 5: Model with no fixed effects to show the original effects of variants on adjusted shear force

The model was used for adjusted shear force to show that the individual variant effects and to show the effect of the known tenderness associated polymorphisms,

CAPN1-SNP316 and CAPN1-SNP530 by comparing results with model 6.

Yi = μ + αi + Ri

Yijk the response variable

μ the overall mean

th αi the effect of i variant

Ri the residual effect

Model 6: Model with fixed effects to adjust for the known tenderness related variants calpain 1 (CAPN1-SNP316 and SNP530) for adjusted shear force

The model was used for adjusted shear force to show that the variant effect while the known tenderness associated polymorphisms, CAPN1-SNP316 and CAPN1-SNP530, were taken into account. [Note; It was not necessary to take MSTN F94L in account as it was part of the derivation of adjusted shear force as a trait.] By comparing model 5 with the model 1, the variation accounted for by CAPN1-SNP361 and

CAPN1-SNP530 is removed and other significant polymorphisms may be revealed that explain the remaining genetic variation.

Yijk = μ + αi + βj + γk+ Rijk

Yijk the response variable

μ the overall mean

128

th αi the effect of i calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC )

th βj the effect of j calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA)

th γk the effect of k variant

Rijk the residual effect

Model 7: Model with all fixed effects to estimate the additive and dominance effects

The model was used for significant variants from models 3 to proportion the genetic variance accounted for by the variant into the additive and dominance components for all traits except adjusted shear force.

Yijk= μ + αi + βj + γk + bXo + bXq + Rijk

Y ijk the response variable

(all traits except adjusted shear force value for LD and ST muscles)

μ the overall mean

th αi the effect of i cohort (six levels)

th βj the effect of j breed (Jersey or Limousin)

th γk the effect of k sire (three sires)

bXo regression coefficient of the additive effect of the variant

bXq regression coefficient of the dominance effect of the variant

Rijk the residual effect

Model 8: Model with all fixed effects and the known tenderness related variants myostatin F94L and calpain 1 (CAPN1-SNP316 and SNP530) to estimate the additive and dominance effects

The model was used for significant variants from models 4 to proportion the genetic variance accounted for by the variant into the additive and dominance components for all traits except adjusted shear force.

129

Yijklmn= μ + αi + βj + γk + λl +δm + θn + bXo + bXq + Rijklmn

Y ijklmn the response variable

(all traits except adjusted shear force value for LD and ST muscles)

μ the overall mean

th αi the effect of i cohort (six levels)

th βj the effect of j breed (Jersey or Limousin)

th γk the effect of k sire (three sires)

th λl the effect the l myostatin F94L genotype (AA, CA, CC)

th δm effect of the m calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC )

th θn effect of the n calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA)

bXo regression coefficient of the additive effect of the variant

bXq regression coefficient of the dominance effect of the variant

Rijklmn the residual effect

Model 9: Model to estimate the additive and dominance effects for adjusted shear force

The model was used for significant variants from model 5 to proportion the genetic variance accounted for by the variant into the additive and dominance components for adjusted shear force.

Yijk= μ + b1X1j + b2X2k + Rijk

Yijk the response variable (adjusted shear force value for LD and ST

muscles)

μ the overall mean

bXj the regression coefficient of the additive effect of the variant

bXk the regression coefficient of the dominance effect of the variant

Rijk the residual effect

130

Model 10: Model with calpain 1 effects to estimate the additive and dominance effects for adjusted shear force

The model was used for significant variants from model 6 to proportion the genetic variance accounted for by the variant into the additive and dominance components for adjusted shear force.

Yij= μ + αi + βj + bXk + bXl + Rij

Yij the response variable (adjusted shear force value for LD and ST

muscles)

μ the overall mean

th αi the effect of i calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC )

th βj the effect of j calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA)

bXk the regression coefficient of the additive effect of the variant

bXl the regression coefficient of the dominance effect of the variant

Rij the residual effect

Model 11: Model with all variants included simultaneously

In this model, cohort, breed and sire were fitted into the model as fixed factors.

Myostatin F94L genotype, calpain 1 genotypes (CAPN1-SNP316 and CAPN-

SNP530) and all investigated DNA variants were fitted as random factors simultaneously in a mixed model using ASREML, a statistical approach for fitting linear mixed models (Gilmour et al., 2006). [Note: for adjusted shear force, it was not necessary to include cohort, breed, sire and MSTN genotype as factors because they are already accounted for in the trait.] This model allowed a direct comparison of the amount of genetic variation accounted for by all the DNA polymorphisms and may reveal additional variants which may explain that variation. Typically, any

131 variant explaining 5% or more of the variation in the Davies gene mapping herd when all variants are analysed simultaneously are likely to have significant effects when tested further (W. Pitchford, personal communication).

5.2.1 Statistical analysis of DNA variant associations

Associations between the candidate gene DNA variants and traits were tested by general linear regression using Genstat. The aim of the association analyses was to verify the relationship between the selected variants of the candidate genes and tenderness traits. Traits included tenderness and related meat quality measurements

(Table 5.3).

Table 5.3 Traits analysed in association studies. Abbreviation Description wbld_adjusted adjusted shear force for LD muscle wbst_adjusted adjusted shear force for ST muscle Ageing rate of LD muscle or logwbld1_26 ageing rate for LD muscle Ageing rate of ST muscle or logwbst1_26 ageing rate for ST muscle ST_Compression compression of ST muscle ST_Hydroxyproline collagen content of ST muscle LD Weight muscle weight of LD muscle ST Weight muscle weight of ST muscle pHld pH value of LD muscle pHst pH value of ST muscle Clld cooking loss in LD muscle Clst cooking loss in ST muscle

5.2.2 Trait summary

The basic statistics were determined for all the traits (Table 5.4). The results demonstrated several specific muscle phenomena. Similar to the derived traits of adjusted shear force and ageing rate, the LD muscle had a larger coefficient of variation than ST in the shear force related traits, cooking loss and pH, suggesting

132 that the number of the factors that are involved in these traits in LD muscle is greater than those for the ST muscle. The ageing rate and muscle weights were also in opposite direction for the two muscles.

Table 5.4 Basic statistics for all the traits

Traits Mean Maximum Minimum Coefficient Of Variation wbld1 4.87 13.20 2.60 27.39 wbld5 4.42 10.30 2.50 23.38 wbld12 4.19 10.60 1.80 24.77 wbld26 3.96 9.90 2.10 24.94 wbst1 5.30 10.30 3.50 16.68 wbst5 5.07 7.40 2.60 16.65 wbst12 4.93 8.70 2.70 17.05 wbst26 4.67 8.40 2.70 16.95 wbld_adjusted 4.37 9.02 2.64 15.95 wbst_adjusted 4.76 7.23 3.58 8.09 ageing rate for LD 0.19 0.86 -0.29 88.91 ageing rate for ST 0.13 0.46 -0.36 104.72 st_compression 1.96 2.83 1.15 15.05 st_hydroxyproline 6.44 9.89 3.59 16.87 LDWeight 6.28 11.48 3.11 24.00 STWeight 2.49 10.00 1.14 33.63 clld 21.82 39.65 14.46 8.68 clst 26.25 31.05 19.04 6.89 pHld 5.63 6.70 5.43 2.17 pHst 5.69 6.38 5.45 1.72

The least squares means of all traits categorised by cohort, breed, sire and MSTN

F94L demonstrated the effects on the traits (Table 5.5). In particular, the effects of breed were large for the shear force related traits. The breed changed the least square means in the opposite direction between muscles. Also the effect of the MSTN F94L genotype on the ST muscle was observed when the results were compared between the two muscles.

133

Table 5.5 Least square means of all traits categorised by cohort, breed, sire and MSTN F94L genotype.

cohort breed sire MSTN

96H 96S 97H 97S 98H 98S XJ XL 361 368 398 AA AC CC wbld1 4.06 3.84 5.57 4.84 5.87 4.37 4.77 5.02 4.99 4.89 4.77 4.72 4.95 4.87 wbld5 3.77 3.56 4.79 4.72 5.06 3.94 4.16 4.74 4.49 4.41 4.37 4.29 4.39 4.54 wbld12 3.48 3.28 4.80 4.60 4.48 3.73 3.95 4.49 4.23 4.25 4.10 4.07 4.16 4.33 wbld26 3.04 2.94 4.56 4.33 4.22 3.78 3.74 4.25 3.99 4.07 3.84 3.85 3.99 4.01 wbst1 4.65 4.54 5.68 5.44 5.47 5.38 5.50 5.08 5.32 5.37 5.22 4.75 5.33 5.54 wbst5 4.38 4.20 5.24 5.20 5.33 5.35 5.23 4.89 5.08 5.08 5.06 4.50 5.08 5.36 wbst12 4.17 3.98 5.05 5.18 5.18 5.24 5.12 4.72 4.97 4.96 4.89 4.44 4.94 5.19 wbst26 3.92 3.89 4.98 4.72 4.84 5.03 4.87 4.46 4.70 4.73 4.61 4.21 4.74 4.82 st_compression 2.19 1.85 NA NA 1.98 1.94 1.97 1.94 1.93 2.02 1.90 1.79 1.98 1.98 st_hydroxyproline 6.54 6.64 NA NA 6.01 6.57 6.47 6.42 6.34 6.42 6.61 5.66 6.48 6.70 LDWeight 7.17 7.41 6.50 6.83 5.49 5.04 5.72 6.94 6.50 6.19 6.13 6.68 6.24 6.13 STWeight 2.54 2.61 2.31 2.73 2.14 2.66 2.11 2.97 2.61 2.44 2.44 3.15 2.42 2.31 clld 21.52 21.66 21.24 21.67 23.35 21.51 21.45 22.28 21.87 21.68 21.96 21.64 21.95 21.72 clst 24.88 24.18 27.64 27.52 25.57 25.79 26.08 26.49 26.27 26.28 26.24 25.43 26.38 26.48 pHld 5.59 5.57 5.74 5.57 5.70 5.58 5.63 5.62 5.64 5.62 5.62 5.62 5.63 5.62 pHst 5.69 5.71 5.73 5.73 5.64 5.63 5.68 5.69 5.71 5.68 5.68 5.68 5.68 5.70 ageing rate for LD 0.29 0.26 0.19 0.11 0.32 0.14 0.23 0.17 0.22 0.18 0.21 0.22 0.21 0.19 ageing rate for ST 0.17 0.16 0.13 0.14 0.12 0.07 0.12 0.13 0.13 0.13 0.12 0.13 0.12 0.14 wbld_adjusted 3.58 3.40 4.92 4.62 4.90 3.95 4.00 4.46 4.27 4.27 4.15 4.12 4.25 4.32 wbst_adjusted 4.16 4.03 5.12 5.01 5.09 5.13 4.95 4.56 4.80 4.80 4.68 4.33 4.87 5.08 NA: not available due to insufficient number of samples. 96~98: year 1996~1998; H:heifer; S:steer; XJ: Jersey backcross; XL: Limousin backcross; 361,368,398: sire families; AA,AC,CC: genotypes of MSTN F94L.

134

5.3 Results

There were 10 variants in the sequenced candidate genes chosen for genotyping

(Chapter 4). In addition, the patented DNA marker for compression in the LOX gene

(Barendse, 2002) was investigated as another potential major variant affecting compression. Only the patented region was sequenced for the LOX gene in the 3 sires and the grandparents as there was no QTL for tenderness near the gene. Two

SNPs were identified in the LOX intron 5 (Appendix 16) that were heterozygous in sires 361 and 368. The patented SNP (SNP1) was chosen for genotyping. The second

SNP (SNP2) could not be genotyped as it was too close to SNP1 for accurate allele assignment using any of the different genotyping technologies utilised herein. No other LOX SNPs were identified or genotyped. There were another 9 variants in six genes chosen for analysis based on the potential roles of the genes in tenderness and their positions relative to tenderness QTL (described above).

As a consequence of the low minor allele frequencies for some variants, the genotype frequency of one homozygote was occasionally also small despite at least one sire being heterozygous for each variant (Appendix 17A). The number of genotypes can be limiting when there are only a few animals with a particular genotype and the results must be interpreted accordingly. Therefore, herein, in those cases where there were less than 10 animals with a particular genotype, it is noted. In those instances where a genotype frequency was particularly low (eg CAPN4-two base insertion/deletion, CAPN4-3 base repeat, CAPN5-SNP16, MYO1G-SNP2,

MYO1G-SNP3, IGF1R-SNP1, FST-SNP7, FSTL1-SNP1 and FSTL1_SNP2), there were large standard errors for the least squares means because there were very few progeny representing that genotypic class. In these instances, the size of effects could not be accurately estimated. Nevertheless, the results are still worth noting because

135 the effect must be large to be significant even with such low numbers. As an example, the major gene variant that causes yellow fat in cattle, BCO2 W80X, was first discovered in these same Jersey-Limousin backcross progeny with only 4 animals in the homozygous recessive genotypic class (Tian et al. 2010). The variant was then confirmed in 6 other populations as causative (Tian et al. 2010). Recent work suggests that rare variants are responsible for much of the genetic variation observed in animals (Liu and Leal, 2010). Therefore, significant results should not be ignored even with low numbers of animals in some genotypic classes.

In the same vein, for association studies, P-values must be considered in light of the number of statistical analyses performed. Normally, a P-value of less than 0.05 was considered significant herein. If the P-value was less than 0.1 but greater than 0.05

(that is, 0.05 < P < 0.1), then the association was still noted. For although the certainty would be 90% rather than 95% in such cases, these variants could still be included in the panel for verification.

5.3.1 Effects of CAPN major variants on tenderness related traits

The effect of MSTN and CAPN1 on tenderness within the population of animals utilized herein has been previously described from the standpoint of explaining the specific ageing day shear force QTL on BTA 2 and BTA 29, respectively

(Esmailizadeh et al., 2008; Lines et al., 2009; Morris et al., 2006). However, the effects of MSTN and CAPN1 on tenderness related traits, other QTL and their interactions have not been fully investigated.

The effect of CAPN1 on tenderness is largely ascribed to differences in ageing rate

(Dayton et al., 1976; Koohmaraie 1994; Koohmaraie 1996). However, using a general linear regression model (model 4) with cohort, breed, sire, myostatin F94L

136 genotype and CAPN1-SNP316/530 genotypes (and no other variant genotype) as fixed factors, and ageing rate as the dependent variable (for LD or ST muscle), no relationship between ageing rate and the CAPN1-SNP316 or CAPN1-SNP530 genotypes was observed. That is, ageing rate was not affected by the CAPN1 genotypes. This finding is not consistent with the theory that CAPN1-SNP316 affects tenderness by differences in ageing rate. The most likely explanation is that the calpain effects on ageing rate were greatest at very early stages of tenderization in this particular experiment. All steaks were collected from the abattoir on day 1 after slaughter and stored in a chiller at 1℃. The temperature of the meat decreased from

37℃ to approximately 15℃ in day 1. All samples were collected from the carcass for the assessment during day 1 and then the carcasses were shifted to the freezer at minus 20℃. So by day 2, the meat samples had cooled to 0.5℃. Hence, the calpain system may have been active during the period before the technical “day 1” because the temperature of the meat would still allow calpain activity. This could especially be the case with carcasses being heavy and fat after at least 6 months of feedlot finishing. At end of day 2, the low temperature (-20℃) of the samples would have reduced the activity of the calpain system. However, by then, much of the ageing process may have already occurred although further ageing definitely did occur

(Figures 5.1 and 5.2).

Notably, using a general linear regression model (model 4) with cohort, breed, sire, myostatin F94L genotype and CAPN1-SNP316 and SNP530 genotypes as fixed factors and shear force at the four ageing days (1, 5, 12 and 26) as the dependent variables, the CAPN1-SNP316 genotype was significant associated with shear force

(Figures 5.1 and 5.2) at all ageing days. Consequently, the effect of the CAPN1-

SNP316 genotype was observed at each specific ageing period even though the

137 specific effect of CAPN1-SNP316 genotypes on ageing rate was not significant for the overall period from day 1 to day 26.

3 CAPN1_SNP316/CC

wbld(KgF) 2 CAPN1_SNP316/GC CAPN1_SNP316/GG 1

0 0 10 20 30 Days post-slaughter

Figure 5.1 Effects of CAPN1-SNP316 on LD shear force measured at 4 time points (day 1, 5, 12 and 26 post-slaughter). Bars represent standard errors.

The effect of CAPN1-SNP316 on tenderness has been shown in the LD muscle previously (Page et al., 2002). However, dissimilar effects of CAPN1-SNP316 on 2 different muscles (LD and ST) were observed herein (Figures 5.1 and 5.2). The GC and CC genotypes of CAPN1-SNP316 had a different effect from the GG genotypes of CAPN1-SNP316 on the LD muscle, in general. In contrast, for the ST muscle, the

GC and GG genotypes of CAPN1-SNP316 had a different effect from the CC genotype. Thus, it can be concluded that the effects of CAPN1-SNP316 on tenderness may be muscle specific (Figures 5.1 and 5.2).

138

3 CAPN1_SNP316/CC

wbst (KgF) 2 CAPN1_SNP316/GC CAPN1_SNP316/GG 1

0 0 10 20 30 Days post-slaughter

Figure 5.2 Effects of CAPN1-SNP316 on ST shear force measured at 4 time points (day 1, 5, 12 and 26 post-slaughter). Bars represent standard errors. When investigating relationships between CAPN1 SNPs (SNP316 and SNP530) and the tenderness related traits, it was observed that the CAPN1-SNP316 and CAPN1-

SNP530 were associated with 7 traits and 4 traits, respectively (Table. 5.6; models 3 and 5). CAPN1-SNP316 showed significant results across the whole period of tenderization for both types of muscles, LD and ST.

Table 5.6 Relationships between CAPN1 SNPs and tenderness related traits.

Traits CAPN1-SNP316 CAPN1-SNP530 wbld_adjusted <.001*** 0.029* wbst_adjusted <.001*** 0.392 ageing rate for LD 0.239 0.057† ageing rate for ST 0.222 0.155 ST compression 0.532 0.016* ST hydroxyproline 0.967 0.32 LD weight 0.972 0.124 ST weight 0.897 0.647 LD pH 0.103 0.129 ST pH 0.047* 0.102 LD cooking loss 0.688 0.021* ST cooking loss 0.191 0.585 †(P<0.1); * (P<0.05); ** (P<0.005); *** (P<0.001)

139

There were no relationships between other traits except the pH of the ST muscle and the CAPN1-SNP316. CAPN1-SNP316 significantly explained 3.7%, 2.1% 2.5%,

1.5%, 1.5% of the variation in wbld 1, wbld 26, wbst1, wbst26 and pHst, respectively. CAPN1-SNP316 also explained 4.8% of the variation in adjusted shear force for both muscles. On the other hand, the CAPN1-SNP530 only showed an effect on shear force at day 1 and adjusted shear force for the LD muscle and explained 2.1% and 2.7% of the variation, respectively. CAPN1-SNP530 did not affect the shear force of the ST muscle. However, CAPN1-SNP530 did explain 3.1% of the variation in the compression of the ST muscle and 1.9% of the variation in LD cooking loss. Although the variants in the CAPN1 genes were not associated with all tenderness traits in both muscles, the CAPN1 SNP genotypes were included in the analytical models (models 4, 6, 8, 10 and 11) for all traits for consistency.

5.3.2 Effects of candidate gene variants on tenderness related traits

The association of the candidate gene DNA variants with the tenderness traits was investigated with and without the known major MSTN F94L, CAPN1 SNP316 and

CAPN1 SNP530 (models 3~6). Comparison of the results allows one to deduce whether or not the association of a given variant is independent of MSTN and/or

CAPN1 and thus, provide insight into the most likely biological mechanism underlying the associations.

5.3.2.1 Effects of candidate gene variants on cooking loss

There were 2 variants in candidate genes associated with cooking loss in the LD muscle (Table 5.7), CAPN5-SNP17 and SST-SNP2. Cooking loss for ST muscle was not observed to be associated with any of the candidate gene variants.

140

Table 5.7 Tests of significance for candidate gene variants on cooking loss in ST and LD muscles.

Name of variant LD cooking loss ST cooking loss MSTN-F94L 0.374 <.001*** CAPN1-SNP530 0.021* 0.585 CAPN5-SNP17 0.035* 0.089† MYO1G-SNP2 0.064† 0.897 SNIP1-SNP3 0.626 0.099† SST-SNP2 0.048* 0.65 †(P<0.1); * (P<0.05); ** (P<0.005); *** (P<0.001)

CAPN5-SNP17 and SST-SNP2 explained 1.5% and 0.9% of the variation for the cooking loss in the LD muscle, respectively. Cattle with the GG genotype for the

CAPN5-SNP17 had the greatest cooking loss (Figure 5.3). The cooking loss of meat from the homozygous GG cattle was 3% greater than that of homozygous AA cattle.

The additive effect was significant (P=0.01). The estimated allelic substitution effect for the G allele was 0.322 ± 0.127 %.

Cattle with the GG genotype for the SST2 had a greater cooking loss in the LD muscle (Figure 5. 4). The cooking loss of meat from the homozygous CC cattle was

2.7% less than that of heterozygous CT cattle. The additive effect was just significant

(P=0.048). The estimated allelic substitution effect for the C allele was -0.597 ±

0.301 %.

141

22.4 b 22.2 b 22

21.8 a

21.6

21.4 clld (%) clld 21.2 21 20.8 20.6 AA GA GG CAPN5-SNP17 genotypes

Figure 5.3 Effect of CAPN5-SNP7 genotypes on LD cooking loss. Bars represent standard errors. Letters indicate significant differences between genotypes.

22.8 a 22.6 22.4

22.2

) 22 (% b

21.8 clld clld 21.6 21.4 21.2 21 TC CC SST-SNP2 genotypes

Figure 5.4 Effect of SST-SNP7 genotypes on LD cooking loss. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.3.2.2 Effects of candidate gene variants on pH

There were associations between some of the variants and the pH of both the LD and

ST muscles (Table 5.8). CAPN5-SNP16 affected the pH of the LD muscle (P=0.014).

For CAPN5-SNP16, the meat from heterozygous CT cattle of CAPN5-SNP16 was

0.65 % higher in pH units than that of homozygous CC cattle (Figure 5.5). There was

142 a dominance effect (0.064±0.031) for CAPN5-SNP16. In addition, the MYO1G-

SNP2 was significantly associated with both the LD and ST pH (P<0.001 and

P=0.014). The meat of homozygous MYO1G-SNP2 TT cattle was 2.4% pH units higher than that of homozygous CC cattle and heterozygous CT cattle in the LD muscle (Figure 5.6) and 1.7% in the ST muscle (Figure 5.7). There were additive

(0.0696±0.0187) and dominance effects (-0.080±0.021) for the MYO1G-SNP2 in the

LD muscle and additive (0.051±0.018) and dominance effects (-0.051±0.020) for the

MYO1G-SNP2 in the ST muscle.

Table 5.8 Tests of significance of candidate gene variants on pH. Name of variant LD pH ST pH 0.014* CAPN5-SNP16 NS n=3(TT) <0.001*** 0.014 * MYO1G-SNP2 n=8(TT) n=8(TT) †(P<0.1); * (P<0.05); ** (P<0.005); *** (P<0.001); NS: not significant

Although CAPN5-SNP16 and MYO1G-SNP2 were associated with pH, these two

SNPs did not show any other effects on tenderness. The reason that these two genes did not contribute to the shear force is likely to be because the variation accounted for by these two genes on pH was too small to directly affect tenderness (1.6% for

CAPN5-SNP16 on the LD muscle and 2.9% and 1.9% MYO1G-SNP2 on the LD and

ST muscles, respectively).

143

5.7 a

5.65 b a,b 5.6

5.55

5.5

pH value pH LD formuscle 5.45

5.4 CC CT TT CAPN5-SNP16 genotypes

Figure 5.5 Effect of CAPN5-SNP16 genotypes on LD pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.85

5.8 a 5.75 5.7 5.65 b b 5.6

pH vaule pH LD formuscle 5.55 5.5 CC TC TT MYO1G-SNP2 genotypes

Figure 5.6 Effect of MYO1G-SNP2 genotypes on LD pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

144

5.85 b 5.8

5.75 a a 5.7

5.65

pH vaule for LD muscleLD for vaule pH 5.6 CC TC TT MYO1G-SNP2 genotypes

Figure 5.7 Effect of MYO1G-SNP2 genotypes on ST pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.3.2.3 Effects of candidate gene variants on adjusted shear force

A general linear regression model (model 5) was used to investigate the relationships between the 20 variants in the candidate genes and adjusted shear forces for LD and

ST muscles. The follistatin-related protein 1 gene SNP, FSTL1-SNP2, was the only variant associated with adjusted shear force for the ST muscle (Table 5.9). There were no variants associated with adjusted shear force for the LD muscle.

Table 5.9 Candidate gene variants associated with adjusted shear force. Variants LD adjusted shear force ST adjusted shear force FSTL1_SNP2 NS 0.048* †(P<0.1); * (P<0.05); ** (P<0.005); *** (P<0.001); NS: not significant, n=5(GG)

The FSTL1-SNP2 explained 1.7% of the variation in adjusted shear force for the ST muscle. Cattle with the GG genotype for the FSTL1-SNP2 had higher shear force values for the ST muscle (Figure 5.8). The meat of homozygous GG cattle was 7.8% tougher than that of homozygous AA cattle. The additive effect was significant

(P=0.02). The estimated allelic substitution effect for the G allele was 0.1843±0.0971 kgF. However, the size of the effect is likely to be over-estimated as there were low

145 numbers of animals with the FSTL1-SNP2 GG genotype (5).

5.4 5.3 b 5.2 5.1 5 4.9 a 4.8 a 4.7

wbst_adjusted (F/Kg) 4.6 4.5 4.4 AA AG GG FSTL1-SNP2 genotypes

Figure 5.8 Effect of FSTL1-SNP2 genotypes on ST adjusted shear force. Bars represent standard errors. Letters indicate significance differences between genotypes.

5.3.2.4 Effect of candidate gene variants on ageing rate

The same model (model 3) was used for the general linear regression analysis of ageing rate. There were 3 genes significantly associated with ageing rate of the LD muscle (Table 5.10), SST-SNP2 MYO1G-SNP2 and FSTL1-SNP2 (Figures 5.9-5.11).

Table 5.10 Tests of significance of candidate gene variants on LD ageing rate.

Significance Name of Variant (F-probability) MYO1G-SNP2 0.042*, n=8(TT) FSTL1-SNP2 0.048*, n=5(GG) SST-SNP2 0.034* LOX-SNP1 0.096† FSTL1-SNP1 0.064†, n=4(TT) †(P<0.1) * (P<0.05); ** (P<0.005); *** (P<0.001)

The cattle with the CC genotype of SST-SNP2 had 41% greater ageing rate than cattle with the CT genotype. The SST-SNP2 only had two genotypes in the 146 experimental population (Figure 5.9). However, the estimated allelic substitution effect for the C allele of the SST-SNP2 genotype was 0.0627± 0.0295 (ln kgF).

0.25 a

0.2 b

0.15

0.1

0.05

Ageing Ageing ratein kgF log formuslce LD CC TC SST-SNP2 genotypes

Figure 5.9 Effect of SST-SNP2 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

The cattle with TC genotype of MYO1G-SNP2 had a 24% greater ageing rate than animals with the CC genotype (Figure 5.10). The large standard errors are due to the low number of animals with the TT genotype (8) of MYO1G-SNP2 in the sample population. Furthermore, the cattle with the AG genotype of FSTL1-SNP2 had a

28% greater ageing rate than cattle with the AA genotype (Figure 5.11). Again, the large standard errors are due to the low number of animals with the GG genotype (5) of FSTL1-SNP2 in the sample population. When the additive and dominance effects were accounted for at the same time in the model, there were no significant additive or dominance effects for MYO1G-SNP2 or FSTL1-SNP2.

147

0.35 a,b 0.3 a 0.25 0.2 b 0.15 0.1 0.05 0

Ageing rate in log kgF for LD muslce CC TC TT MYO1G-SNP2 genotypes

Figure 5.10 Effect of MYO1G-SNP2 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

0.35

0.3 a,b b 0.25 a 0.2

0.15

0.1

0.05

Ageing Ageing ratein ln for kgF LD muscle 0 AA AG GG FSTL1-SNP2 genotypes

Figure 5.11 Effect of FSTL1-SNP2 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

Although the LOX-SNP1 had no significant effect on ageing rate for the LD muscle

(P=0.096), the cattle with the CC genotype had significantly higher ageing rates

(36%) for the LD muscle than the cattle with the TT genotype (Figure 5.12). The estimated allelic substitution effect for C allele of the LOX-SNP1 genotype was

0.0322± 0.0154 (ln kgF).

148

0.3 a 0.25 a,b 0.2 b

0.15

0.1

0.05

0 Ageing Ageing ratein ln for kgF LD muslce CC CT TT LOX-SNP1 genotypes

Figure 5.12 Effect of LOX-SNP1 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

Moreover, the effect of FSTL1-SNP1 on ageing rate for the LD muscle was close to significance (P=0.064). The cattle with the CT genotype had 9.5% greater ageing rates than the cattle with the CC genotype of FSTL1-SNP1 (Figure 5.13). The number of the animals with the TT genotype (4) of FSTL1-SNP1 was low in the sample population. When the additive and dominance effects were accounted for at the same time in the model, there were no significant additive or dominance effects for the FSTL1-SNP1.

0.35 0.3 a,b 0.25 b 0.2 a 0.15 0.1 0.05 0

Ageing Ageing ratein ln for kgF LD muscle CC TC TT FSTL1-SNP1 genotypes

Figure 5.13 Effect of FSTL1-SNP1 genotypes on LD ageing rate. Bars represent

149 standard errors. Letters indicate significant differences between genotypes.

There were no variants in the candidate genes significantly associated with ageing rate in the ST muscle. However, two SNPs of follistatin precursor-like 1 gene

(FSTL1-SNP1 and -SNP2) were close to significance (P<0.1). The animals with CC genotype of FSTL1-SNP1 had a significantly lower ageing rate (32%) than animals with the TC genotype (P=0.035) (Figure 5.14). The meat from the animals with AA genotype of FSTL1-SNP2 showed a significantly lower ageing rate (34%) than the

AG genotype (P=0.034) (Figure 5.15). The homozygous TT and GG genotypes for the SNP1 and SNP2 of the FSTL1 gene had large standard errors due to the small number of these two genotypes in the population (4 for TT of SNP1 and 5 for GG of

SNP2). So the existence of any additive or dominance effects was unlikely to be discovered and could not be accurately estimated.

0.3

0.25 a,b

0.2 b 0.15 a 0.1

0.05

Ageing Ageing ratein ln for kgF ST muscle 0 CC TC TT FSTL1-SNP1 genotypes

Figure 5.14 Effect of FSTL1-SNP1 genotypes on ST ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

150

0.3

0.25 a,b

0.2 b 0.15 a 0.1

0.05

0 Ageing Ageing ratein ln for kgF ST muscle AA AG GG FSTL1-SNP2 genotypes

Figure 5.15 Effect of FSTL1-SNP2 genotypes on ST ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.3.2.5 Effect of candidate gene variants on compression

The analysis of the compression data was based on the general linear regression model as described above (model 3). The results showed that none of the variants were significantly associated with compression for the ST muscle, although SNIP1-

SNP3 was nearly significant (P=0.063). For the additive effect, the estimated allelic substitution effect for the T allele was -0.097 ± 0.425 kg. Interestingly, compression is believed to be related to collagen content or the stability of the collagen matrix.

However, LOX and LOXL1, two genes encoding proteins related to maintaining the stability of collagen, did not show any effect on compression. This implies that the cross-links produced by LOX and LOXL1 may not be important for compression in the ST muscle or that there were no genetic variants of these genes in the cattle population herein which specifically affected compression.

5.3.2.6 Effect of candidate gene variants on hydroxyproline in ST muscle

Model 3 was used for the general linear regression analysis of hydroxyproline content of the ST muscle as a measure of collagen concentration. Only the IGF1-

151

SNP1 was significantly associated with hydroxyproline content for the ST muscle.

The TT genotype of IGF1-SNP1 was significantly associated with a higher amount of hydroxyproline than the CC genotype. For the additive effect of IGF1-SNP1, the estimated allelic substitution effect for C allele was -0.26 ± 0.127 mg/g. The overall effect of another SNP, MYO1G-SNP3, was close to significance (P=0.08, n=3 AA).

The additive effect of the MYO1G-SNP3 was not quite significant (P=0.059). The dominance allele substitution effect was -1.047±0.552 mg/g.

5.3.2.7 Effect of candidate gene variants on muscle weights

The candidate genes variants were also analysed for any potential associations with muscle weights in order to determine if increased muscle mass may explain any effects on tenderness. Using model 3, there were two variants (CAPN5-SNP21 and

IGF1-SNP1) highly significantly associated with ST muscle weight but less associated with LD muscle weight (Table 5.11). The CAPN5-SNP21 was highly significant for ST muscle weight (P=0.004). The cattle with the AA genotype had

10% heavier ST muscles than the cattle with the GG genotype (Figure 5.16). For the additive effect, the estimated allelic substitution effect for G allele of CAPN5-SNP21 was -0.1229 ± 0.0435 Kg.

Table 5.11 Candidate gene variants associated with muscle weight.

Variants ST muscle weight LD muscle weight CAPN5-SNP21 0.004** 0.065† IGF1-SNP1 0.03* 0.098† †(P<0.1); * (P<0.05); ** (P<0.005); *** (P<0.001)

152

2.7 a

2.6

2.5 b 2.4 b

2.3

2.2 ST muslce ST weight (Kg) 2.1

2 AA GA GG CAPN5-SNP21 genotypes

Figure 5.16 Effect of CAPN5-SNP21 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

The IGF1-SNP1 also significantly affected ST muscle weight (P=0.03). The cattle with CT genotype produced 8% more ST muscle than the cattle with CC genotype

(Figure 5.17). There was only a significant dominance effect. For the dominance effect, the allelic substitution was 0.1186 ± 0.0512 kg.

2.6 c 2.55 a,c 2.5 2.45 2.4 a,b 2.35 2.3 2.25

ST muslce weight muslce weight ST (kg) 2.2 2.15 2.1 CC CT TT IGF1-SNP1 genotypes

Figure 5.17 Effect of IGF1-SNP1 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

Although there were no variants which were highly significant for LD muscle weight, the two variants affecting ST muscle weight, CAPN5-SNP21 and IGF1-SNP1, also

153 had relatively low P values for LD muscle weight (P=0.065 and P=0.098, respectively). The cattle with the AA genotype of CAPN5-SNP21 had significantly greater LD muscle weights than cattle with the GG genotype (Figure 5.18). The cattle with the AA genotype of CAPN5-SNP21 had 6.7% more LD muscle than the cattle with the GG genotype. The additive effect CAPN5-SNP21 was 0.204 ± 0.101 kg for the A allele substitution. The cattle with the CC genotype of the IGF1-SNP1 had significantly lower LD muscle weight (6%) than cattle with the CT genotype

(Figure 5.19). The estimated allelic substitution effect for C allele of IGF1-SNP1 genotype was -0.187± 0.106 kg.

6.8

6.6 a

6.4 a,b b 6.2

5.8 LD LD muslceweight (kg) 5.6

5.4 AA GA GG CAPN5-SNP21 genotypes

Figure 5.18 Effect of CAPN5-SNP21 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

154

6.6 a,b b

6.4

6.2 a

5.8 LD LD weight muslce (kg) 5.6

5.4 CC CT TT IGF1-SNP1 genotypes

Figure 5.19 Effect of IGF1-SNP1 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.3.3 Effects of candidate gene variants on tenderness related traits dependent of myostatin and calpain 1 gene variants

In order to determine whether the effects of the candidate gene variants observed above were independent or related to the effects of myostatin and calpain 1, additional analytical models were used (models 4 and 6). These models took the SNP genotypes of these major genes in account.

5.3.3.1 Effect of candidate gene variants on adjusted shear force with CAPN1 genotypes

A general linear regression model (model 6) was used for investigating the relationships between 20 variants of candidate genes and adjusted shear force of the

LD and ST muscles. This model was used to identify the effect of the variants only when the effects of CAPN1 on tenderness were fitted into the model as myostatin was already accounted for in the adjusted shear force trait.

There were 3 variants affecting adjusted shear force using this model (Table 5.12).

SNIP1-SNP3 was the only variant affecting adjusted shear force in the LD muscle.

155

The cattle with the TT genotype of SNIP1-SNP3 had 11% tougher meat than the cattle with the other genotypes (Figure 5.20). The SNIP1-SNP3 explained 2.2% of the variation in the adjusted shear force for LD. The C allele was completely dominant. For the additive effect, the estimated allelic substitution effect for T allele was 0.237 ± 0.084 kgF (P=0.005). The dominance effect was -0.207 ± 0.099 kgF

(P=0.039).

Table 5.12 Tests of significance of variants on adjusted shear force in LD and ST muscles with the effects of CAPN1 SNP316 and SNP530 in the model.

Name of Variant Muscle Significance SNIP1-SNP3 LD 0.019* MYO1G-SNP2 ST 0.032*, n=8(TT) FST-SNP7 ST 0.020*, n=3(AA) †( P<0.10); * (P<0.05); ** (P<0.005); *** (P<0.001)

4.8 b 4.6

4.4 a a 4.2

wbld_adjusted(F/Kg) 4

3.8 CC CT TT SNIP1-SNP3 genotypes

Figure 5.20 Effect of SNIP1-SNP3 genotypes on LD adjusted shear force. Bars represent standard errors. Letters indicate significant differences between genotypes.

There were 2 variants affecting adjusted shear force of the ST muscle, MYO1G-

SNP2 and FST-SNP7. The cattle with the CC genotype of MYO1G-SNP2 had 2.6% tougher meat than the cattle with the TC genotype (Figure 5.21). However, there were only 8 animals with the TT genotype in the population. This led to a large

156 standard error for the TT genotype data. The MYO1G-SNP2 accounted for 2.0% of the variation in the adjusted shear force of the ST muscle. There were no additive and dominance effects found when both additive and dominance effects were fitted into the model simultaneously.

4.9

a 4.8 a,b 4.7 b

4.6

4.5

wbst_adjusted(F/Kg) 4.4

4.3 CC TC TT MYO1G-SNP2 genotypes

Figure 5.21 Effect of MYO1G-SNP2 genotypes on ST adjusted shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. The cattle with the AG genotype for the FST-SNP7 had 9.8% tougher meat than the cattle with the GG genotype (Figure 5.22). The AA genotype was only present in three cattle, however, leading to a very large standard error. The FST-SNP7 accounted for 2.2% of the variation in the adjusted shear force of the ST muscle. The dominance effect was 0.285 ± 0.122 kgF (P=0.02).

5 b

4.8 a a,b 4.6

4.4

4.2

wbst_adjusted(F/Kg) 4

3.8 AA AG GG FST-SNP7 genotypes

Figure 5.22 Effect of FST-SNP7 genotypes on ST adjusted shear force. Bars

157 represent standard errors. Letters indicate significant differences between genotypes.

5.3.3.2 Effect of candidate gene variants on ageing rate with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes

For the ageing rate on the LD muscle, only MYO1G-SNP2 had a significant relationship with ageing rate when the MSTN and CAPN genotypes were included in the model (model 4) (Table 5.13). MYO1G-SNP2 explained 2.1% of the variation in the ageing rate. The cattle with the CC genotype have a 40% lower ageing rate than the cattle with the CT genotype (Figure 5.23). However, a large standard error for TT was observed due to the low number (8) of animals with the TT genotype in the population. Thus, no significant difference between the ageing rate of the CC and the

TT genotypes was obvious. However, a dominance effect was observed.

Table 5.13 Tests of significance of the candidate variants of on LD ageing rate with the genotypes of MSTN F94L, CAPN1 SNP316 and SNP530 in the model.

Name of Variant Significance MYO1G-SNP2 0.019*, n=8(TT)

SNIP1-SNP3 0.078†

SST-SNP2 0.084† CAPN5-SNP16 0.140, n=3(TT) FSTL1-SNP2 0.126, n=5(GG) †( P<0.10); * (P<0.05); ** (P<0.005); *** (P<0.001)

158

0.4

a,b 0.35 0.3 b 0.25 a 0.2 0.15 0.1

0.05 Ageing Ageing rate LD for muscle (log) 0 CC TC TT MYO1G-SNP2 genotypes

Figure 5.23 Effect of MYO1G-SNP2 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

SNIP1-SNP3 and SST-SNP2 did not significantly affect ageing rate although their P- values were close to significance (P=0.078 and P=0.08, respectively). The cattle with the CC genotype of SNIP1-SNP3 significantly aged faster than the cattle with the CT genotype (Figure 5.24). The SNIP1-SNP3 and SST-SNP2 explained 1.3% and 0.8% of the total variation in ageing rate of the LD muscle, respectively (Figure 5.25).

0.3 a 0.25 b a,b 0.2

0.15

0.1

0.05 Ageing Ageing rate LD for muscle (log) 0 CC CT TT SNIP1-SNP3 genoytpes

Figure 5.24 Effect of SNIP1-SNP3 genotypes on LD ageing rate. Bars represent

159 standard errors. Letters indicate significant differences between genotypes.

0.25

0.2 a

0.15

0.1

0.05 Ageing rateLD Ageing muscle for(log) 0 TC CC SST-SNP2 genotypes

Figure 5.25 Effect of SST-SNP2 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

CAPN5-SNP16 and FSTL1-SNP2 did not show any significant effects on ageing rate.

However, there were significant differences between the averages of the genotypes observed. The cattle with TT genotype of the CAPN5-SNP16 had higher ageing rates than the cattle with the other genotypes (Figure 5.26). The cattle with the AA genotype of FSTL1-SNP2 had lower ageing rates than the cattle with the AG genotypes (Figure 5.27).

160

0.6

b 0.5

0.4

0.3 a a 0.2

0.1 Ageing Ageing rate LD for muscle (log) 0 CC CT TT CAPN5-SNP16 genotypes

Figure 5.26 Effect of CAPN5-SNP16 genotypes on LD ageing rate for LD muscle. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: different Y-axis scale due to size of effect.

0.3 a,b b 0.25 a 0.2

0.15

0.1

0.05 Ageing Ageing rate LD for muscle (log) 0 AA AG GG FSTL1-SNP2 genotypes

Figure 5.27 Effect of FSTL1-SNP2 genotypes on LD ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

For the ageing rate on ST muscle, IGF1R-SNP1 showed an effect close to significance (P=0.056). The cattle with the AA genotype of the IGFR-SNP1 had a significantly lower ageing rate than the cattle with other genotypes (Figure 5.28).

However, there were only 2 cattle with the AA genotype in the population. Hence, to conclude that this SNP significantly affects ageing rate of the ST muscle is

161 questionable.

0.25

0.2 b

0.15 b

0.1 a 0.05

0 AA AC CC -0.05

Ageing Ageing rate ST for muscle (log) -0.1

-0.15 IGF1R-SNP1 genotypes

Figure 5.28 Effect of IGF1R-SNP1 genotypes on ST ageing rate. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.3.3.3 Effect of candidate gene variants on compression with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes

None of the variants in the candidate genes were associated with compression when

MSTN and CAPN1 were fitted into the model (model 4). In the model, MSTN explained 4% of the variation in compression for ST muscle and CAPN1-SNP530 explained 1.5% of the variation.

A significant effect of MSTN F94L on shear force and compression had been shown previously (Lines, 2006). With the cohort, breed and sire as fixed factors, the mean square for the regression of ST day 26 shear force (wbst26) on MSTN F94L was 4.48 kgF2. In another model, compression was fitted as a covariate as well to adjust for the effect of the compression on shear force, and the mean square for the regression of ST day 26 shear force on MSTN F94L decreased to 2.60 kgF2. The residual mean square did not change when the covariate of compression was fitted. Hence, the

162 change in the mean square for MSTN F94L can be accredited to the effect of compression. That is, approximately half of the effect of MSTN on shear force is through an effect on compression.

5.3.3.4 Effect of candidate gene variants on hydroxyproline in ST muscle with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes

The MSTN and CAPN1 genotypes were fitted into the model for hydroxyproline content (model 4). The IGF1-SNP1 still significantly affected the content of hydroxyproline in the ST muscle (P=0.015). IGF1-SNP1 accounted for 3.2% of the variation in hydroxyproline concentration. For the additive effect of IGF1-SNP1, the estimated allelic substitution effect for the C allele was -0.238 ± 0.124 mg/g

(P=0.055). A dominance effect was observed (P=0.05) and the size of the effect was

-0.282 ± 0.143 mg/g.

The CAPN4-3 base repeat was a new variant to be significantly associated with hydroxyproline content (P=0.003, n=6(GCC)8/(GCC)8) when the MSTN and CAPN1 genotypes were included in the model. The CAPN4 insertion/deletion explained

2.8% of the variation in hydroxyproline content. The cattle homozygous for the deletions had a lower content of hydroxyproline in the ST muscle than the cattle without any deletions. An additive effect of the CAPN4 insertion/deletion was not found when the dominance effect was fitted into the model simultaneously.

5.3.3.5 Effect of candidate gene variants on muscle weights with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes

The candidate gene variants were again analysed for effects on muscle weights in case any associations with tenderness were due to changes in muscle mass. The

IGF1-SNP1 was the only variant significantly associated with LD muscle weight

163

(P=0.014) when the MSTN and CAPN genotypes were included in the model (model

4). However, the effects of CAPN5-SNP21 and SST-SNP2 on LD muscle weight were close to significance (Table 5.14).

Table 5.14 Tests of significance of variants on LD muscle weight with the effects of MSTN F94L, CAPN1 SNP316 and SNP530 in the model.

Name of Variant Significance IGF1-SNP1 0.014*

CAPN5-SNP21 0.075†

SST-SNP2 0.078† †( P<0.10); * (P<0.05); ** (P<0.005); *** (P<0.001)

The variation in the LD muscle weight accounted for by the IGF1-SNP1 was 1.3%.

The cattle with the CC genotype had 8% lower LD muscle weight than the cattle with the other genotypes (Figure 5.29). For the additive effect of the IGF1-SNP1

(P=0.041), the estimated allelic substitution effect for C allele was -0.236 ± 0.115 kg.

The dominance effect was 0.282 ± 0.129 kg.

6.6 b b 6.4 6.2 6 a 5.8 5.6

LD LD weight muslce (Kg) 5.4 5.2 CC CT TT IGF1-SNP1 genotypes

Figure 5.29 Effect of IGF1-SNP1 genotypes on LD muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

CAPN5-SNP21 and SST-SNP2 were not highly significant perhaps due to the relatively small numbers of specific genotypes of the CAPN5-SNP21 and SST-SNP2 in the population (Figure 5.30 and 5.31). The number of animals with the GG

164 genotype of CAPN5-SNP21 and the TC genotype of SST-SNP2 was 39. However, the cattle with the GG genotype of CAPN5-SNP21 had significantly lower LD muscle weights than the cattle with the AA genotype. For the additive effect of the

CAPN5-SNP21, the estimated allelic substitution effect for G allele was -0.223 ±

0.105 Kg (P=0.035). The CAPN5-SNP21 only explained 0.8% and the SST-SNP2 only explained 0.5% of the variation in the LD muscle weight though.

6.8

6.6 a 6.4 a,b 6.2 b 6 5.8

LD LD weight muslce (Kg) 5.6 5.4 AA GA GG CAPN5-SNP21 genotypes

Figure 5.30 Effect of CAPN5-SNP21 genotypes on LD muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

6.9 b 6.8

6.7 6.6 6.5 6.4 a 6.3 6.2 6.1

LD LD weight muslce (Kg) 6 5.9 5.8 CC TC SST-SNP2 genotypes

Figure 5.31 Effect of SST-SNP2 genotypes on LD muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes. For ST weight, CAPN5-SNP21 and SNIP1-SNP3 were significantly associated with

165

ST muscle weight with the MSTN and CAPN1 genotypes in the model (Table 5.15).

The cattle with the AA genotype of CAPN5-SNP21 had 9% heavier ST muscles than the cattle with other genotypes (Figure 5.32). An additive effect was observed. The estimated allelic substitution effect for G allele was -0.1098 ± 0.0405 Kg (P=0.007).

The CAPN5-SNP21 explained 0.9 % of the variation in the ST muscle weight.

Table 5.15 Tests of significance of candidate gene variants on ST muscle weight with the effects of MSTN F94L, CAPN1 SNP316 and SNP530 genotypes in the model.

Name of Variant Significance CAPN5-SNP21 0.008* SNIP1-SNP3 0.016* FST-SNP7 <0.001***, n=3(AA) †( P<0.10); * (P<0.05); ** (P<0.005); *** (P<0.001)

2.7 a

2.6

2.5 b b 2.4

2.3

ST muslce weight muslce weight ST (Kg) 2.2

2.1 AA GA GG CAPN5-SNP21 genotypes

Figure 5.32 Effect of CAPN5-SNP21 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

The cattle with the TT genotype of SNIP-SNP3 had significantly higher ST muscle weights than the cattle with the other genotype. The cattle with the TT genotype of

SNIP-SNP3 had 18% heavier ST muscle weight than the 20 cattle with the CC genotype (Figure 5.33), The T allele was recessive. For the additive effect, the

166 estimated allelic substitution effect for T allele was 0.2249 ± 0.078 Kg (P=0.004). A dominance effect was also observed (-0.1772 ± 0.087 Kg), and the SNIP-SNP3 explained 1.1% of the variation of the ST muscle weight.

3.5 b

3 a a 2.5

1.5

1 ST muslce ST weight (Kg) 0.5

0 CC CT TT SNIP1-SNP3 genotypes

Figure 5.33 Effect of SNIP-SNP3 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes.

The cattle with the AA genotype of FST-SNP7 had significantly higher ST muscle weights than the cattle with the other genotype of the FST-SNP7. The 3 cattle with the AA genotype of FST-SNP7 had 97% heavier ST muscle weight than the 293 cattle with the GG genotype and 68 cattle with AG genotype (Figure 5.34), The A allele was recessive. For the additive effect, the estimated allelic substitution effect for T allele was 1.215 ± 0.148 Kg (P<0.001). A dominance effect was also observed

(-1.244 ± 0.159 Kg), and the FST-SNP7 explained 7.7% of the variation of the ST muscle weight.

167

6 a 5

3 b b 2

ST muslce ST weight (Kg) 1

0 AA AG GG FST-SNP7 genotype

Figure 5.34 Effect of FST-SNP7 genotypes on ST muscle weight. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: different Y-axis scale due to size of effect.

5.3.3.6 Effect of candidate gene variants on cooking loss and pH with MSTN F94L, CAPN1 SNP316 and SNP530 genotypes

A general linear regression model (model 4) was used with the 3 major known variants (MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530) fitted in the model to examine the effects of the candidate genes on cooking loss and pH.

The CAPN5-SNP17 had on an effect on the pH of both the ST and LD muscles as well as cooking loss in the LD muscle (Table 5.16). The MYO1G-SNP2 also showed significant effects on pH of both the ST and LD muscles (Table 5.16).

Table 5.16 Tests of significance of variants on cooking loss and pH with the effects of MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530 in the model. Variants ST pH LD pH ST cooking loss LD cooking loss CAPN5-SNP17 0.005* 0.009* NS 0.049* 0.028* <.001*** MYO1G-SNP2 NS NS n=8(TT) n=8(TT) SNIP1-SNP3 NS NS 0.084† NS †( P<0.10); * (P<0.05); ** (P<0.005); *** (P<0.001); NS: not significant

168

The cattle with the GG genotype of CAPN5-SNP17 had significantly lower pH in the

ST muscle than the cattle with the AA genotype. The cattle with the AA genotype of

CAPN5-SNP17 had a 0.8% higher pH in ST muscle than the cattle with the GG genotype (Figure 5.35). The genotypes of the CAPN5-SNP17 explained 2.7% of the variation in ST pH. The G allele was recessive. For the additive effect, the estimated allelic substitution effect for G allele was -0.02 ± 0.007 pH units (P=0.001). A dominance effect was not observed.

5.74 a 5.72

5.7 b b 5.68

pHst pHst (unit) 5.66

5.64

5.62 AA GA GG CAPN5-SNP17 genotypes

Figure 5.35 Effect of CAPN5-SNP17 genotypes on ST pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

The cattle with the GG genotype of CAPN5-SNP17 also had a significantly lower pH in the LD muscle than the cattle with the AA genotype. The cattle with the AA genotype of CAPN5-SNP17 had a 0.8% higher pH in LD muscle than the cattle with the GG genotype (Figure 5.36). The genotypes of the CAPN5-SNP17 explained

1.9% of the variation in LD pH. The G allele was recessive. For the additive effect, the estimated allelic substitution effect for G allele was -0.024 ± 0.008 pH units

(P=0.005). A dominance effect was not observed.

169

5.7 5.68 a

5.66

5.64 b b

5.62 pHld (unit) pHld 5.6 5.58 5.56 AA GA GG CAPN5-SNP17 genotypes

Figure 5.36 Effect of CAPN5-SNP17 genotypes on LD pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

This SNP also affected cooking loss. The cattle with the AA genotype of CAPN5-

SNP17 had significantly lower cooking loss in LD muscle than the cattle with the

GG genotype. The cattle with the GG genotype of CAPN5-SNP17 had a 3.2% higher cooking loss in the LD muscle than the cattle with the AA genotype (Figure 5.37).

The genotypes of the CAPN5-SNP17 explained 1.6% of the variation in LD cooking loss. The C allele was recessive. For the additive effect, the estimated allelic substitution effect for T allele was 0.345 ± 0.14% (P=0.015). A dominance effect was not observed.

22.4 22.2 b 22 a,b

21.8

21.6 a

21.4 clld (%) clld 21.2 21 20.8 20.6 AA GA GG CAPN5-SNP17 genptypes

Figure 5.37 Effect of CAPN5-SNP17 genotypes on LD cooking loss. Bars represent standard errors. Letters indicate significant differences between genotypes.

170

The cattle with the CC genotype of MYO1G-SNP2 also had a significantly lower pH in the ST muscle than the cattle with the TT genotype. The cattle with the TT genotype of CAPN5-SNP17 had a 1.6% higher pH in the ST muscle than the cattle with the CC genotype (Figure 5.38). The genotypes of the MYO1G-SNP2 explained

1.9% of the variation in ST pH. The C allele was recessive. For the additive effect, the estimated allelic substitution effect for T allele was 0.048 ± 0.018 pH units

(P=0.0012). A dominance effect was observed (-0.052±0.021) (P=0.014).

5.85 b 5.8

5.75

5.7 a a

pHst pHst (unit) 5.65

5.6

5.55 CC TC TT MYO1G-SNP2 genotypes

Figure 5.38 Effect of MYO1G-SNP2 genotypes on ST pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

The cattle with the CC genotype of MYO1G-SNP2 had significantly lower pH in the

LD muscle than the cattle with the TT genotype. The cattle with the TT genotype of

CAPN5-SNP17 had a 2.4% higher pH in the LD muscle than the cattle with the CC genotype (Figure 5.39). The genotypes of the MYO1G-SNP2 explained 2.9% of the variation in LD pH. The C allele was recessive. For the additive effect, the estimated allelic substitution effect for T allele was 0.096 ± 0.019 pH units (P<0.001). A dominance effect was observed (-0.081±0.023) (P<0.001).

171

5.85 5.8 b 5.75

5.7 5.65 a a

pHld (unit) pHld 5.6 5.55 5.5 5.45 CC TC TT MYO1G-SNP2 genptypes

Figure 5.39 Effect of MYO1G-SNP2 genotypes on LD pH. Bars represent standard errors. Letters indicate significant differences between genotypes.

5.3.4 Simultaneous effects of candidate gene variants on tenderness

The individual effects of all the candidate gene variants were examined by using an

ASREML mixed model to take into account all of the effects of the variants simultaneously (Model 11). Based on experience with analysis of other traits, any variant explaining 5% (in contrast to 1% when fitted as individual fixed effects) or more of the variation was considered important to contributing to the trait (W.

Pitchford, unpublished data). When the variation explained by the polymorphism is non-zero, but less than 5%, these variants still could have real effects on the traits.

However, the possibility of a false positive result increases,

Ten traits were analysed (Table 5.17). The pH and cooking loss were included to determine if these traits are related to or contributed to the 3 major tenderness traits of adjusted shear force, ageing rate and compression of the LD and ST muscles.

From the analysis, the CAPN1-SNP316 showed the largest effect on adjusted shear force for both the LD and ST muscles (Appendices 24 and 25). This is consistent

172 with the known effects of CAPN1-SNP316 on tenderness. SNIP1-SNP3 is the only variant in the candidate genes that significantly affected adjusted shear force of the

LD muscle (Table 5.12). However, the variation in adjusted shear force on LD muscle accounted for by SNIP1-SNP3 was only 0.75%. The MYO1G-SNP5 explained 2.11% of the variation in adjusted shear force of the LD muscle. These findings were not in total agreement with the results from the single variant analyses from the fixed models. In the fixed model analysis (model 5), the MYO1G-SNP5 had no effect on adjusted shear force although SNP1-SNP3 showed a significant effect on adjusted shear force in LD muscle and explained 2.2% of the variation (P=0.019).

For the adjusted shear force of the ST muscle using the fixed model analysis, FST-

SNP7 and MYO1G-SNP2 were the only two variants associated with this trait.

However, in the ASREML mixed model (Gilmour et al., 2006), the effect of the

FST-SNP7 was small (0.71%, n=3, AA) and the effect of the MYO1G-SNP2 was not present (n=8, TT). Instead, FST-SNP5 and FSTL1-SNP1 (n=4, TT) individually contributed 3.25% of the variation. LOXL1-SNP1 and CAPN5-SNP21 also explained small fractions of the variation.

For ageing rate, there were two variants, MYO1G-SNP2 (n=8, TT) and SST-SNP2, which explained more than 5% of the variation in ageing rate in the LD muscle using the ASREML mixed model. Except for the effect of MYO1G-SNP2 on ageing rate and cooking loss in the LD muscle, there was no variant that affected more than 2 traits in the same muscle. Within the same gene, MYO1G-SNP3 (n=3, AA) explained a large proportion of the variation of the compression in the ST muscle (13.09%).

Interestingly, MYO1G-SNP3 greatly affected compression but not adjusted shear force. On the other hand, the MYO1G-SNP3 was not associated with any of the traits

173 using the individual fixed model analyses.

174

Table 5.17 Variation accounted for by candidate gene variants in mixed model.*

pHld pHst Clld Clst Variants Adjusted LD Adjusted ST Ageing rate LD Ageing rate ST Compression ST LOXL1-SNP1 0.00% 1.34% 1.31% 0.00% 0.00% 0.00% 0.00% 0.00% 1.36% CAPN4- 3 base repeat 0.00% 0.00% 0.00% 0.33% 0.00% 0.00% 0.00% 0.83% 3.40% CAPN1-SNP316 13.76% 11.58% 0.00% 1.01% 0.00% 4.63% 3.72% 0.00% 0.00% CAPN1-SNP530 0.00% 0.00% 1.51% 0.00% 0.00% 0.00% 0.00% 0.00% 0.17% FST-SNP5 0.00% 3.25% 0.00% 0.56% 0.00% 0.00% 0.00% 0.00% 0.00% FST- SNP7 0.00% 0.71% 0.00% 3.65% 0.00% 0.00% 0.56% 0.00% 0.00% FSTL1-SNP1 0.00% 3.25% 4.44% 3.37% 0.07% 0.14% 0.00% 0.00% 1.19% MYL7-SNP1 0.00% 0.00% 0.00% 0.00% 0.00% 1.09% 0.00% 0.25% 0.55% MYO1G-SNP2 0.00% 0.00% 7.79% 0.00% 0.00% 0.35% 0.64% 6.71% 0.00% MYO1G-SNP3 0.00% 0.00% 0.00% 0.00% 13.09% 0.00% 0.00% 0.00% 0.00% SST-SNP2 0.00% 0.00% 5.53% 0.00% 0.00% 0.13% 0.00% 0.00% 0.00% CAPN5-SNP16 0.00% 0.00% 0.00% 0.00% 0.00% 1.05% 0.00% 0.00% 0.00% CAPN5-SNP21 0.00% 0.22% 0.00% 0.00% 0.00% 0.00% 4.25% 0.00% 0.00% CAPN5-SNP17 0.00% 0.00% 0.00% 0.00% 0.00% 2.95% 9.06% 3.84% 2.04% IGF1-SNP1 0.00% 0.00% 0.00% 0.32% 0.00% 0.00% 0.00% 0.00% 0.00% IGF1-SNP2 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 1.45% LOX-SNP1 0.00% 0.02% 1.45% 0.00% 0.00% 0.16% 0.06% 0.00% 0.00% IGF1R-SNP1 0.00% 1.01% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% SNIP1-SNP3 0.75% 0.00% 1.38% 0.00% 4.49% 1.34% 0.00% 0.00% 3.40% MYO1G-SNP5 2.11% 1.28% 0.00% 1.03% 0.00% 0.00% 0.00% 0.00% 0.00% *Model 11

175

5.4. Discussion

Shear force is regarded as the major objective measurement for tenderness and the best method that imitates biting (Boccard et al., 1981). Compression is seen as more representative of the chewing process (Lepetit and Culioli, 1994). Other characteristics of the meat which are considered to be potentially associated with tenderness or toughness include cooking loss, pH, collagen content and muscle weight.

There was no relationship between collagen content and shear force or compression discovered (Chapter 2). Therefore, it is not a surprise that although the CAPN4 3 base repeat and the IGF1-SNP1 were found to be associated with collagen content in the ST muscle, these two variants did not affect shear force or compression. The results indicate that collagen content may not be a critical factor for tenderness or the change in the collagen content by the CAPN4 3 base repeat and the IGF1-SNP1 was not large enough to influence shear force or compression.

Furthermore, the variants affecting the pH of LD and ST muscles, CAPN5-SNP16 and MYO1G-SNP2 (identified by using model 3), did not show any relationship with shear force or compression. The results are consistent with the earlier observations because the pH was not found to be associated with compression and only explained a small amount of the variation in shear force (Chapter 2). Notably, the limited pH range in muscles herein is not likely to cause a large change in proteolysis (Yu and

Lee, 1986; Purchas, 1990).

The effects of 20 variants in 12 candidate genes on these tenderness related traits were examined with and without the MSTN F94L, CAPN1-SNP316 and CAPN1-

SNP530 genotypes in the model. In general, previous association studies identifying

176

DNA variants affecting tenderness have not taken the effects of major genes, such as myostatin and calpain 1, into account (Frylinck et al., 2009; Curi et al., 2010). By taking these major genes into account, more of the residual variation in tenderness traits can be explained by additional genes. However, in order to determine whether there is an interaction between these major genes and the new candidate gene variants, additional analyses must be conducted (Chapter 6).

Each trait was related to one or more variants in the candidate genes except for the compression and ageing rate of the ST muscle. The muscle specific effect of the variants was shown in the results. For example, CAPN1 contributed to tenderness during the very early stages of ageing and its effects lasted throughout the ageing period. However, the effect of CAPN1-SNP316 on shear force was muscle specific in that the CAPN1-SNP316 alleles had an opposite effect on the LD and ST muscles

(Figures 5.1 and 5.2).

Most of the other variants that were significant for shear force, pH, ageing rate and muscle weight were also specific in terms of the muscles. There was no single variant that significantly affected both the LD and ST muscles for any trait except for the effects of CAPN5-SNP17 and MYO1G-SNP2 on pH (Table 5.18). This is presumably due to subtle differences in the biological pathways for most of the traits in the LD and ST muscles.

177

Table 5.18 Comparison of SNPs significantly associated with tenderness traits with a size of effect of 1% or greater.

IGF1- FST- SNIP1 MYO1G CAPN4-3 CAPN5- CAPN5 Traits SNP1 SNP7 -SNP3 -SNP2 base repeat SNP17 -SNP21 LD adjusted *, ++ shear force ST adjusted *, ++ *, ++ shear force ST ageing rate

LD ageing rate †,+ *, ++

ST *, +++ **, ++ hydroxyproline LD weight *, +

ST weight *, ++++ *, + *, +

ST pH *, + **, ++

LD pH ***, ++ **, +

ST cooking loss LD cooking *, + loss P-value = † < 0.1, * < 0.05, ** < 0.01, *** < 0.001 Overall size of effect = + =1-2%, ++ = 2-3%, +++ = 3-4, ++++ < 4

In general, the larger or heavier the muscle, the more tender the meat (Bailey et al.,

1982). Muscle weight is thought to be important to tenderness for two reasons.

Firstly, it is believed that a change in muscle weight implies an effect of hypertrophy or hyperplasia, which supposedly would affect tenderness (Renand, et al., 2001;

Allais, et al., 2010). Secondly, a change of the muscle weight may possibly alter the portion of collagen which is seen as a resistant force to shearing (Albrecht et al.,

2006; Lines et al., 2009). The results indicate that the AA genotype of CAPN5-

SNP21 resulted in 9% and 7% heavier ST and LD muscle weights, respectively. The animals with AA genotype of the SNIP1-SNP3 had 18% heavier ST muscles than other genotypes. Neither of these two variants showed any effect on shear force in response to the change in muscle weight. However, the animals with the AA

178 genotype of the FST-SNP7 had an increased ST muscle weight of 97% of the ST muscle weight. This was associated with adjusted shear force in ST muscle although not with wbst1 and wbst26 (Appendices 18-23). Unfortunately, there were only 3 animals with the AA FST-SNP7 genotype. So the number of animals with the AA genotype of the FST-SNP7 was too low to be confident that the size of the effects will be observed in other populations.

SNIP1-SNP3, FST-SNP5 and FST-SNP7 are variants of proteins in the myostatin pathway of muscle development (Kim, et al., 2000; Sidis, et al., 2006; Funkenstein, et al., 2009). These three variants may be expected to mediate the effect of myostatin on muscle development. Myostatin is believed to cause hyperplasia which increases the muscle size and weight by increasing the number of muscle cells (Swatland and

Kieffer, 1974; Wegner et al., 2000; Lines et al., 2009). This increase in muscle mass is believed to be associated with the observed increase in tenderness (Swatland and

Kieffer, 1974; Wegner et al., 2000; Bellinge et al., 2005). Since both FST and SNIP1 are part of the myostatin pathway, they were expected to affect muscle weight, and thus, tenderness. Indeed, when the MSTN and CAPN1 variants were included in the model, both SNIP1 and FST affected the muscle weight of the ST muscle. FST also affected the adjusted shear force of the ST muscle. Interestingly though, the SNIP1-

SNP3 variant had no effect on the ST adjusted shear force but did affect the LD adjusted shear force. This is in spite of the fact that SNIP1-SNP3 did not affect the

LD muscle weight. The results are not entirely consistent with the expectation that

SNIP1 and FST would affect muscle weight and shear force in a manner similar to

MSTN. The observations suggest that SNIP1 may affect tenderness through other unknown pathways and/or that the MSTN effects on tenderness may not involve hyperplasia.

179

Lastly, IGF1 is also believed to cause hypertrophy or hyperplasia in muscle

(Fernandez et al., 1995; Musaro et al., 2001). In this case, the IGF1-SNP1 did affect muscle weight and hydroxyproline, but did not affect shear force, ageing rate or compression. These findings suggest that an increase in muscle size and/or change in collagen content do not appear to automatically increase tenderness in that muscle.

There may not been enough variation in the muscle weight or collagen content associated with the IGF1-SNP1 to be detected as an effect on tenderness. However, the key may be the proportional change of collagen to myofibres in the increased muscle mass.

A related observation was that only half of the effect of the MSTN gene on tenderness appears to be due to compression. The effect of MSTN on shear force tenderness at day 26 of ageing was 4.6%. However, the effect of MSTN on tenderness only decreased to 2.3% when the compression was fitted into the model as a covariate. It is not obvious how MSTN affects tenderness given that muscle weight and collagen content did not explain all of the variation accounted for by

MSTN (Lines et al., 2009).

The analyses showed that ageing rate was independent of adjusted shear force. The variants affecting ageing rate were also different from the variants affecting adjusted shear force with the exception of SNIP1-SNP3. SNIP1-SNP3 showed an effect on ageing rate (P=0.078) and shear force (adjusted and non-adjusted) of the LD muscle.

Therefore, the mechanism underlying the effect of SNIP1-SNP3 on ageing rate and adjusted shear force of the LD muscle may be different from the other variants.

The effects of the three major genes (MSTN F94L, CAPN1-SNP316 and CAPN1-

180

SNP530) were used in the models to investigate the influence of these three variants on candidate variants (Table 5.19). The models with/without the three major genes demonstrated that a significant effect of a gene can be offset or affected by another gene. This implies that some of the effects of genes on the traits may be undiscovered without the consideration of other genes in the model.

Table 5.19 Comparison of significant candidate gene variants on tenderness traits with/without the effects of MSTN F94L, CAPN1-SNP316 and SNP530 genotypes in the model. SNPs without MSTN + SNPs with MSTN + CAPN Traits CAPN genotypes genotypes LD adjusted shear force SNIP1-SNP3*

MYO1G-SNP2*, n=8(TT) ST adjusted shear force FSTL1-SNP2*, n=5(GG) FST-SNP7*, n=3(AA) MYO1G-SNP2*, n=8(TT) MYO1G-SNP2*, n=8(TT) SST-SNP2* SST-SNP2† LD ageing rate FSTL1-SNP1†, n=4(TT) SNIP1-SNP3† FSTL1-SNP2*, n=5(GG) LOX-SNP1† ST ageing rate

ST compression SNIP1-SNP3 †

IGF1-SNP1* ST hydroxyproline IGF1-SNP1* MYO1G-SNP3†, n=3(AA) CAPN4-3 base repeat* CAPN5-SNP21† CAPN5-SNP21† LD weight IGF1-SNP1† IGF1-SNP1* SST-SNP2† CAPN5-SNP21* CAPN5-SNP21* ST weight IGF1-SNP1* SNIP1-SNP3*

FST-SNP7***, n=3(AA) CAPN5-SNP16* CAPN5-SNP17* LD pH MYO1G-SNP2***, n=8(TT) MYO1G-SNP2***, n=8(TT) MYO1G-SNP2*, n=8(TT) MYO1G-SNP2*, n=8(TT) LD pH CAPN5-SNP17* CAPN5-SNP17* CAPN5-SNP17* LD cooking loss SST-SNP2* MYO1G-SNP2†, n=8(TT) SNIP1-SNP3† SNIP1-SNP3† ST cooking loss CAPN5-SNP17† †( P<0.10); * (P<0.05); ** (P<0.005); *** (P<0.001)

181

The CAPN1-SNP316 genotype has been included in a molecular marker-assisted selection system for tenderness (http://www.nbcec.org/validation/igenity/ tenderness.html), and incorporating other known tenderness-related genes in the prediction of tenderness would be a reasonable extension. However, care must be taken if there are interactions between these genes (Chapter 6).

By comparing the inclusion of the MSTN F94L and CAPN1 SNPs in the models, three phenomena were observed (Table 5.19). Firstly, the models with the MSTN

F94L and CAPN1 SNPs showed that some of the variants were not associated with traits without the MSTN F94L and CAPN1 SNPs in the models. Namely, the effects of the SNIP1-SNP3, MYO1G-SNP2 and FST-7 on LD adjusted shear force, CAPN4-

3 base repeat on hydroxyproline in ST, IGF1-SNP1 on LD weight, and CAPN5-

SNP17 on LD pH were only detected with of the MSTN F94L and CAPN1 SNPs in the model. This is most likely due to the fact that once the variation explained by these major gene SNPs had been removed, the other remaining variation can then be explained by the candidate genes.

Secondly, some effects of the variants in the models without MSTN F94L and

CAPN1 SNPs disappeared when the MSTN F94L and CAPN1 SNPs were included in the model. Namely, the effects of the FSTL1-SNP2 on ST adjusted shear force, FST-

SNP7 on ST ageing day 1 shear force, FSTL1-SNP2 and SST-SNP2 on ageing rate in

LD, IGF1-SNP1 on ST weight, CAPN5-SNP16 on LD pH and SST-SNP2 on LD cooking loss disappeared. It is likely to be that the MSTN F94L and CAPN1 SNPs are acting epistatically with these candidate gene variants and therefore, the variation can be accounted for by including either the MSTN and CAPN1 variants or the candidate gene variants in the model.

182

Thirdly, the results for some SNPs did not change whether or not the MSTN and

CAPN1 variants were included in the models. Namely, the effects of SNIP1-SNP3 on

LD ageing day 26 shear force, FSTL1-SNP1 and FSTL1-SNP2 on ST ageing day 1 shear force, MYO1G-SNP2 on ST adjusted shear force and LD ageing rate, IGF1-

SNP1 on hydroxyproline of ST, CAPN5-SNP21 on ST weight, MYO1G-SNP2 on LD and ST pH and CAPN5-SNP17 on LD cooking loss were unchanged. This indicates that the effects of these variants are independent of the effects of these three major tenderness genes for their respective traits.

The mixed model (model 11) provided an opportunity to infer the effect of each variant when the other variants were also fitted into the model simultaneously. The mixed model theoretically offers a more comprehensive result in that all the variants are considered together. Adjusted shear force was slightly affected by some variants in the mixed model. Although they did not explain more than 5% of variation, the importance of SNIP1, MYO1G and FST variants on adjusted shear force was evident.

The variants of the MYO1G gene demonstrated a significant role on several traits including ageing rate, compression and cooking loss. However, based on recent results which indicate that MYO1G is primarily expressed in haematopoietic cells

(Olety, et al., 2010), it does not seem likely that MYO1G itself is directly affecting these traits. It is more likely that the MYO1G–SNP2 is in strong linkage disequilibrium with the causative QTN. If the linkage disequilibrium is strong, then the MYO1G–SNP2 can still serve as a DNA marker for selection. However, in terms of understanding the biology of tenderness, further analysis of genes in the region will be required.

183

Even though the variants examined herein did not always show a significant effect on a trait, the importance of a given variant cannot be easily excluded. The effect may appear in different herds or breeds. For example, two studies have demonstrated the importance of the activity of cathepsins, particularly cathepsin B on tenderness (Calkins and Seideman, 1988; Johnson et al., 1990). Johnson et al.,

(1990) showed that the pure Angus steers produced meat that was more tender and had a higher concentration of the cathepsin than cattle that were ½ Angus or ¼

Angus. Therefore, cathepsin B is considered to be an important candidate for affecting tenderness. However, the variants in cathepsin B did not have any effect on tenderness in the cattle population herein. Nevertheless, the effect of cathepsin B on tenderness in other cattle breeds or populations cannot be excluded as there are several possible explanations for the lack of association herein. The most obvious of the explanations is that the DNA variant(s) causing the effect are not present in the

Jersey-Limousin backcross population.

Likewise, one cannot assume that those DNA variants shown to affect tenderness traits herein will be associated with tenderness in other beef cattle populations or other breeds. This is because beef cattle are relatively outbred and the linkage disequilibrium between DNA markers has been estimated to be a low as 10 kb (de

Roos et al., 2008). Consequently, the likelihood that a given marker and the causative variant will segregate together across breeds is greatly reduced. The marker itself must be causative or very tightly linked to the causative variant (within

10 kb). As a result, all the variants discovered to affect tenderness herein must be verified in other larger cattle populations. The most obvious approach to confirm the effects of these variants will be to impute the genotypes into larger datasets based on genome-wide association scans (Bolormaa et al., 2011). These scans involve

184 anonymous SNP markers and not necessarily SNPs within candidate genes. However, once the number of markers is sufficiently dense (eg genotype data from 800K SNP chips instead of 50K SNP chips), then imputation will determine which anonymous

SNPs are in linkage disequilibrium with the variants. The variants are verified if these anonymous SNPs also affect tenderness.

185

Chapter 6: Gene Interactions

186

6.1. Introduction

Traits may be not only affected by the additive effects of unrelated single genes and the environment. Epistatic interactions between genes may also have large effects on traits that are equal to or even greater than the individual gene effects. While QTL with large effects can be detected by mapping, the QTL effects that are influenced by other loci or the genetic background have been generally ignored. However, as early as 2000, Brockmann et al., (2000) showed epistatic effects on growth and obesity in mice. Carlborg et al., (2003) demonstrated in chickens that a QTL for the growth was affected by another QTL.

In their review, Carlborg and Haley (2004) postulated that gene interactions would affect the behaviour of QTL. They discussed the neglect of epitasis as a phenomenon in complex trait studies. Nevertheless, a variety of the methods have been developed to try to tackle the interaction issue in QTL mapping wherein analytical models that incorporate gene interactions are used (Kao et al., 2002; Bogdan et al., 2004; Isobe et al., 2007; Zak et al., 2007; Estelle et al., 2008; Duthie et al., 2010).

Genes with large effects on tenderness, such as MSTN or CAPN1, have been reported previously (Casas et al., 2000; Page et al., 2002). However, there have been only a few reports on the interactions between these genes related to tenderness. Only the interaction between µ-calpain gene (CAPN1) and calpastatin gene (CAST) has been shown to have significant effects (P<0.05) on tenderness (Casas et al., 2006b;

Barendse et al., 2007). Therefore, in order to provide better predictions for marker- assisted selection and to better elucidate potential pathways affecting tenderness, gene interaction analyses were conducted herein.

187

6.2 Methods

The studies were performed by using three models (Models 12~14) for identifying epistasis between the variants in the candidate genes and while accounting for the three critical DNA variants (MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530) affecting tenderness. Three models were used for the studies of interactions.

Model 12: Model to identify effects of interactions between candidate genes on adjusted shear force

Model 12 was designed to show the interaction effects between the candidate gene variants for adjusted shear force. For adjusted shear force, the model did not include cohort, breed, sire or myostatin genotype because these were incorporated into the derivation of the trait.

Yikl = μ + αi + (γλ)kl + Rikl

Yikl the response variable

μ the overall mean

th αi the effect of i calpain 1 (CAPN1-SNP316) genotype (GG,GC,CC )

th th (γλ)kl interaction between the γ of the k variant and the λ of the l variant

Rijkl the residual effect

Model 13: Model 12: Model to identify effects of interactions between candidate genes on other tenderness traits

Model 13 was designed to show the interaction effects between candidate gene variants and the CAPN1-SNP316, CAPN1-SNP530 and MSTN F94L genotypes for all traits other than adjusted shear force.

Yijklmno = μ + αi + βj + γk + λl +δm + θn + τo+ (λτ) lo+ (δτ) mo+ (θτ) no + Rijklmno,

where

Yijklmno = response variable,

188

μ = overall mean,

th αi = the effect of i cohort (six levels),

th βj = effect of j breed of dam (Jersey or Limousin)

th γk = effect of k sire (three sires)

th λl = effect of l myostatin F94L genotype (AA, CA, CC)

th δm = effect of m calpain 1 (CAPN1-SNP361) genotype (GG,GC,CC )

th θn = effect of n calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA),

th τo = effect of o SNP variant

th (λτ) lo = effect of the interaction between l myostatin F94L genotype (AA,

CA, CC) and oth SNP variant,

th (δτ) mo = effect of the interaction between m calpain 1 (CAPN1-SNP361)

genotype (GG,GC,CC ) and oth SNP variant,

th (θτ) no = effect of the interaction between n calpain 1 (CAPN1-SNP530)

genotype (GG,AG,AA) and oth SNP variant and,

Rijklmno = residual effect,

Model 14: Model to identify effects of interactions between candidate genes and major genes on adjusted shear force

Model 14 was designed to show the interaction effects between candidate gene variants and the CAPN1-SNP316, CAPN1-SNP530 and MSTN F94L genotypes for adjusted shear force. For adjusted shear force, the model did not include cohort, breed, sire or myostatin genotype because these were incorporated into the derivation of the trait.

Ymno = μ + δm + θn + τo+ (λτ) lo+ (δτ) mo+ (θτ) no + emno,

where

Ymno = response variable,

189

μ = overall mean,

th δm = effect of m calpain 1 (CAPN1-SNP361) genotype (GG,GC,CC ),

th θn = effect of n calpain 1 (CAPN1-SNP530) genotype (GG,AG,AA),

th τo = effect of o SNP variant,

th (λτ) lo = effect of the interaction between l myostatin F94L genotype (AA,

CA, CC) and oth SNP variant,

th (δτ) mo = effect of the interaction between m calpain 1 (CAPN1-SNP361)

genotype (GG,GC,CC ) and oth SNP variant,

th (θτ) no = effect of the interaction between n calpain 1 (CAPN1-SNP530)

genotype (GG,AG,AA) and oth SNP variant and,

emno = residual effect.

Traits Seven major groups of traits were used to study of the interactions between the genes.

1. Tenderness: adjusted shear force for LD muscle, adjusted shear force for ST muscle, shear force at day 1 and day 26 for LD and ST muscles (wbld_adjusted, wbst_adjusted, wbld1, wbld26, wbst1 and wbst26 in kgF).

2. Ageing rate: ageing rate in LD muscle (lnkgF) and ageing rate in ST muscle

(lnkgF)

3. Chewiness: compression in ST muscle (kgF)

4. Collagen content: hydroxyproline concentration in ST muscle (mg/g)

5. Muscle weight: LD weight and ST weight (kg)

6. pH: ultimate pH of LD and ST muscles (pHld and pHst in pH units)

7. Cooking loss: cooking loss of LD and ST muscles (clld and clst in percentage)

6.3 Results

The studies were performed by using three models to identify epistasis between the

190 variants in the candidate genes and also the three critical DNA variants (MSTN F94L,

CAPN1-SNP316 and CAPN1-SNP530) affecting tenderness. Investigating the interactions between candidate genes with different functions has the potential to detect new pathways affecting tenderness. In some cases, the number of animals in specific genotypic classes were low (Appendix 17B) and the size of effect could not be accurately estimated. In those cases where there were less than 10 animals with a particular combination but the size of effect could be estimated, it was noted.

Nevertheless, the size of effect must be large or the interaction would not be detected.

Therefore, these cases were still noted.

The interactions between genes were analysed by two approaches. The first approach was to examine the interactions between all SNPs in the candidate genes for adjusted shear force as this trait was of primary interest (Model 12). The second approach investigated the interactions between candidate genes and the 3 SNPs with known major tenderness effects, MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530

(Appendix 26) (Models 13 and 14) for all traits. Two models were required for the second approach because for adjusted shear force, some of the fixed effects were already taken into account (namely, cohort, sire, breed and myostatin genotype).

From the analysis of the individual SNP effects (chapter 5), the 3 major known SNPs

(MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530) were not associated with ageing rate but were associated with shear force. The MSTN F94L genotype affected shear force in the ST muscle, as well as compression, hydroxyproline content, muscle weight of both muscles and cooking loss in the ST muscle. The CAPN1-

SNP316 genotype showed an effect on shear force for both LD and ST muscles and

ST pH. The CAPN1-SNP530 genotype was associated with adjusted shear force in

191 the LD muscle, shear force at day 1 of ageing in the LD muscle, compression in the

ST muscle and cooking loss in the LD muscle.

6.3.1 Candidate gene interactions with MSTN F94L, CAPN1-SNP316 and CAPN1-SNP530 on tenderness traits

All the candidate gene variants were examined for interactions with the 3 known major DNA variants affecting tenderness (MSTN F94L, CAPN1-SNP316 and

CAPN1-SNP530) using models 13 and 14. Even though an individual candidate gene variant or known major DNA variant did not affect tenderness related traits, they frequently were found to exert an effect through gene interactions as described below.

6.3.1.1 Adjusted shear force

From the analysis of the candidate gene epistasis on adjusted shear force, two interactions and three interactions were discovered for adjusted shear force of the LD and ST muscles, respectively (Table 6.1). These interactions did not affect any other traits that were not related to shear force. The number of the SNIP1-SNP3 significant interactions with other genes was higher than for other candidate genes. Therefore, the effect of SNIP1-SNP3 on tenderization appears to be greater than other variants.

Table 6.1 Significant interactions between genes affecting adjusted shear force Trait DNA variants P-values wbld_adjusted SNIP1-SNP3 and CAPN1-SNP316 0.002 wbld_adjusted SNIP1-SNP3 and CAPN1-SNP530 0.045 wbld_adjusted FST-SNP7 and CAPN1-SNP530 0.015 wbst_adjusted SNIP1-SNP3 and CAPN1-SNP316 <0.001 wbst_adjusted FST-SNP7 and CAPN1-SNP530 0.054

There were no interactions with the MSTN F94L genotype which affected the adjusted shear force for either the LD or ST muscles. However, the SNIP1-SNP3 did significantly interact with CAPN1-SNP316 on adjusted shear force of both the LD

192 and ST muscles.

Interestingly, SNIP1-SNP3 did not show an individual effect on adjusted shear force for the ST muscle. However, the effect of the interaction with CAPN1-SNP316 was significant for the ST muscle. CAPN1-SNP316 alone explained 5.2% of the variation in adjusted shear force for the ST muscle. The effect of the interaction between

CAPN1-SNP316 and SNIP1-SNP3 accounted for 9.5% of the variation in adjusted shear force for the ST muscle (Table 6.2). Therefore, the interaction showed a greater effect on tenderness than the individual CAPN1-SNP316. Animals with the TT allele of SNIP1-SNP3 had 20% tougher meat than other genotypes of SNIP1-SNP3. The average for the combination of TT genotype of SNIP1-SNP3 and GG genotypes of

CAPN1-SNP316 showed 23% higher adjusted shear force for the ST muscle than the other combinations of CC and GC genotypes for CAPN1-SNP316 (Figure 6.1).

Notably, the combination of TT genotype of SNIP1-SNP3 and CC/GC genotypes of

CAPN1-SNP316 produced relatively lower adjusted shear force values for the ST muscle than the other genotype combinations. This was not expected given the TT genotype of SNIP1-SNP3 alone makes meat tougher.

Table 6.2 Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and two variants, CAPN1-SNP316 and SNP530, on adjusted shear force of ST muscle. Proportion of the total DNA variants P-value sum of squares (%) CAPN1-SNP316 5.21 <.001 CAPN1-SNP530 0.01 0.982 SNIP1-SNP3 0.26 0.621 SNIP1-SNP3 and MSTN F94L 1.83 0.342 CAPN1-SNP316 and SNIP1-SNP3 9.45 <.001 CAPN1-SNP530 and SNIP1-SNP3 1.93 0.130

193

4 CAPN1-SNP316/CC 3 CAPN1-SNP316/GC CAPN1-SNP316/GG 2 wbst_djusted(KgF) 4 7 7 5

0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 6.1 Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on adjusted shear force of ST muscle. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

For the LD muscle, the CAPN1-SNP316 alone explained 4.7% of the variation of adjusted shear force and the effect of the interaction between CAPN1-SNP316 and

SNIP1-SNP3 accounted for 4.9% of the variation in adjusted shear force (Table 6.3).

Furthermore, the effect of this combination of the two genotypes was greater than the individual effect of SNIP1-SNP3 (2.3%). The interaction between TT genotype of the SNIP1-SNP3 and GG genotype of the CAPN1-SNP316 produced meat with 6 kgF for adjusted shear force (Figure 6.2). This implies that this combination of genotypes can produce unacceptably tough meat.

Table 6.3 Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and two variants (CAPN1-SNP316 and SNP530) on

adjusted shear force of LD muscle.

Proportion of the total DNA variants P-value sum of squares (%) CAPN1-SNP316 4.76 <.001 CAPN1-SNP530 1.31 0.093 SNIP1-SNP3 2.32 0.015 SNIP1-SNP3 and MSTN F94L 1.13 0.661 CAPN1-SNP316 and SNIP1-SNP3 4.90 0.002 CAPN1-SNP530 and SNIP1-SNP3 2.71 0.045

194

4 CAPN1-SNP316/CC 3 CAPN1-SNP316/GC CAPN1-SNP316/GG 2 wbld_djusted(KgF) 4 7 7 5 1

0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 6.2 Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on adjusted shear force of LD muscle. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

The SNIP1-SNP3 also interacted with CAPN1-SNP530 on adjusted shear force of the LD muscle and the interaction explained 2.7% of the variation in the adjusted shear force. The interaction between the TT genotype of the SNIP1-SNP3 and the

GG genotype of the CAPN1-SNP530 produced 35% tougher meat than the other combinations of genotypes of the SNIP1-SNP3 with the GG genotype of the

CAPN1-SNP530 (Figure 6.3). The polymorphism CAPN1-SNP530 itself was not associated with adjusted shear force (Chapter 5). Hence, this result highlights the importance of interactions between genes and the effect of epitasis. The relatively low adjusted shear force values were decreased by the interaction between the TT genotype of the SNIP1-SNP3 and the AA genotype of the CAPN1-SNP530 (n=1).

However, given the small number of animals with the combination of the TT genotype of the SNIP1-SNP3 and the genotypes of CAPN1-SNP530, these effects on adjusted shear force value for LD muscle should be confirmed.

195

4 CAPN1-SNP530/AA 3 CAPN1-SNP530/AG CAPN1-SNP530/GG

wbld_djusted(KgF) 2

1 1 8 0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 6.3 Interaction of SNIP1-SNP3 and CAPN1-SNP530 genotypes on adjusted shear force of LD muscle. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

The FST-SNP7 interacted with CAPN1-SNP530 to affect adjusted shear force in both the LD and ST muscles (Table 6.4 and 6.5). However, the individual effect of the FST-SNP7 was only observed for the ST muscle, not the LD muscle. The proportions of the total sum of squares for the adjusted shear force for LD and ST muscles explained by the interaction between FST-SNP7 and CAPN1-SNP530 were

2.45% and 1.7%, respectively.

Table 6.4 Proportion of the total sum of squares accounted for by the interaction between FST-SNP7 and two variants (CAPN1-SNP316 and SNP530) on adjusted shear force of LD muscle. Proportion of the total DNA variants P-value sum of squares (%) CAPN1-SNP316 4.69 <.001 CAPN1-SNP530 1.37 0.094 FST-SNP7 0.69 0.305 FST-SNP7 and MSTN F94L 1.98 0.334 CAPN1-SNP316 and FST-SNP7 1.08 0.156 CAPN1-SNP530 and FST-SNP7 2.45 0.015

196

Table 6.5 Proportion of the total sum of squares accounted for by the interaction between FST-SNP7 and two variants (CAPN1-SNP316 and SNP530) on adjusted shear force of ST muscle. Proportion of the total DNA variants P-value sum of squares (%) CAPN1-SNP316 5.20 <.001 CAPN1-SNP530 0.01 0.982 FST-SNP7 2.30 0.020 FST-SNP7 and MSTN F94L 1.42 0.556 CAPN1-SNP316 and FST-SNP7 1.04 0.168 CAPN1-SNP530 and FST-SNP7 1.70 0.054

6.3.1.2 Ageing rate

The two major genes myostatin and calpain 1 alone had no effect on ageing rate

(described in chapter 5). However, two interactions significantly affected ageing rate of the LD muscle and 4 interactions significantly affected ageing rate of the ST muscle (Table 6.6). All the interactions except two (SST-SNP2 and CAPN1-SNP316 plus CAPN4 3 base repeat and MSTN F94L) only affected ageing rate. These two interactions (SST-SNP2 with CAPN1-SNP316 and CAPN4 3 base repeat with MSTN

F94L) were also associated with ST and LD day 1 shear forces (wbst1 and wbld1), respectively (Appendix 22).

Table 6.6 Significant interactions between major genes affecting ageing rate. Trait DNA variants P-value LD ageing rate SST-SNP2 and CAPN1-SNP316 0.045 LD ageing rate LOX-SNP1 and CAPN1-SNP530 0.016 ST ageing rate CAPN4 3 base repeat and MSTN F94L 0.003 ST ageing rate IGF1-SNP1 and CAPN1-SNP530 0.003 ST ageing rate LOX-SNP1 and CAPN1-SNP530 0.007 ST ageing rate MYO1G-SNP3 and CAPN1-SNP530 0.033

The interaction between SST-SNP2 and CAPN1-SNP316 explained 1.66% of the total sum of the squares for ageing rate of the LD muscle (Table 6.7). The CT

197 genotype of SST-SNP2 and the GG genotype of CAPN1-SNP316 tended to lower the ageing rate. The interaction between these two genotypes had accumulated effects and the interaction between SST-SNP2/CT and CAPN1-SNP316/GG genotypes gave the lowest ageing rate for the LD muscle (Figure 6.4).

Table 6.7 Proportion of the total sum of squares accounted for by the interaction between SST-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for LD muscle. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 0.14 0.766 CAPN1-SNP316 0.88 0.19 CAPN1-SNP530 1.32 0.084 SST-SNP2 0.80 0.082 MSTN F94L.SST-SNP2 0.47 0.409 CAPN1-SNP316 and SST-SNP2 1.66 0.045 CAPN1-SNP530 and SST-SNP2 0.34 0.526

0.6

0.5

0.4

0.3 SST2/CC 0.2 SST2/TC

0.1 2 Ageing Ageing rate LD for muscle (log Kg) 0 CAPN1-SNP316/CC CAPN1-SNP316/GCCAPN1-SNP316/GG

Figure 6.4 Interaction of SST-SNP2 and CAPN1-SNP316 genotypes on ageing rate for LD muscle. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

The interaction between LOX-SNP1 and CAPN1-SNP530 explained a relatively high proportion of the total sum of squares (3.14%) (Table 6.8) Interestingly, the interaction between the CT genotype of LOX-SNP1 and the AA genotype of CAPN1-

198

SNP530 resulted in a relative higher ageing rate than most of other interactions

(Figure 6.5). This implies that in the cattle with the AA genotype of CAPN1-SNP530, the LOX-SNP1 may have a positive dominance effect.

Table 6.8 Proportion of the total sum of squares accounted for by the interaction between LOX-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for LD muscle. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 0.13 0.776 CAPN1-SNP316 0.86 0.184 CAPN1-SNP530 1.41 0.063 LOX-SNP1 0.95 0.155 MSTN F94L and LOX-SNP1 0.53 0.716 CAPN1-SNP316 and LOX-SNP1 0.59 0.677 CAPN1-SNP530 and LOX-SNP1 3.14 0.016

0.45

0.4 0.35 0.3

0.25 CAPN1-SNP530/AA 0.2 CAPN1-SNP530/AG 0.15 CAPN1-SNP530/GG 0.1

Ageing Ageing rate LD for muscle (log Kg) 0.05 7 0 LOX-SNP1/CC LOX-SNP1/CT LOX-SNP1/TT

Figure 6.5 Interaction of LOX-SNP1 and CAPN1-SNP316 genotypes on ageing rate for LD muscle. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

Of all the interactions that were significantly associated with ageing rate for the ST muscle, the CAPN1-SNP530 interacted with three candidate gene SNPs. On the other hand, the CAPN1-SNP316 had no interaction affecting ageing rate for the ST muscle herein. Furthermore, the interaction between MSTN F94L and CAPN4-3 base repeat was the only one significantly affecting ageing rate involving MSTN F94L.

199

The interaction between MSTN F94L and the CAPN4-3 base repeat explained 4.81% of the total sum of squares, which is the largest effect on ageing rate discovered in this study (Table 6.9).

Table 6.9 Proportion of the total sum of squares accounted for by the interaction between CAPN4 -3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 0.69 0.316 CAPN1-SNP316 1.47 0.087 CAPN1-SNP530 0.77 0.276 CAPN4 -3 base repeat 0.58 0.585 MSTN F94L and CAPN4 -3 base repeat 4.81 0.003 CAPN1-SNP316 and CAPN4 -3 base repeat 0.49 0.897 CAPN1-SNP530 and CAPN4 -3 base repeat 1.61 0.375

The interaction between IGF1-SNP and CAPN1-SNP530 had the largest significant effect on ageing rate with the involvement of CAPN1-SNP530. The interaction between IGF1-SNP and CAPN1-SNP3530 explained 4.77% of the total sum of squares (Table 6.10). The interaction between LOX1-SNP1 and CAPN1-SNP3530 explained 4.43% of the total sum of squares (Table 6.11). The interaction between

MYO1G-SNP3 and CAPN1-SNP3530 explained 2.12% of the total sum of squares

(Table 6.12). This contribution of MYO1G-SNP3 to ageing rate for ST muscle was the smallest observed.

Table 6.10 Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 0.66 0.327 CAPN1-SNP316 1.51 0.080 CAPN1-SNP530 0.74 0.290 IGF1-SNP1 0.92 0.212 MSTN F94L and IGF1-SNP1 0.90 0.551 CAPN1-SNP316 and IGF1-SNP1 0.72 0.488 CAPN1-SNP530 and IGF1-SNP1 4.77 0.003

200

Table 6.11 Proportion of the total sum of squares accounted for by the interaction between LOX1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 0.78 0.279 CAPN1-SNP316 1.37 0.107 CAPN1-SNP530 0.78 0.277 LOX1-SNP1 0.06 0.905 MSTN F94L and LOX-SNP1 0.77 0.635 CAPN1-SNP316 and LOX-SNP1 0.36 0.879 CAPN1-SNP530 and LOX-SNP1 4.34 0.007

Table 6.12 Proportion of the total sum of squares accounted for by the interaction between MYO1G-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ageing rate for ST muscle. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 0.87 0.246 CAPN1-SNP316 1.78 0.057 CAPN1-SNP530 0.65 0.352 MYO1G-SNP3 0.19 0.729 MSTN F94L and MYO1G-SNP3 0.80 0.276 CAPN1-SNP316 and MYO1G-SNP3 0.39 0.531 CAPN1-SNP530 and MYO1G-SNP3 2.12 0.033

6.3.1.3 Compression

MSTN F94L and CAPN1-SNP530 had significant individual effects on compression in the ST muscle (Chapter 5). However, there were no significant interactions involving CAPN1-SNP530 found (Table 6.13). The interaction between FSTL1-

SNP1 and MSTN F94L showed a result similar to the interaction between FSTL1-

SNP2 and MSTN F94L. Both of these interactions explained more of the variation in compression than MSTN F94L itself (Table 6.14 and 6.15).

Table 6.13 Significant interactions between genes affecting compression. Trait DNA variants P-values ST compression FSTL1-SNP1 and MSTN F94L 0.004 ST compression FSTL1-SNP2 and MSTN F94L 0.004 ST compression IGF1-SNP1 and CAPN1-SNP316 0.021

201

The genotype of FSTL1-SNP1 was nearly in complete linkage disequilibrium with the genotype of FSTL1-SNP2. Hence, the effect of the interaction between FSTL1-

SNP1 and MSTN F94L and the interaction between FSTL1-SNP2 and MSTN F94L could not be easily distinguished. The CAPN1-SNP316 and IGF1-SNP1 were not individually associated with compression. However, the interaction between IGF1-

SNP1 and CAPN1-SNP316 accounted for 3.68% of the total sum of squares for the compression. The effect of the interaction is even greater than the effect of CAPN1-

SNP530 alone (Table 6.16).

Table 6.14 Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on compression. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 4.23 0.003 CAPN1-SNP316 0.78 0.337 CAPN1-SNP530 2.72 0.023 FSTL1-SNP1 0.44 0.536 MSTN F94L and FSTL1-SNP1 4.85 0.004 CAPN1-SNP316 and FSTL1-SNP1 0.63 0.622 CAPN1-SNP530 and FSTL1-SNP1 0.55 0.463

Table 6.15 Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on compression. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 4.58 0.003 CAPN1-SNP316 1.09 0.238 CAPN1-SNP530 2.04 0.069 FSTL1-SNP1 0.62 0.441 MSTN F94L and FSTL1-SNP2 5.18 0.004 CAPN1-SNP316 and FSTL1-SNP2 0.87 0.512 CAPN1-SNP530 and FSTL1-SNP2 0.42 0.574

202

Table 6.16 Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on compression. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 4.25 0.004 CAPN1-SNP316 0.68 0.400 CAPN1-SNP530 2.37 0.042 IGF1-SNP1 0.00 0.995 MSTN F94L and IGF1-SNP1 0.97 0.620 CAPN1-SNP316 and IGF1-SNP1 3.68 0.021 CAPN1-SNP530 and IGF1-SNP1 1.20 0.515

6.3.1.4 Hydroxyproline content of ST muscle

MSTN F94L was the only one of the three major DNA variants that affected hydroxyproline concentration. MSTN F94L interacted with three candidate variants to affect hydroxyproline content (Table 6.17). Although two of these interactions

(IGF1-SNP2 with MSTN F94L and CAPN4 3 base repeat with MSTN F94L) did not reach the P<0.05 significance threshold, their P-values are still small enough that these two interactions should not be ignored.

Table 6.17 Significant interactions between genes affecting hydroxyproline content. Trait DNA variants P-value ST hydroxyproline CAPN5-SNP16 and MSTN F94L 0.046 ST hydroxyproline IGF1-SNP2 and MSTN F94L 0.051 ST hydroxyproline CAPN4-3 base repeat and MSTN F94L 0.054

These three interactions explained the total sum of squares ranging from 2.45% to

3.1% (Table 6.18, 6.19 and 6.20). The combined effects of these interactions were as large as MSTN F94L alone (5.8%). The DNA variants in the three candidate genes were not found to be associated with hydroxyproline content individually. CAPN1-

SNP316 and SNP530 had no effect on hydroxyproline content even though CAPN1 interacted with other candidate genes.

203

Table 6.18 Proportion of the total sum of squares accounted for by the interaction between CAPN5-SNP16 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on hydroxyproline content. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 5.75 <.001 CAPN1-SNP316 0.04 0.950 CAPN1-SNP530 1.35 0.182 CAPN5-SNP16 0.17 0.809 MSTN F94L and CAPN5-SNP16 2.45 0.046 CAPN1-SNP316 and CAPN5-SNP16 0.16 0.816 CAPN1-SNP530 and CAPN5-SNP16 1.08 0.434

Table 6.19 Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on hydroxyproline content. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 5.58 <.001 CAPN1-SNP316 0.04 0.944 CAPN1-SNP530 1.34 0.178 IGF1-SNP2 0.12 0.855 MSTN F94L and IGF1-SNP2 3.04 0.051 CAPN1-SNP316 and IGF1-SNP2 1.32 0.332 CAPN1-SNP530 and IGF1-SNP2 1.92 0.176

Table 6.20 Proportion of the total sum of squares accounted for by the interaction between CAPN4 3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on hydroxyproline content. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 5.59 0.001 CAPN1-SNP316 0.04 0.946 CAPN1-SNP530 1.33 0.192 CAPN4 -3 base repeat 0.48 0.754 MSTN F94L and CAPN4 -3 base repeat 3.10 0.054 CAPN1-SNP316 and CAPN4 -3 base repeat 0.10 0.992 CAPN1-SNP530 and CAPN4 -3 base repeat 0.93 0.675

204

6.3.1.5 Muscle weights

There are 6 interactions for LD weight and 4 interactions for ST weight discovered that involved the major DNA variants. MSTN F94L was the only one of the three major variants with an individual effect on muscle weight. On the other hand,

CAPN1 only affected muscle weight by interacting with other genes (Table 6.21).

All the candidate genes affecting muscle weight through these interactions were related to protein turnover except LOXL1.

Table 6.21 Significant interactions between genes affecting muscle weight. Traits DNA variants P-values LD Weight CAPN5-SNP21 and MSTN F94L 0.027 LD Weight IGFR1-SNP1 and MSTN F94L 0.026 LD Weight FST-SNP5 and MSTN F94L 0.010 LD Weight FST-SNP7 and MSTN F94L 0.002 LD Weight MYO1G-SNP5 and MSTN F94L 0.001 LD Weight LOXL1-SNP1 and CAPN1-SNP530 0.043 ST Weight CAPN5-SNP21 and MSTN F94L <0.001 ST Weight SST-SNP2 and MSTN F94L 0.029 ST Weight FST-SNP5 and CAPN1-SNP316 0.024 ST Weight CAPN4-3 base repeat and CAPN1-SNP530 0.012

The interaction between CAPN5-SNP21 and MSTN F94L explained 1.67% of the total sum of squares for LD weight, which is marginally higher than MSTN F94L alone (1.5%) (Table 6.22). The interaction between CAPN5-SNP21 AA genotype and MSTN F94L AA genotype produced a greater LD muscle weight than the other interactions. Within the cattle carrying the MSTN F94L AA genotype, the cattle carrying the AA genotype of the CAPN5-SNP21 increased the muscle weight by almost 20% more than the cattle carrying the GG genotype of the CAPN5-SNP21

(Figure 6.6).

205

Table 6.22 Proportion of the total sum of squares accounted for by the interaction between CAPN5-SNP21 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight.

Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 1.50 0.007 CAPN1-SNP316 0.08 0.769 CAPN1-SNP530 0.76 0.082 CAPN5-SNP21 0.80 0.071 MSTN F94L and CAPN5-SNP21 1.67 0.027 CAPN1-SNP316 and CAPN5-SNP21 0.19 0.872 CAPN1-SNP530 and CAPN5-SNP21 0.93 0.190

4 MSTN/AA MSTN/AC 3 MSTN/CC

LD LD weight muslce (Kg) 2

1 7 7 0 CAPN5-SNP21/AA CAPN5-SNP21/GA CAPN5-SNP21/GG

Figure 6.6 Interaction of CAPN5-SNP21 and MSTN F94L genotypes on LD weight. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

All other interactions with MSTN F94L on LD weight had higher proportions of the total sum of squares for LD weight (range from 1.54 to 2.77%) than MSTN alone

(1.3%) (Table 6.23). The effects of these interactions (such as FST-SNP5 and MSTN

F94L or FST-SNP7 and MSTN F94L or MYO1G-SNP5 and MSTN F94L) contributed double the effect of MSTN F94L on LD weight (Table 6.24, 6.25 and 6.26). However, due to the unbalanced numbers of cattle in some classes of genotypes, the estimates of the effects on LD weight for these interactions could not be predicted. Hence, the

206 results of different combinations of genes on LD weight could not be accurately estimated. Nevertheless, the significant effects of those interactions are still important, especially given that the effect of MSTN F94L itself on LD weight is not very large.

Table 6.23 Proportion of the total sum of squares accounted for by the interaction between IGFR1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 1.31 0.013 CAPN1-SNP316 0.05 0.860 CAPN1-SNP530 0.82 0.066 IGFR1-SNP1 0.08 0.764 MSTN F94L and IGFR1-SNP1 1.40 0.026 CAPN1-SNP316 and IGFR1-SNP1 0.59 0.140 CAPN1-SNP530 and IGFR1-SNP1 0.69 0.102

Table 6.24 Proportion of the total sum of squares accounted for by the interaction between FST-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 1.35 0.012 CAPN1-SNP316 0.05 0.845 CAPN1-SNP530 0.86 0.059 FST-SNP5 0.48 0.204 MSTN F94L and FST-SNP5 1.76 0.010 CAPN1-SNP316 and FST-SNP5 0.49 0.360 CAPN1-SNP530 and FST-SNP5 0.21 0.714

Table 6.25 Proportion of the total sum of squares accounted for by the interaction between FST-SNP7 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 1.31 0.013 CAPN1-SNP316 0.05 0.859 CAPN1-SNP530 0.82 0.064 FST-SNP7 0.38 0.279 MSTN F94L and FST-SNP7 2.54 0.002 CAPN1-SNP316 and FST-SNP7 0.26 0.42 CAPN1-SNP530 and FST-SNP7 0.09 0.743

207

Table 6.26 Proportion of the total sum of squares accounted for by the interaction between MYO1G-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 1.32 0.012 CAPN1-SNP316 0.05 0.852 CAPN1-SNP530 0.81 0.065 MYO1G-SNP5 0.35 0.301 MSTN F94L and MYO1G-SNP5 2.77 0.001 CAPN1-SNP316 and MYO1G-SNP5 0.26 0.617 CAPN1-SNP530 and MYO1G-SNP5 0.50 0.491

The interaction between LOXL1-SNP1 and CAPN1-SNP530 explained 1.6% of the total sum of squares for LD weight, which is slightly higher than MSTN F94L alone

(1.5%) (Table 6.27). For the interaction between LOXL1-SNP1 and CAPN1-SNP530, there was no a clear trend or pattern found (Figure 6.7). However, the interaction between the GG genotype of LOXL1-SNP1 and GG genotype of CAPN1-SNP530 had an unexpected result in that these animals had higher LD muscle weights than other genotypes of LOXL1-SNP1 in the CAPN1-SNP530 GG class. The reason this was unexpected is that the animals with the CAPN1-SNP530 GG genotypes tended to have lower LD weights in general.,

Table 6.27 Proportion of the sum of total squares accounted for by the interaction between LOXL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 1.12 0.032 CAPN1-SNP316 0.05 0.868 CAPN1-SNP530 0.82 0.080 LOXL1-SNP1 0.05 0.852 MSTN F94L and LOXL1-SNP1 0.68 0.379 CAPN1-SNP316 and LOXL1-SNP1 0.91 0.231 CAPN1-SNP530 and LOXL1-SNP1 1.60 0.043

208

6 5 4 CAPN1-SNP530/AA 3 CAPN1-SNP530/AG

2 CAPN1-SNP530/GG LD LD weight muslce (Kg)

1 8 6 0 LOXL1-SNP1/AA LOXL1-SNP1/AGLOXL1-SNP1/GG

Figure 6.7 Interaction of LOXL1-SNP1 and MSTN F94L genotypes on LD weight. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

For ST weight, the MSTN F94L variant had a greater effect than on the LD muscle.

On average, the effect of MSTN F94L accounted for 8% of the total sum of squares.

On the other hand, the significant interactions contributing to ST weight did not have any large effects (range from 0.73% to 2.21%).

The CAPN5-SNP21 had a very small effect (0.94%) on ST weight. The interaction between MSTN F94L and CAPN5-SNP21 though explained 2.21% of the total sum of squares for ST weight (Table 6.28). The AA genotype of the MSTN F94L produced more ST muscle than the other two genotypes of MSTN F94L. However, the interaction between MSTN F94L AA and CAPN5-SNP21 GG genotypes demonstrated a negative effect of the GG genotype of CAPN5-SNP21 on ST weight

(Figure 6.8). The GG genotype of CAPN5-SNP21 counteracted the effect of the AA genotype of MSTN F94L. The effect of AA genotype of MSTN F94L on ST weight decreased the percentage of the GG genotype of CAPN5-SNP21 by 34%. Notably, the shear force did not change despite this change of muscle weight.

209

Table 6.28 Proportion of the total sum of squares accounted for by the interaction between CAPN5-SNP21 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 7.90 <.001 CAPN1-SNP316 0.08 0.653 CAPN1-SNP530 0.27 0.225 CAPN5-SNP21 0.94 0.006 MSTN F94L and CAPN5-SNP21 2.21 <.001 CAPN1-SNP316 and CAPN5-SNP21 0.02 0.996 CAPN1-SNP530 and CAPN5-SNP21 0.43 0.326

3.5

2.5

2 MSTN/AA 1.5 MSTN/AC MSTN/CC

1 ST muslce ST weight (Kg)

0.5 7 7 0 CAPN5-SNP21/AA CAPN5-SNP21/GA CAPN5-SNP21/GG

Figure 6.8 Interaction of CAPN5-SNP21 and MSTN F94L genotypes on ST weight. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

The interaction between MSTN F94L and SST-SNP2 explained only 0.73% of the total sum of squares for ST weight (Table 6.29). Because of the very small effect of this gene and the small difference between genotypes of SST-SNP2 in the various

MSTN F94L genotype classes (Figure 6.9), the importance of this interaction is likely to be minimal.

210

Table 6.29 Proportion of the total sum of squares accounted for by the interaction between SST-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 7.84 <.001 CAPN1-SNP316 0.05 0.798 CAPN1-SNP530 0.34 0.187 SST-SNP2 0.03 0.575 MSTN F94L and SST-SNP2 0.73 0.029 CAPN1-SNP316 and SST-SNP2 0.01 0.960 CAPN1-SNP530 and SST-SNP2 0.23 0.321

3.5

2.5

2 SST-SNP2/CC 1.5 SST-SNP2/TC

1 ST muslce ST weight (Kg)

0.5 8 0 MSTN/AA MSTN/AC MSTN/CC

Figure 6.9 Interaction of SST-SNP2 and MSTN F94L genotypes on ST weight. Bars represent standard errors. Numbers indicate where n<10 in genotypic class.

The other two interactions (FST-SNP5 with CAPN1-SNP316 and CAPN4 3 base repeat with CAPN1-SNP530) explained only 0.94% and 1.42% of the effect on ST weight, respectively (Table 6.30 and 6.31). However, due to the unbalanced numbers of cattle in some classes of genotypes, the estimates of ST weight for these interactions could not be accurately predicted.

211

Table 6.30 Proportion of the sum of squares accounted for by the interaction between FST-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 8.22 <.001 CAPN1-SNP316 0.05 0.793 CAPN1-SNP530 0.26 0.269 FST-SNP5 0.16 0.434 MSTN F94L and FST-SNP5 0.13 0.719 CAPN1-SNP316 and FST-SNP5 0.94 0.024 CAPN1-SNP530 and FST-SNP5 0.11 0.767

Table 6.31 Proportion of the sum of squares accounted for by the interaction between CAPN4-3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST muscle weight. Proportion of the total P- DNA variants sum of squares (%) value MSTN F94L 8.18 <.001 CAPN1-SNP316 0.05 0.786 CAPN1-SNP530 0.26 0.261 CAPN4 -3 base repeat 0.24 0.465 MSTN F94L and CAPN4 -3 base repeat 0.17 0.774 CAPN1-SNP316 and CAPN4 -3 base repeat 0.38 0.548 CAPN1-SNP530 and CAPN4 -3 base repeat 1.42 0.012

6.3.2 Interactions between all candidate genes including 3 major DNA variants (MSTN F94L, CAPN1-SNP316 and SNP530) on adjusted shear force

All the candidate genes were selected based on their function which may be involved in tenderness. The interactions between candidate genes on adjusted shear force may be able to unveil the complex relationships between genes that cannot be observed by examining single gene effects alone. Therefore, the web of the interactions between all the genes was examined for adjusted shear force as adjusted shear force represented the best measure of tenderness. The 20 variants from candidate genes and three major tenderness variants (MSTN F94L, CAPN1-SNP316 and SNP530) were examined in the analytical model in which all of the interactions were examine simultaneously (model 12).

212

6.3.2.1 LD adjusted shear force

For the LD muscle, there was a clear pattern in the matrix of gene interactions affecting adjusted shear force (Appendix 24). Adjusted shear force was mainly affected by variants interacting with SNIP1-SNP3. In fact, 9 out of the 22 variants significantly interacted with SNIP1-SNP3 on adjusted shear force for the LD muscle.

The SNIP1-SNP3 interactions included all types of candidate genes. This implies that SNIP1-SNP3 may be involved in several different physiological pathways.

However, there was no interaction between MSTN F94L and SNIP1-SNP3. This is not consistent with the knowledge that SNIP1 is a mediator of myostatin gene expression.

6.3.2.2 ST adjusted shear force

For the ST muscle adjusted shear force, the pattern of gene interactions was obviously different from the LD muscle (Appendix 25). The SNIP1-SNP3 significantly interacted with only two variants, CAPN1-SNP316 and CAPN5-SNP17, to affect adjusted shear force. Instead, the FST-SNP7 interacted with the largest number of candidate variants (9 of 22) on adjusted shear force of the ST muscle.

Moreover, interactions between FST and MSTN, IGF, IGFR1, LOX or LOXL1 were not observed. This implies that FST effects on the shear force of the ST muscle may not involve the MSTN, IGF or LOX pathways. Thus, the broad picture indicates differences in genes and physiological pathways controlling tenderness between the

LD and ST muscles.

213

Table 6.32 Proportion of the sum of squares accounted by all significant interactions.

LD adjusted shear ST adjusted shear Interaction LD ageing rate ST ageing rate ST compression ST hydroxyproline LD weight ST weight force force

SNIP1-SNP3 and CAPN1-SNP316 4.90 9.45

FST-SNP7 and CAPN1-SNP530 2.45 1.7

SNIP1-SNP3 and CAPN1-SNP530 2.71

SST-SNP2 and CAPN1-SNP316 1.66

LOX-SNP1 and CAPN1-SNP530 3.14 4.34 3.93

CAPN4 -3 base repeat and MSTN-F94L 4.81 3.10

IGF1-SNP1 and CAPN1-SNP530 4.77

LOXL1-SNP1 and CAPN1-SNP530 0.06 1.6

FSTL1_SNP2 and MSTN-F94L 5.18

FSTL1-SNP1 and MSTN-F94L 4.85

IGF1-SNP1 and CAPN1-SNP316 3.68

CAPN5-SNP10 and MSTN-F94L 2.45

IGF1-SNP2 and MSTN-F94L 3.04

CAPN5-SNP21 and MSTN-F94L 1.67 2.21

IGF1R-SNP1 and MSTN-F94L 1.40

SST-SNP2 and MSTN-F94L 0.73

FST-SNP5 and MSTN-F94L 1.76

FST-SNP7 and MSTN-F94L 2.54

MYO1G-SNP5 and MSTN-F94L 2.77

MYO1G-SNP2 and CAPN1-SNP316 0.094

FST-SNP5 and CAPN1-SNP316 0.94

214

6.4 Discussion

The results from the interaction analyses demonstrated that even though a variant of the candidate gene is not associated with a trait, it may still be significantly affecting the trait by interacting with another gene (Table 6.32). For example, the FST-SNP7 and SNIP1-SNP3 had no individual effects on adjusted shear force of the LD or ST muscles, respectively. However, these two genes significantly interacted with

CAPN1-SNP530 and CAPN1-SNP316 to affect adjusted shear force of the LD and

ST muscles, respectively. In addition, these 2 variants, FST-SNP7 and SNIP1-SNP3, interacted with the largest number of other candidate genes to affect adjusted shear force of these muscles. Moreover, in many instances, the interaction effects were larger than the individual SNP effects. For example, the interaction between CAPN1-

SNP316 and SNIP1-SNP3 had a larger effect on adjusted shear force of the ST muscle than the known tenderness related variant CAPN1-SNP316 alone.

These findings indicate the importance of gene interactions. Also, the results show that the effect of a gene on a given trait may be underestimated by solely observing its individual effect. Even though the variants such as SNIP1-SNP3 and FST-SNP7 do not affect adjusted shear force, their effects still exist through the interactions with variants of other genes. Therefore, the number of the interactions involving a variant and the strength of those interactions on a trait will also determine the importance of a gene.

Notably, the CAPN1-SNP316, is one of three major gene variants in the

GeneSTAR® tenderness panel used in the marker-assisted selection system for producing tender meat. However, the studies herein indicate the effects of the

CAPN1 (SNP316 and SNP530) on shear force are affected by epistatic interactions

215 with SNIP1-SNP3, FST-SNP7, FSTL1-SNP1 and SNP2, LOXL1-SNP1 and IGF1-

SNP2 and that these interactions can have unintended consequences. For example, the TT genotype of the SNIP1-SNP3 significantly increased the estimate of the GG genotype of the CAPN1-SNP316 for the adjusted shear force of both LD and ST muscles. However, the TT genotype of the SNIP1-SNP3 also significantly decreased the shear force of the GC genotype of the CAPN1-SNP316 in the ST muscle. Thus, the accuracy of the CAPN1-SNP316 in estimating the shear force of the meat may be limited and biased. Taking the effects of the variants interacting with the CAPN1-

SNP316 into account could be quite beneficial in marker-assisted selection schemes.

The epistatic interactions suggest that the tenderness of the LD and ST muscles may be controlled by different genes within different biological pathways, and not necessarily having effects through the expected networks. For example, the gene products of SNIP1 and FST are part of the myostatin pathway, which mediates the development of muscle in animals (Kim et al., 2000; Funkenstein et al., 2009; Sidis et al., 2006) (Chapter 7). The myostatin F94L gene variant is known to be associated with increased muscling and meat tenderness in cattle (Sellick et al., 2007;

Esmailizadeh et al., 2008; Lines et al., 2009). However, the polymorphisms in the

SNIP1 and FST genes (SNIP1-SNP3 and FST-SNP7) did not significantly interact with MSTN-F94L for shear force. Furthermore, the SNIP1-SNP3 did not show any interaction with MSTN-F94L on the muscle weight for either the LD or ST muscles.

On the other hand, the FST-SNP7 did interact with MSTN-F94L for LD muscle weight.

This demonstrates two points. Firstly, the SNP3 variant in SNIP1 may not play a large direct role in the regulation of the myostatin gene since the interaction between

216

SNIP1-SNP3 and MSTN-F94L did not appear to affect muscle development. SNIP1 is a known regulator of myostatin gene expression (Kim et al., 2000). However, the

SNIP1-SNP3 specific allele effects may not relate to a sufficiently large alteration in myostatin gene expression such that the myostatin pathway is affected appreciably.

Secondly, the effect of FST-SNP7 on shear force may not involve the pathway of myostatin. The FST-SNP7 was shown herein to interact with MSTN-F94L on LD muscle weight and was itself correlated with adjusted shear force in ST muscle. Yet, the change in the muscle LD weight from the interaction between FST-SNP7 and

MSTN-F94L did not affect the shear force of the LD muscle (Chapter 6.3.1.5).

Moreover, there was no significant effect of the interaction of FST-SNP7 and MSTN-

F94L on the ST shear force or muscle weight. Apparently, the mechanism by which

FST affects tenderness in the muscles may be through another unknown pathway.

Ten significant interactions on muscle weights were identified in the study. However, there were no significant interactions affecting muscle weights which also affected shear force, compression or ageing rate. Therefore, again there was no evidence supporting the hypothesis that muscle size affects tenderness (Seideman et al., 1988;

Crouse et al., 1991).

pH affects tenderness by changing the activities of various enzymes such as calpains and cathepsin related proteolysis and ATPases related muscle contraction (Yu and

Lee, 1986; Purchas, 1990; Lawrie 1992; Kanawa et al., 2002). The change of the cooking loss in beef and chicken is positively corrected with shear force (Harrell et al., 1978; Barbantia et al., 2005; Li et al., 2007b). However, there were not any interactions involving pH or cooking loss discovered unlike the other tenderness related traits. There may not have been sufficient variation in these traits for such

217 interactions to be discovered or the mechanisms that alter pH and cooking loss do not involve interactions between this particular set of gene variants.

The significant interactions were trait-specific with a few exceptions (Table 6.32).

The interaction between LOX-SNP1 and CAPN1-SNP530 affected ageing rate for both the LD and ST muscles. The interaction between these 2 SNPs also affected the content of the hydroxyproline in ST muscle. The effect of this interaction on these two traits in ST muscle could be coincidental because the ageing rate in ST muscle is not associated with the hydroxyproline or collagen content of the ST muscle. This is also true for the interaction between CAPN4 -3 base repeat and MSTN-F94L, which affected ageing rate in ST muscle and hydroxyproline. However, there were two interactions (between the CAPN4 -3 base repeat with MSTN-F94L and SST-SNP2 with CAPN1-SNP316) that showed their effects on ST ageing rate and day 1 shear force and on LD ageing rate for LD muscle and day 1 shear force, respectively.

Much of the variation in the ageing rates for the ST and LD muscles was explained by the ST day 1 shear force (wbst1 explained 28% of the variation of ageing rate in

ST muscle, p < 0.001) and the LD day 1 shear force (Appendix 22) (wbld1 explained

20% of the variation of ageing rate in LD muscle, p < 0.001). Thus, the fact that these two interactions are affecting these two traits is not likely to be a coincidence but is likely to involve the same mechanism.

Muscle-specific effects were also observed in most of the interactions (34 of 40 in total) (Table 6.32). There were only six exceptions (namely, SNIP1-SNP3 and

CAPN1-SNP316 on adjusted shear force and day 26 of ageing, FST-SNP7 and

CAPN1-SNP530 on adjusted shear force, SNIP1-SNP3 and CAPN1-SNP530 on day

1 of ageing, LOX-SNP1 and CAPN1-SNP530 on ageing rate and CAPN5-SNP21 and

218

MSTN-F94L on muscle weights). This indicates again that the tenderness traits are likely to be controlled by different mechanisms for the LD and ST muscles. It also raises another important point and that is, even if interactions are considered in marker-assisted selection schemes, the effects in one muscle are unlikely to be the same in other muscles. However, in those cases where the muscles are likely to be affected similarly (e.g. SNIP1-SNP3 and CAPN1-SNP316 interaction effects on shear force in the LD and ST muscles), then incorporating gene interaction effects into marker-assisted selection programs may be merited if the aim is to produce superior genotypes, not just those with the greatest breeding value (additive effect only).

Comparing the total of the proportions of the sum of the squares of all the significant individual SNPs for the traits with that of the significant interactions demonstrated that the interaction effects could contribute as much or more to the traits as the individual SNPs (Table 6.33). This is true even if the effects of the 3 known major variants (MSTN F94L, CAPN1-SNP316 and SNP530) are included.

Table 6.33 Comparison of the total of the proportions of the sum of the square of all the significant individual SNPs for the traits with that of the significant interactions. Traits Individual SNPs (%) Interactions (%) LD adjusted shear force 2.2 10.1 ST adjusted shear force 4.2 11.2 LD ageing rate 2.1 4.8 ST ageing rate 0 16.0 ST compression 0 13.7 ST hydroxyproline content 6.0 12.5 LD weight 1.3 11.7 ST weight 2.3 5.3

219

There were some extreme cases. For example, the three major variants (MSTN F94L,

CAPN1-SNP316 and SNP530) themselves and candidate variants did not have any individual effects on ageing rate for ST muscle, yet the interaction effects contributed up to 16% of the variation. Such observations underscore the potential importance of interactions.

Furthermore, for those traits affected by the three major known tenderness variants

(MSTN F94L, CAPN1-SNP316 and SNP530), the size of the effects of these variants may be under estimated, as the variants were commonly involved in significant interactions of large effect. Hence, the impact of the three major tenderness variants could be larger than expected.

220

Chapter 7: General Discussion

221

7.1 Introduction

Since the development of QTL (quantitative trait loci) mapping methods, (Soller et al., 1976; Lander and Botstein, 1989; Haley and Knott 1992; Zeng, 1993; Jansen,

1994; Fan and Xiong, 2002; Anderson et al., 2006), researchers have been able to locate the genomic positions of genes potentially controlling the traits of interest in livestock. QTL mapping has shown that genomic regions with large effects can be rather easily identified. However, the regions with small effects may be missed due to the inherent limitations of the QTL mapping methods (Lander and Botstein, 1989).

Furthermore, various factors in the experimental design may affect the efficiency of

QTL discovery, including breed, number of informative meioses, number and type of markers, linkage disequilibrium, allele frequencies, etc. The discovery of a QTL theoretically depends on the genetic variation within the breed or population, the number and location of the markers genotyped and the number of informative meioses (Lander and Botstein 1989; Darvasi et al., 1993; Casas et al., 2005; Curi et al., 2010; Zhang et al., 2010). Even though there are no QTL found within a given experimental cattle mapping herd for a trait, it is possible that QTL may be discovered in a larger population as the QTL may be of small effect. Alternatively,

QTL may be discovered in another breed or herd as the QTL may not be segregating within the given mapping population.

The lack of linkage disequilibrium (LD) between markers (e.g. microsatellites or single nucleotide polymorphisms) has been an issue limiting the accuracy of the

QTL prediction in beef cattle association studies. The LD between microsatellite markers interval decays quickly at distances greater than 40 centiMorgans (cM)

(Farnir, 2000). Using SNP markers expedites the decay of the LD with the distance between markers decreasing to less than 100 kb, even in purebred Holstein

222

(Sargolzaei et al., 2008). Therefore, Mckay et al., (2007) suggested that LD cannot be found if the distance between SNP markers is larger than 0.5 cM. Hence, for cattle, a minimum of 30,000~50,000 SNP loci are suggested for whole genome association studies. Commercial gene chips with 50K SNP markers provide a minimum marker density (37 kb on median interval) at least for dairy cattle research (Matukumalli et al., 2009). Khatkar et al., (2007) suggested that the number of the SNPs needed for forming haplotype blocks for association studies would be 250,000 for cattle. de

Roos et al., (2008) reported that 300,000 SNPs are required for linkage disequilibrium across divergent breeds.

7.2 QTL mapping for tenderness

QTL for tenderness have been mapped in other species such as pig (Rohrer et al.,

2006; Harmegnies et al., 2006; Edwards et al., 2008; Choi et al., 2011; Kim et al.,

2011) and sheep (Johnson et al., 2005; Karamichou et al., 2006). Information from other species can be invaluable as many of the same genes that affect tenderness in cattle also have been shown to affect tenderness across species. For example, tenderness has been associated with calpain 1 variants in chickens (Zhang et al.,

2008), and calpastatin variants in pigs (Ciobanu et al., 2004; Meyers and Beever

2008; Lindholm-Perry et al., 2009; Nonneman et al., 2011).

However, as mapped QTL are very broad and typically covered 40 cM, it is difficult to use comparative mapping to align these QTL accurately with the bovine genome.

Fortunately, in the beef industry, meat quality traits, including beef tenderness, have been investigated by using QTL mapping methods to a surprising extent given the need for large herds and the difficulty of measuring meat quality traits (Keele et al.,

1999; Casas et al., 2000, 2001, 2003; Rexroad et al., 2001; Drinkwater et al., 2006;

223

Gutie´rrez-Gil et al., 2007; Alexander et al., 2007; Davis et al., 2007; Esmailizadeh et al., 2011). QTL for tenderness have been published and collated into databases, such as the Cattle QTL database (CattleQTLdb) (http://www.animalgenome.org/cgi- bin/QTLdb/BT/draw_traitmap?trait_ID=1186).

However, even though significant QTL are identified, these can also be affected by epistasis and lead to biased results (Carlborg et al., 2006). Generally, epistasis has not been taken into account the QTL mapping although some QTL mapping has been conducted in chicken where epistatic effects were considered in order to more accurately estimate the size of effect of each QTL for the selection (Carlborg et al.,

2003; Carlborg and Haley, 2004).

Using the results of QTL mapping, tenderness genes such as calpain 1 (Page et al.,

2002), calpain 3 (Barendse et al., 2008), calpastatin (Barendse, 2002), lysyl oxidase

(Barendse, 2002), and myostatin (Casas et al., 2000) have been identified. Additional

QTL and gene variants within these QTL that affect beef tenderness and traits related to tenderness have been also been discovered herein.

The QTL for shear force discovered herein were different from those identified in a previous analysis of the same herd (Lines, 2006), with the exception of two QTL.

This difference is not unexpected because the shear force values used in the study herein were adjusted for the MSTN F94L genotype and interaction effects between

MSTN F94L and other fixed factors. The adjusted shear force values should be a more accurate description of the biological variation than individual ageing points because they utilise information from four time points (1, 5, 12, and 26 days of ageing post-slaughter). Furthermore, the effects of the CAPN-SNP316 and CAPN-

SNP530 were examined in the QTL mapping analysis. Thus, the QTL mapping

224 results for adjusted shear force herein should provide better estimates. In addition, ageing rate was derived as an additional “trait” and additional QTL were identified

(Table 3.10).

The QTL findings for calpain 1 in different studies and herein show that if the effect of the gene is reasonably large and occurs across breeds (Casas, et al. 2003; Casas, et al. 2005; Gutie´rrez-Gil, et al., 2007; Esmailizadeh et al., 2011), QTL mapping with relatively small sample sizes and numbers of markers can lead to identification of major genes. For minor genes (that is, genes of small effect), the sample size and number of markers must be larger. Thus, the performance of the QTL mapping analyses is quite dependent on the size of the gene effects. From another point of view, the QTL mapping is able to provide believable positions of QTL because all the discovered QTL should represent genes with large effects on the trait. Therefore, all new QTL found herein may contribute to the further discovery of the new genes for shear force and ageing rate.

Not all the significant candidate gene variants that affected adjusted shear force were located within adjusted shear force QTL. For example, three candidate gene variants,

SNIP1-SNP3 on chromosome 3, MYO1G-SNP2 on chromosome 4 and FST-SNP7 on chromosome 20, were associated with adjusted shear force (Table 5.11). Yet, none of these significant variants were detected in the QTL mapping analyses for adjusted shear force (Chapter 3.3). A reasonable explanation is that the single gene effects discovered in this association study were not sufficiently large to be observed in the

QTL mapping. Alternatively, it may be related to an insufficient number of informative meioses for the nearest microsatellite markers within the mapping herd.

It may also be related to the inherent limitations of the regression QTL mapping

225 method used.

The regression approach can cause biases and overestimate residual variance producing lower likelihood ratios, which makes the discovery of putative QTL difficult (Xu, 1995). Lander and Botstein (1989) emphasized that a putative QTL may not be identified when a QTL with small effect is closely linked with markers or when a QTL with large effect is only loosely linked with the adjacent markers.

These problems can be overcome in part by increasing the number of the markers with a high heterozygosity or using marker haplotypes (Grapes et al., 2006; Hayes et al., 2007). However, in most livestock QTL experiments, fine mapping QTL of small or moderate effects to small chromosomal regions is limited by the lack of large numbers of informative animals with adequate phenotypes. Although association studies using SNP chips with hundreds of thousands of markers provide better resolution, an even greater number of animals are required (Hayes et al., 2007). The cost of performing such whole genome scans and gathering tenderness data for so many animals can be prohibitive. Therefore, to date, QTL mapping is still worthwhile. However, it should be noted though that using SNP chips for genotyping should be better than mapping with microsatellites once it is cost effective to utilise chips with 500K or more markers because more markers will increase informativeness. To truly improve linkage mapping and association studies though, a greater number of animals with good phenotypes will provide more power than additional SNP markers (Yan et al., 2006).

226

7.3 Breed effects on tenderness

In this study, two extreme breeds, Limousin and Jersey, were used, and therefore, breed effects were expected. The statistical analyses confirmed this expectation. For tenderness, the breed effects were particularly noticeable in the ST muscle. When the effects of cohort and sire were taken account in the analysis, the breed effects explained 15% and 5% of the variation in the day 26 shear force values for the ST and LD muscles, respectively. The other traits were also affected by breed but to a lesser extent.

These breed effects are genetic effects in that the performance of one cattle breed differs from the performance of another cattle breed in the same environment because the two breeds have different combinations of alleles or allele frequencies.

As an example, the inadequate segregation and consequent small effect of the

CAPN1-SNP316 variant in Bos indicus was shown to be caused by the low minor allele frequency of the SNP316 in non-taurine breeds (Casas et al., 2005; Curi et al.,

2010). The segregation of CAPN1-SNP316 in Bos taurus was adequate because of the higher minor allele frequency, resulting in a sufficient number of informative meioses of each genotype to observe the effect of the allele (Page et al., 2004).

The combination of alleles can affect the performance in two ways though. Firstly, there can be breed differences in allele frequency (eg CAPN1-SNP316). Secondly, the alleles of one gene may interact with other genes to control the performance. For example, SNIP1-SNP3, which had no effect on shear force, interacted with CAPN1-

SNP316 to affect shear force significantly. Thus, it is the combination of alleles within a breed that affects the performance. A key gene may be inhibited or activated by a regulator gene allele only present in a single breed. Therefore, some genes may

227 be found to be associated with a trait in one breed, but do not affect the same trait to the same extent in another breed. These inconsistencies between breeds have been observed in the QTL mapping studies in different breeds (discussed in Chapter 3) and at the SNP level (Page et al., 2004; Casas et al., 2005; Curi et al., 2010).

Investigating the difference between breeds at the DNA level benefits the discovery of the mechanisms controlling tenderness.

For adjusted shear force, the interactions between ageing day (days 1, 5, 12 and 26) and breed (Limousin and Jersey) (Table 2.1, Figure 2.2) and between breed and muscle (LD and ST) were observed. The effect of the breed implies a different combination of genes is influencing a trait in Jersey versus Limousin. Hence, the difference caused by breeds may be reasonably ascribed to differences in allele frequencies. For example, in the experimental herd, the number of animals with the

CC genotype of the CAPN1-SNP360, which is associated with tender meat, was larger in the Jersey backcross than the Limousin backcross (Table 7.1). This may lead to the least square means for XJ being slightly lower than that for XL. At day 1 of ageing, the least squares means of the adjusted shear force is still higher in the

Jersey backcross than the Limousin backcross. It may be that the effect of CAPN1 is not present in both muscles or that other unknown factors are involved.

Table 7.1 Number of animals with CAPN1-SNP361 genotypes for each backcross. CAPN1- CAPN1- CAPN1- Cross SNP316/CC SNP316/GC SNP316/GG XJ 18 (10%) 98 (53%) 70 (37%) XL 4 (3%) 49 (33%) 97 (64%)

The interaction between breed and muscle may have a similar cause. The myostatin

AA genotype of the MSTN F94L results in more tender ST meat in these cattle (Lines

228 et al., 2009). In the experimental herd, there are 54 progeny with the AA genotype of the MSTN F94L in the backcross Limousin compared with zero in the Jersey backcross (Table 7.2). This will result in the lower shear forces in the ST muscle of the backcross Limousin progeny and can cause the interaction between breed and muscle.

Table 7.2 Number of animals with MSTN F94L genotypes for each backcross. Genotype XJ XL MSTN F94L AA 0 (0%) 54 (34%) MSTN F94L AC 100 (50%) 89 (56%) MSTN F94L CC 100 (50%) 16 (10%)

7.4. DNA variants affecting tenderness

Besides the non-synonymous variants within exons, the variants within introns, untranslated regions (UTRs), and flanking regions are worthy of attention as they may have important functions, such as mRNA stability, microRNA transcription or as regulatory element binding sites (Tokuhiro et al., 2003; Clop et al., 2006; Kawase et al., 2007; Bartel, 2009). At least one known tenderness gene is affected by microRNAs, namely the regulation of myostatin by mir1 and mir206 (Clop et al.,

2006), and demonstrates another potential mechanism by which tenderness may be controlled genetically. Thus, in addition to identifying potential candidate gene pathways, it is equally important to identify possible DNA regulatory elements, such as transcription factors and microRNA binding sites, near and within the candidate tenderness genes.

Intronic variants that were significantly associated with tenderness related traits were identified herein (Chapter 5). The most obvious mechanisms by which intronic

229 variants can affect a phenotype is by interfering with splice sites (Kawase et al.,

2007) or transcription factor binding sites (Tokuhiro et al., 2003). In our study, some of the intronic variants with significant effects on tenderness traits, such as FST-

SNP5 and SNP7, were not in the splicing sites or known transcription factor binding sites. It is most likely that there is another causative variant affecting the trait near these significant SNPs which is in linkage disequilibrium.

Significant intronic SNPs may not only be a marker but may have a functional role in controlling the trait besides splicing and transcription factor binding though. One such role would be as a microRNA binding site. MicroRNAs, which are short non- coding RNA of 20~23 nucleotides, have been identified as functional fragments in non-coding regions that can regulate genes. For example, as mentioned above, Clop et al., (2006) showed that the microRNAs mir1 and mir206 bind a fragment in the

3’UTR of MSTN to inhibit the expression of MSTN in Texel sheep, resulting in enhanced muscularity. Hence, it is possible that the significant variants in introns or untranslated regions can affect the coding sequence of a microRNA contained therein or affect the ability of a specific microRNA to bind to the sequence of the variant in a target gene. Therefore, determining if the variants in introns or untranslated regions are within microRNA coding sequences or microRNA binding sites is a valuable approach to establish whether the variants within introns and untranslated regions are likely to have a direct effect.

The variants with large effects discovered herein were examined for potential microRNA sites. However, no known sites were found. The only fragment that aligned elsewhere was a short fragment (GCCTAATAATAATAA/GTATCA) around

FST-SNP5 (http://www.ensembl.org/Bos_taurus/blastview, 2009), which was found

230 to complement a sequence in the 3’ UTR in the interleukin-1 alpha precursor (IL-1 alpha) gene (Figure 7.1).

Figure 7.1 Alignment of FST-SNP5 with IL-1. Short fragment around FST-SNP5 (GCCTAATAATAATAA/GTATCA) aligns to the 3’ UTR of the interleukin-1 alpha precursor gene (IL-1). The complementary region is highlighted with blue colour.

There is no evidence yet to support that the fragment including FST-SNP5 is a microRNA site (Figure 7.2). Nevertheless, examining the existence of the short RNA sequences including significant SNP may be a reasonable step to identifying potential roles of these non-coding SNPs. Therefore, genes known to affect tenderness should be examined more exhaustively for any potential microRNA sites that may affect expression of these genes.

Figure 7.2 miRNA predicted from FST intron 3. miRNA predicted from FST intron 3 sequence where the FST-SNP5 is located. Arrow indicates the site of the FST-SNP5 (A/C). (Predicted from RNAfold WebServer at http://rna.tbi.univie.ac.at//cgi-bin/ RNAfold.cgi?PAGE=3&ID=NRi0o2s vb9.

231

For example, Iwanowska et al. (2011) demonstrated that polymorphisms within the

3’UTR of the CAPN4 gene are associated with shear force and muscle degradation.

The 3’UTR affects the stability of the mRNA and is often the target of microRNA binding, which affects gene expression (Wang, et al., 2006; Liu, et al., 2010). Either could be the case for the CAPN4 mRNA. Unfortunately, to date, there are no good tools for the efficient discovery nucleotide changes that affect mRNA stability or microRNA binding sites. Therefore, the potential role of this SNP in CAPN4 should be further investigated using gene expression studies in experimental herds with tenderness phenotypes.

Another example is the CAPN1 Gene. Using the human microRNA database, the human CAPN1 gene was found to have 6 potential miRNA binding sites were found in the 3’UTR of the human CAPN1 gene (Table 7.3). No such sites were identified in the cattle CAPN1 gene herein. This is not surprising though because not all microRNAs have been characterised in cattle. However, once available, it could be worth verifying if there are any microRNA coding or binding sequences in the cattle

CAPN1 gene and any other genes of large effect.

Table 7.3 MicroRNA and microRNA binding sites in the 3’UTR of the human CAPN1 gene.

microRNA microRNA sequence microRNA binding sites

hsa-miR-1207-3p 3'cuuuacucccGGUCGACu 5' 5' ccaggccaccCCAGCUGc 3'

hsa-miR-1301 3' cuUCAGUGAGGGUCCGUCGACGUu 5' 5' ccAGGC-CACCC---CAGCUGCAa 3'

hsa-miR-1204 3' uauuaccUCUGGUCCGGUGcu 5' 5' ccacaccACACCAGGCCACcc 3'

hsa-miR-1293 3' cgUGUUUAGAGGUCUGGUGGGu 5' 5' ccACACCACACCAGGCCACCCc 3'

hsa-miR-608 3' ugccucgacagGGUUGUGGUGGGGa 5' 5' accacaccacaCCAG-GCCACCCCa 3'

hsa-miR-138 3' gccGGACUAAGUGUUGUGGUCga 5' 5' ccuCCUACCACACCACACCAGgc 3' (Data from the microRNA Organisation website at http://www.microrna.org/microrna/ getMrna.do?gene=uc001odf.1&organism=9606#).

232

7.5. Signalling pathways affecting tenderness

Myostatin is an important gene affecting meat yield and tenderness in cattle (Casas et al., 2000; Karim et al., 2000; Sellick et al., 2007; Lines et al., 2009). In the myostatin pathway, FST, SMAD1, SMAD2, SMAD4 and SNIP1 are involved in controlling the activity of MSTN (Figure 7.3) (Kim et al., 2000; Sidis et al., 2006;

Funkenstein et al., 2009). The binding of the myostatin to the ActRIIB receptor induces a phosphorylation cascade that results in the SMAD transduction system suppression of the expression of myogenic genes. Follistatin prevents the binding of myostatin to ActRIIB (Dominique and Gérard, 2006). SNIP1 controls of the activity of myostatin by competing with the SMAD4 co-repressor (Kim, et. al., 2000).

SNIP1

Figure 7.3 Myostatin pathway and potential interactions of the candidate gene proteins. Diamonds represent myostatin in the activated and inactivated forms. Myostatin is inhibited from binding its receptor by follistatin (Foll). A cascade effect after myostatin binding results in the SMAD complex translocating to the nucleus and blocking gene expression of myogenic proteins in competition with SNIP1. (Adapted from Dominique and Gérard, 2006.)

233

Based on this knowledge, three SNPs, FST-SNP5, FST-SNP7 and SNIP1-SNP3, were expected to interact with MSTN F94L in the association studies. Interestingly, none of these SNPs interacted with MSTN F94L to affect shear force, but they did interact to affect the muscle weights of the LD and ST. Furthermore, although FST-

SNP5 and FST-SNP7 affected tenderness of the ST muscle like MSTN F94L, unlike

MSTN F94L, FST-SNP5 and FST-SNP7 alone did not affect muscle weights. On the other hand, SNIP1-SNP3 did not show any effect on tenderness in the ST muscle, but did have an effect on the LD muscle tenderness instead. The observations in this study do not correlate with the hypothesis herein that if MSTN affects shear force in the ST muscle, then related genes in the same pathway would have the same effect on shear force in ST muscle. Moreover, since there was no interaction between FST and MSTN F94L on shear force, the results suggest that the mechanism by which

MSTN affects shear force may be independent of the mechanism by which FST and

SNIP1 affect shear force.

SNIP1 is not only an inhibitor of MSTN, but also is a strong inhibitor of CBP/p300

(Kim et al., 2000). CBP/p300 is a complex of closely related transcriptional co- activating proteins. The CBP/p300 is involved in many signalling pathways that interact with numerous transcription factors (Figure 7.4) (Vo and Goodman, 2001).

Since SNIP1 can influence the activity of CBP/p300, SNIP1 may affect the expression of any gene that CBP/p300 controls, and therefore, affect tenderness through many other pathways. Examining the various CBP/p300 pathways involving

SNIP1 could help determine the mechanism by which SNIP1 affects tenderness.

234

A NOTE: This figure/table/image has been removed to comply with copyright regulations. It is included in the print copy of the thesis held by the University of Adelaide Library.

Figure 7.4 CBP/p300 complex and interacting proteins. (From Goodman and Smolik, 2000.)

One of the most interesting SNIP1 pathways, in this regard, involves the proteasomal system. Lin et al., (2002) reported that CBP/p300 and SNIP1 play an important role in the transduction pathway of the bone morphogenetic proteins (BMPs), which are members of the TGF-β superfamily, like myostatin. In the BMP pathway, the bone morphogenetic proteins play a role in the activation of the proteasomal system

(Figure 7.5) (Gruendler et al., 2001). The SMADs (particularly, Smad1 and Smad4) and SNIP1 interact with the proteasomal system to control transcription of downstream target genes (Guo et al., 2000; Lin et al. 2002).

235

Figure 7.5 Signalling pathway of BMPs involving proteasomes. (From Lin et al., 2002.)

Initially, Smad1, proteasome β subunit HsN3 and Az induce the formation of 20S proteasomes that mature into 26S proteasomes (Figure 7.6) (Lin et al. 2002). Smad1, proteasome β subunit HsN3, Az and Smad4 also translocate into the nucleus and bind various DNA transcription factors. Then, CBP/p300 and SNIP1 complexes attach to the Smad1/4/HsN3/Az complex where SNIP1 acts as a stabiliser of the

Smad1/4/HsN3/Az complex (Figure 7.5). In order to remove this complex and allow

CBP/p300 to trigger transcription, the degradation of the SNIP1 by the proteasomes is essential (Griffin et al., 2000; Kim et al., 2001; Lin et al. (2002). The removal of

236

SNIP1 promotes the degradation of the Smad1/4/HsN3/Az complex, permitting

CBP/p300 to help activate the transcription of genes.

Figure 7.6 Formation of proteasomes with the Smad1/pro-HsN3/Az complex. (From Lin et al., 2002.)

Because the role of proteasomes includes the ability to digest proteins within the muscle structure, potential effects of proteasomes on tenderness have been suggested

(Robert et al., 1999; Dutaud et al., 2006) and an association recently demonstrated

(Oury, et al., 2009). However, if proteasomes can also affect the function of important modulators, such as SNIP1, then proteasomes may play a wider role.

Whether the interplay between SNIP1, SMADs, CBP/p300, and proteasomes could drive the expression of genes related to tenderness is not clear as the downstream targets of this pathway are largely unknown. Investigating the potential role of proteasomes on tenderness should be considered. Of note, the proteasome beta 10 subunit gene (PSMB10) is located at 34 Mb on BTA 18 and is a strong positional candidate gene based on the QTL mapping results (Chapter 3).

237

7.6. Role of myostatin on tenderness

It has been generally assumed that the MSTN effects on tenderness are related directly to the muscle hyperplasia and/or hypertrophy associated with MSTN variants in cattle (Renand, et al., 2001; Allais, et al., 2009: Albrecht et al., 2006; Lines et al.,

2009). For example, in double muscled cattle, it is believed that the increase in tenderness is because the amount of collagen is reduced and/or because the muscle fibre diameter has changed (Wegner et al., 2000; Bellinge et al., 2005; Warner, et al.,

2010). However, there is another potential pathway by which MSTN can affect tenderness.

In the MSTN variants, the ratio of muscle fibre types altered (Wegner et al., 2000;

Martyn et al., 2004: Haynes et al., 2012). Wegner et al., (2000) showed that more muscle fibre Type IIb was found in double muscled cattle. Type IIb muscle fibres are thought to be more sensitive to the calpain system (Koohmaraie, 1996), and are often related to larger fibre sizes and tenderness (Rosser et al., 1992). There is direct evidence of a relationship between fibre types and tenderness and that is, the higher portion of Type IIb fibres, the more tender the meat (Totland et al., 1988; Muroya et al., 2010). Hence, MSTN might be able to partially affect tenderness through changes in the fibre types. Analyses conducted by AgResearch determined that the ratio between fibre Type IIa and IIb is significantly associated with tenderness in a subset of animals from the sister herd of Jersey-Limousin backcross cattle to those herein

(N. Cullen, unpublished). The MSTN F94L variant significantly changed the ratio between fibre Types IIa and IIb. When the MSTN F94L genotype was fitted into the model, the mean square for the fibre type (the ratio between fibre Types IIa and IIb) decreased and fibre type was longer significant for tenderness (N. Cullen,

238 unpublished). This implies that the effect of fibre type on tenderness could result, at least in part, from the MSTN F94L genotype.

These results do not prove categorically that the effect of MSTN F94L on tenderness is based on the change of fibre types. However, the results do indicate a connection between MSTN F94L, fibre type and tenderness. Unfortunately, the question of whether the calpain system is influenced by the fibre type remains. Due to the insufficient number of samples with fibre type data representing all the genotypes of

CAPN1-SNP316 nested within the MSTN F94L genotype, this question could not be answered herein. Nevertheless, these analyses provide an alternative pathway by which MSTN can affect tenderness through fibre type changes. To clarify the relationship between myostatin, fibre types and the calpain system, more samples with all the possible MSTN and CAPN1 allele combinations need to be analysed for muscle fibre type and tenderness.

7.7. Muscle specific effects on tenderness

One of the more interesting findings from the results herein was that the genetic effects on tenderness appear to be muscle specific for the most part. The example of the CAPN1-SNP316 highlights this observation. The CAPN1-SNP316 not only showed an effect on shear force but also demonstrated differential effects of the genotype on tenderness traits in different muscles (Chapter 5.3.1). In this case, the

GC genotype of the CAPN1-SNP316 had the same effect as the CC genotype of the

CAPN1-SNP316 on the LD muscle. However, in the ST muscle, the effect of the GC genotype of the CAPN1-SNP316 was the same as the GG genotype of the CAPN1-

SNP316. This inconsistency in genotype effects is noteworthy. When selecting the G allele of CAPN1-SNP316 for tender meat, the outcomes in different muscles will be

239 affected by different dominance effects. Moreover, the findings suggest that the effect of the calpain 1 gene in the heterozygous animals varies in different muscles and even in the dominance of opposite alleles. This unusual phenomenon implies that alteration in the calpain protein structure may or may not change its function based on the susceptibility of the muscle. This highlights the fact that the muscle structure or content can affect the activity of calpain. Based on studies of LD and ST muscles (Vestergaard et al., 2000), there are some differences in the percentages of fibre types, fibre areas and collagen. The differences in muscle content may cause such a dramatic change in the CAPN1-SNP316 heterozygotes through different sensitivities to CAPN1. Therefore, the cause of this phenomenon needs further investigation.

There are other cases of differentially genetic effects between muscles in cattle. One such example is the percentage of Bos indicus genetics. The percentage of Bos indicus genetics is a critical factor affecting the grading results for meat palatability.

The higher the percentage of Bos indicus there is within a breed, the tougher the meat and the lower the meat quality (Stoloeski et al., 2006). However, the percentage of Bos indicus has different effects on the MSA palatability score of different muscles (Shackelford et al., 1995; Thompson 2002), highlighting that the genetic effects on tenderness are muscle specific. In this case, the Bos indicus content effect is opposite to the MSTN genotype effect in that the effect on LD is greater than the effect on ST, whereas the MSTN effect is bigger on the ST muscle than the LD muscle. Bos indicus content affects ageing rate whereas the MSTN genotype affects muscle fibre number and type (Shackelford et al., 1995; Kambadur et al., 1997;

McPherron et al., 1997; O'Connor et al., 1997; Bouley et al., 2005).

240

7.8. Collagen effects on tenderness

Collagen content has been linked to the shear force, compression a chewiness of meat although the effects differ between raw and cooked meat (Dransfield, et al.,

2003; Lepetit, 2007; Christensen, et al., 2011). The stability of collagen is maintained by cross-links, which are produced by LOX and are able to stabilize the structure of the tropocollagen molecules (Bailey, 1984; Lepetit, 2007). However, there was no significant relationship between the variants of LOX and LOXL1 and compression discovered (Chapter 5.3.2.5). Therefore, the cross-links produced by

LOX and LOXL1 may not be important for compression in the ST muscle.

Alternatively, in this experimental population, there may have been no genetic variants in these genes that specifically affect compression. Another observation is that there was no relationship between hydroxyproline content and compression or shear force in the experimental population (Chapter 2), also suggesting that total collagen concentration itself was not a critical parameter for shear force or compression herein as recently observed elsewhere (Ngapo et al., 2002; Dransfield et al., 2003; Riley et al., 2005; Stolowski et al., 2006).

7.9 Epistatic effects on tenderness

Another important finding in this study was the large number of epistatic effects on tenderness that explained a significant proportion of the genetic variation. The epistatic effects on tenderness are not confined to the calpain 1/calpastatin interaction (Casas et al., 2006b; Barendse et al., 2007). The gene interactions revealed other possible physiological pathways for tenderness and the importance of specific genotypes for tenderness. These interactions highlighted the importance of

SNPs which were not necessarily significant on their own (eg. FST-SNP7 on adjusted shear force of LD muscle and LOXL1-SNP1 on LD day 26 shear force).

241

Moreover, in other cases, the interaction effects were similar or greater than the significant individual SNP effects. For example, the effect of the interaction between

SNIP1-SNP3 and CAPN1-SNP316 on adjusted shear force of the ST muscle was greater than the effect of CAPN1-SNP316 alone. Furthermore, the SNIP1-SNP3 interacted epistatically with several genes on the adjusted shear force of the LD muscle. The relationship between SNIP1 and these genes is not clear although a broad mechanism, such as that involving CBP/p300 (described above), could possibly explain how SNIP1 may interact with many genes.

In addition to shear force, there were epistatic effects on other traits, such as muscle weight, hydroxyproline content, compression and ageing rate, observed to different degrees. This leads to three important points. The first point is that as a consequence of such interactions, QTL of moderate effect on tenderness may be overlooked in mapping projects. The epistasis may also help explain why the effects of major genes are not seen in all breeds (eg CAPN1) (Allais et al., 2011) and the results of whole genome scans do not appear to be repeatable across cattle breeds even when the allele frequencies are similar (Bolormaa et al., 2011).

The second point is that the mechanisms controlling these traits are complex. It has been observed in genome wide association studies that most of the genetic variation cannot be readily explained by common variants, the so-called “missing heritability”

(Stranger, et al. 2011). About half of the missing heritability may be due to the small effects of the common variants which have been overlooked (Yang, et al., 2012). It is believed that the remaining missing heritability is due to the effects of rare alleles and incomplete linkage disequilibrium (Marian, 2012; Yang, et al. 2012). Gene- environment interactions, epistasis and epigenetics result in non-additive genetic

242 variation that does not contribute to the heritability estimates. However, these can cause heritability to be overestimated and thus, affect the missing heritability

(Stranger, et al., 2011). The results herein support the view that epistasis may have a larger role in genetic variation than previously assumed (Cordell, 2009). This will confound genome wide association studies and heritability estimates because epistasis is so difficult to detect and analyse (Cordell, 2009).

The last point is that epistatic effects may be large but are not usually captured in breeding programmes. In breeding programmes using marker assisted selection, additive effects of SNPs are estimated for the prediction of the performance of the progeny. The efficiency and accuracy of these breeding programmes may be decreased because of these unknown interactions between SNPs.

An example of epistasis herein is CAPN1-SNP316, an important marker already used for tenderness selection (Page et al., 2002; Page et al., 2004; Casas et al.,

2006a), interacted with SNIP1-SNP3 to affect shear force. The combination of the

TT genotype of SNIP1-SNP3 and GG genotypes of CAPN1-SNP316 increased the adjusted shear force for the LD muscle by 50% more than the other combinations of genotypes for CAPN1-SNP316 (Figures 6.2). The additive effect for T allele of

SNIP1-SNP3 on shear force of LD muscle alone was 0.237 ± 0.084 kgF (Appendix

27) so selection of T allele would not normally be a priority. However, given that the

T allele of SNIP1-SNP3 can decrease the quality of the meat in combination with the

CAPN1 genotypes, selection against this allele is more critical.

The “A” allele of the FST-SNP7 is another such example. The interaction between

FST-SNP7 and CAPN1-SNP530 explained more of the variation in adjusted shear force for the LD muscle than either FST-SNP7 or CAPN1-SNP530 alone. These

243 interaction effects provide evidence of the importance of the epitasis. Thus, it is suggested that in these cases, breeding programmes should be designed with the aim of eliminating these alleles in the population using marker assisted selection with extra weighting given for specific alleles.

As the mechanisms underlying tenderness are unveiled, more epistatic effects are likely to be found. Therefore, more markers will be required for effective marker assisted selection and complex epistatic effects need to be considered. If the epistatic effects are sufficiently large, it may be sensible to fix certain alleles within a breed, thereby bypassing the issue.

The difficulty is that marker assisted selection is rarely used for meat quality traits for which the producer receives no direct benefit or profit. There are some exceptions. For example, Brahman breeders in Australia have been using marker assisted selection to produce better quality meat for export markets

(http://www.mla.com.au/Research-and-development/Final-report-details?projectid =

1478). Nevertheless, in general, marker assisted selection is under-utilised for improving traits such as tenderness. A big part of the reason for this is that it is difficult for beef producers to capture a premium for carcasses that are genetically superior for tenderness in current markets.

Another problem is that may be unrealistic to test large numbers of markers required for accurate estimation in all breeds. Using haplotype information of all important alleles in selection may be a resolution to this problem. The effect of each haplotype can be evaluated in progeny tests of one breed and then confirmed in other breeds.

Theoretically, haplotype information may be able to offer a more comprehensive and less biased prediction (Grapes et al., 2006; Hayes et al., 2007).

244

Importantly, DNA markers can be also used in grading systems (eg Meat Standards

Australia), to provide better prediction of tenderness for a given carcass. In using such systems, epistatic effects can be accounted for when evaluating an individual carcass. The caveat is that the DNA technologies will need to be inexpensive and rapid, and probably would only be worthwhile if many markers are available and can be tested simultaneously.

7.10 Future research

In addition, to confirming the gene associations and interactions with traits, there are four main areas of future research that need attention. Firstly, the relationship between MSTN F94L, CAPN1-SNP316 and fibre types needs to be verified with a sufficient number of animals of different muscle fibre type ratios and genotypes of

CAPN1-SNP316. If there were enough animals segregating for the MSTN F94L and

CAPN1-SNP316 genotypes with muscle fibre type data, the hypothesis that the effect of the MSTN F94L on tenderness is partly through the change of muscle fibre types, which have different sensitivities to calpain , could be confirmed.

Secondly, the role of SNIP1 needs to be clarified. SNIP1 not only showed an individual effect on shear force, but also interacted with many other genes. However, the likely mechanism by which SNIP1 is involved in tenderness is not clear. Further investigation of the downstream targets of SNIP1 using microarrays would clarify potential pathways.

Thirdly, new candidate genes should be genotyped with the Jersey-Limousin population for confirmation of their effects. Of particular interest are those genes that have been recently reported and are located on chromosomes with QTL detected

245 herein. These genes include the homogentisate 1, 2 dioxygenase (HGD) gene (BTA

1) (Zhou et al., 2010) and the myogenic factor 5 (MYF5) gene on BTA 5 (Ujan et al.,

2011a), and the calpain small subunit 1 (CAPNS1) gene on BTA18 (Iwanowska et al., 2011). Another source of candidate genes will be the results from proteomics studies on tenderness that are beginning to be published (Guillemin et al., 2011;

D'Alessandro et al., 2012). Investigating the existence, allele frequencies and effects on tenderness of the candidate gene variants in the sample population herein could help improve the understanding of the mechanisms underlying tenderness.

Finally, examining the potential microRNA sites in the genes affecting tenderness is worthwhile and important. However, this requires more microRNA sequence data specific for cattle. While 506 microRNA genes in human have been confirmed

(Chiang et al., 2010), it has been predicted that there are 721 microRNAs in human

(Griffiths-Jones et al., 2008). Glazov et al., (2009) predicted a similar number (625) of miRNAs in cattle. The number of microRNA target sites is far higher with >

45,000 microRNA target sites estimated in 3’UTRs (Friedman et al., 2009), Since about half of the microRNAs are species specific, much more data must be acquired before accurate prediction of microRNA genes and binding sites can be made.

7.11 Conclusions

In this study, most discovered SNPs had small effects on tenderness related traits although some SNPs had relatively large effects on these traits. These included the

MYO1G-SNP3 which explained 13% of the variation of compression in ST muscle and the CAPN5-SNP11 which explained 9% of the variation of the ageing rate for the LD muscle. Many interactions between the genes with intermediate to large effects were also found. The largest was interaction between SNIP1-SNP3 and

246

CAPN1-SNP316 which explained 9.5% of variation of adjusted share force in ST muscle. These findings may contribute to a better prediction of tenderness. The effects of the known tenderness genes can be adjusted to acquire a more precise estimation by taking into account those variants which have individual effects or those which interact with known tenderness genes. The important variants should be verified by examining their effects on tenderness within other cattle herds, such as the sister mapping herd in New Zealand. If the results from the New Zealand herd are consistent with those in this study, the outcomes in this study will be more convincing.

The interactions observed between genes highlights the difficulty in resolving the mechanisms underlying tenderness from a genetic standpoint. Even though a gene has no individual effect on a trait, the gene may still be able to greatly affect a trait by interacting with other genes. Therefore, the conventional approach to study individual effect of genes is not sufficient to clarify the all pathways related to a given complex trait. However, the patterns of the interactions can delineate possible pathways and provide clues about the mechanisms controlling trait. Hence, if the high density markers are used in such studies, analysing the interactions would be quite helpful in determining the important candidate genes and pathways.

The interaction effects and individual effects provide helpful information for establishing a more accurate marker selection system. For example, the interaction between SNIP1-SNP3 and CAPN1-SNP316 explained ~10% of the variation in the adjusted shear force for the ST muscle. The interaction effect was even larger than the individual effect of CAPN1-SNP316 (~5%). Surprisingly, the SNIP1-SNP3 itself did not affect adjusted shear force of the ST muscle. This demonstrates the

247 importance of finding all the genes which have a large influence by interacting with the major genes used in marker selection systems. This will help reduce the variation in breeding predictions and thus, provide a more consistent beef products.

With regards to specific gene effects herein, the myostatin effects on tenderness do not appear to be dependent on hypertrophy, hyperplasia or collagen effects. Rather,

MSTN F94L appears to be affecting tenderness through changes in the muscle fibre types. However, the change in muscle fibre types did not explain all of the effect of

MSTN F94L on shear force. The MSTN F94L did not interact with other candidate genes on shear force, so no further clues were provided as to the additional pathway(s) by which myostatin may affect tenderness. Therefore, clarifying these other mechanisms still needs more investigation.

248

References

249

Aaslyng, M.D., Bejerholm, C., Ertbjerg, P., Bertram, H.C. and Anderse, H.J. (2003) Cooking loss and juiciness of pork in relation to raw meat quality and cooking procedure. Food Quality and Preference 14:227-288.

Aberle, E. D., Reeves, E. S., Judge, M.D., Hunsley, R.E. and Perry, T.W. (1981) Palatability and muscle characteristics of cattle with controlled weight gain: time on a high energy diet. Journal of Animal Science 52:753-763.

Adams, R.J. and Pollard, T.D. (1986) Propulsion of organelles isolated from Acanthamoeba along actin filaments by myosin-1. Nature 322:754-756.

Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J.D. (1983) Molecular Biology of the Cell, Garland Publishing Inc., New York.

Albrecht, E., Teuscher, F., Ender, K. and Wegner, J. (2006) Growth- and breed- related changes of muscle bundle structure in cattle. Journal of Animal Science 84:2959-2964.

Alexander, L.J., MacNeil, M.D., Geary, T.W., Snelling, W.M., Rule, D.C. and Scanga, J.A. (2007) Quantitative trait loci with additive effects on palatability and fatty acid composition of meat in a Wagyu–Limousin F2 population. Animal Genetics 38:506–513.

Alfnes, F., Rickertsen, K. and Ueland, Ø. (2008) Experimental evidence of risk aversion in consumer markets: the case of beef. Applied Economics 40:3039–49.

Allain, J.C., Le Lous, M., Bazin, S., Bailey, A.J. and Delaunay, A. (1978) Isometric tension developed during heating of collagenous tissues: Relationships with collagen cross-linking. Biochimica et Biophysica Acta 533:147–155.

Allais, S., Journaux, L., Leveziel, H., Payet-Duprat, N., Raynaud, P., Hocquette, J.F., Lepetit, J., Rousset, S., Denoyelle, C., Bernard-Capel, C. and Renand, G. (2011) Effects of polymorphisms in the calpastatin and mu-calpain genes on meat tenderness in 3 French beef breeds. Journal of Animal Science 89:1-11.

Allais, S., Leveziel, H., Payet-Duprat, N., Hocquette, J. F., Lepetit, J., Rousset, S., Denoyelle, C., Bernard-Capel, C., Journaux, L., Bonnot, A. and Renand, G. (2010) The two mutations, Q204X and nt821, of the myostatin gene affect carcass and meat quality in young heterozygous bulls of French beef breeds. Journal of Animal Science 88:446-454.

Anderson, C.A., McRae, A.F. and Visscher, P.M. (2006) A simple linear regression method for QTL linkage analysis with censored observations. Genetics 173:1735-45.

Archile-Contreras, A.C., Mandell, I.B. and Purslow, P.P. (2010) Disparity of dietary effects on collagen characteristics and toughness between two beef muscles. Meat Science 86:491–497.

250

Arthur, J.S., Elce, J.S., Hegadorn, C., Williams, K. and Greer, P.A. (2000) Disruption of the Murine Calpain Small Subunit Gene, Capn4: Calpain Is Essential for Embryonic Development but Not for Cell Growth and Division. Molecular and Cellular Biology 20:4474–4481.

Asghar, A. and Bhatti, A.R. (1987) Endogenous proteolytic enzymes in skeletal muscle: their significance in muscle physiology and during post-mortem ageing event in carcasses. Advances in Food Research 31:343-451.

Ashmore, C.R. and Doerr, L. (1971) Comparative aspects of muscle fiber types in different species. Experimental Neurology 31:408-418.

Aslan, O., Sweeney, T., Mullen, A.M. and Hamill, R.M. (2010) Regulatory polymorphisms in the bovine Ankyrin 1 gene promoter are associated with tenderness and intra-muscular fat content. BMC Genetics 11:111.

Bailey, A.J. (1984) In Recent Advances in the Chemistry of Meat. The Royal Society of Chemistry; London 22-40.

Bailey, A. J., M. B. Enser, M.B., Dransfield, E., Restall, D.J. and Avery N.C. (1982) Muscle and adipose tissue from normal and double-muscled cattle: collagen types, muscle fibre diameter, fat cell size and fatty acid composition and organoleptic properties. In: Muscle Hypertrophy of Genetic Origin and Its Use to Improve Beef Production. Menissier, F. (ed.), Martinus Nijhoff Publishers, The Hague, pp. 178-203.

Bailey, A.J., Paul, R.G. and Knott, L. (1998) Mechanisms of maturation and ageing of collagen. Mechanisms of Ageing and Development 106:1-56.

Bailey, A.J., Restall, D.J., Sims, T.J. and Duance, V.C. (1979) Meat tenderness: Immunofluorescent localisation of the isomorphic forms of collagen in bovine muscles of varying texture. Journal of Science of Food and Agriculture 30:203– 210.

Barbantia, D. and Pasquini, M. (2005) Influence of cooking conditions on cooking loss and tenderness of raw and marinated chicken breast meat. LWT - Food Science and Technology 38:895-90.

Barendse, W. J. (2002) DNA markers for meat tenderness. International patent application PCT/AU02/00122. International patent publication WO 02/064820 A1.

Barendse, W., Harrison, B. E., Bunch, R. J. and Thomas, M. B. (2008) Variation at the calpain 3 gene is associated with meat tenderness in zebu and composite breeds of cattle. BMC Genetics 9:41.

Barendse, W., Harrison, B.E., Hawken, R.J., Ferguson, D.M., Thompson, J.M. Thomas, M. B. and Bunch, R. J. (2007) Epistasis between calpain 1 and its inhibitor calpastatin within breeds of cattle. Genetics 176:2601-2610.

251

Bartel D.P. (2009) MicroRNAs: Target Recognition and Regulatory Functions. Cell 136:215-231 .

Bateman, J. F., Lamande, S.R. and Ramshaw, J.A.M. (1996) Collagen superfamily. In Extracellular Matrix. Ed. W. D. Comper, Harwood Academic Publishers, Amsterdam. 2:22-67.

Beecher, G.R., Cassens, R.G., Hoekstra, W.G. and Briskey, E.J. (1965). Red and white fiber content and associated post-mortem properties of seven porcine muscles. Journal of Food Science 30:969–976.

Bekhit, A.E.D., Devine, C.E, Morton, J.D. and Bickerstaffe, R (2003). Towards unifying meat shear force measurement systems to determine meat tenderness. In: Proceedings 49th International Congress of Meat Science and Technology, Buenos Aires, Brazil, 221-222.

Belew, J.B., Brooks, J.C., McKenna, D.R. and Savell, J.W. (2003) Warner–Bratzler shear evaluations of 40 bovine muscles. Meat Science 64:507–512.

Bellinge, R.H.S., Liberles, D.A., Iaschi, S.P.A., O'Brian, P.A. and Tay, G.K. (2005) Myostatin and its implications on animal breeding: a review. Animal Genetics 36:1-6.

Bendall, J.R., (1967). The elastin content of various muscles of beef animals. Journal of the Science of Food and Agriculture 18: 553-558.

Bendall, J.R. (1973) In “The Structure and Function of Muscle”. Academic Press, New York.

Bhuiyan, M.S.A., Kim, N.K., Cho, Y.M., Yoon, D., Kim, K.S., Jeon, J.T. and Lee, J.H. (2007) Identification of SNPs in MYOD gene family and their associations with carcass traits in cattle. Livestock Science 126:292–297.

Bishop, M.D., Koohmaraie, M., Killefer, J. and Kappes, S. (1993) Rapid communication: Restriction fragment length polymorphisms in the bovine calpastatin gene. Journal of Animal Science 71:2277.

Blanchard, P.J. (1994) The influence of lean tissue growth rate on pork tenderness. Meat Focus International, November issue 457-458.

Boccard, R. (1981) Facts and reﬂections on muscular hypertrophy in cattle: Double muscling or culard. In Developments in Meat Science. R. Lawrie, ed. Elsevier Applied Science, London, UK, and New York, NY.

Boccard, R., Buchter, L., Casteels, M., Cosentino, E., Dransfield, E., Hood, D.E., Joseph, R.L., MacDougall, D.B., Rhodes, D.N., Schon, I., Tinbergen, B.J. and Touraille, C. (1981) Procedures for measuring meat quality characteristics in beef production experiments. Report of a Working Group in the Commission of the European Communities' (CEC) Beef Production Research Programme. Livestock Production Science 8:385-397.

252

Bogdan, M., Ghosh, J.K. and Doerge, R.W. (2004) Modifying the Schwarz Bayesian information criterion to locate multiple interacting quantitative trait loci. Genetics 167:989-999.

Boleman, S.J., Boleman, S.L., Savell, J.W., Miller, R.K., Cross, H.R., Wheeler, T.L., Koohmaraie, M., Shackelford, S.D., Miller, M.F., West, R.L. and Johnson, D.D. (1995) Consumer evaluation of beef of known tenderness levels. In Proceedings 41th International Congress of Meat Science and Technology, San Antonio, Texas, USA. 594–595.

Bolormaa, S., Neto, L.R.P., Zhang, Y.D., Bunch, R.J. Harrison, B.E. Goddard, M.E. and Barendse, W. (2011) A genome-wide association study of meat and carcass traits in Australian cattle. Journal of Animal Science 89:2297-2309.

Bouley, J., Meunier, B., Chambon, C., De Smet, S., Hocquette, J.F. and Picard, B. (2005) Proteomic analysis of bovine skeletal muscle hypertrophy. Proteomics 5:490–500.

Bouton, P.E. and Harris, P.V. (1972) A comparison of some objective methods used to assess meat tenderness. Journal of Food Science 37:539–543.

Bouton, P.E., Harris, P.V., Macfarlane, J.J. and Shorthose, W.R. (1982) Influence of pH on the Warner-Bratzler shear properties of mutton. Meat Science 6:27-36.

Brady, D.E. (1937) A Study of the Factors Influencing Tenderness and Texture of Beef. Journal of Animal Science 20:246-250.

Brameld, J.M., Buttery, P.J., Dawson J.M. and Harper J.M.M. (1998) Nutritional and hormonal control of skeletal-muscle cell growth and differentiation. Proceedings of the Nutrition Society 57: 207-217.

Bratcher, C.L., Johnson, D.D., Littell, R.C. and Gwartney, B.L. (2005) The effects of quality grade, aging, and location within muscle on Warner–Bratzler shear force in beef muscles of locomotion. Meat Science 70:279–284.

Braun, C., Engel, M., Theisinger, B., Welter, C. and Seifert, M. (1999) CAPN 8: isolation of a new mouse calpain-isoenzyme. Biochemical and Biophysical Research Communications 60:671-675.

Brockmann, G.A., Kratzsch, J., Haley, C.S., Renne, U., Schwerin, M. and Karle, S. (2000) Single QTL effects, epistasis, and pleiotropy account for two-thirds of the phenotypic F2 variance of growth and obesity in DU6i × DBA/2 mice. Genome Research 10:1941–1957.

Brooke, M.H. and Kaiser, K.K. (1970) Muscle fiber types: How many and what kind? Archives of Neurology 23:369–379.

Brooks, J.C. and Savell, J.W. (2004) Perimysium thickness as an indicator of beef tenderness. Meat Science 67:329–334.

253

Bryson-Richardson, R.J. and Currie, P.D. (2003) The genetics of vertebrate myogenesis. Nature Genetics 9: 632-646.

Bunter, K., Hermesch, S., Luxford, B.G., Graser, H.U. and Crump, R.E. (2005) Insulin-like growth factor-I measured in juvenile pigs is genetically correlated with economically important performance traits. Australian Journal of Experimental Agriculture 45:783–792.

Burrow, H. M., Moore, S.S., Johnston, D. J., Barendse, W. and Bindon, B.M. (2001) Quantitative and molecular genetic influences on properties of beef: A review. Australian Journal of Experimental Agriculture 41: 893-919.

Burson, D.E., and Hunt, M.C. (1986) Heat-induced changes in the proportion of types I and III collagen in bovine longissimus dorsi. Meat Science 17:153–160.

Cafe, L.M., McIntyre, B.L., Robinson, D.L., Geesink, G.H., Barendse, W., Pethick, D.W., Thompson, J.M. and Greenwood, P.L. (2010) Production and processing studies on calpain-system gene markers for tenderness in Brahman cattle: 2. Objective meat quality. Journal of Animal Science 88:3059-3069.

Caine, W.R., Aalhus, J.L., Best, D.R., Dugan, M.E.R., and Jeremiah, L.E. (2003) Relationship of texture profile analysis and Warner–Bratzler shear force with sensory characteristics of beef rib steaks. Meat Science 64:333–339.

Calkins, C.R. and Seidman, S.C. (1988) Relationship among calcium-dependent protease, cathepsin B and H, meat tenderness and the response of muscle to aging. Journal of Animal Science 66:1186–1193.

Carango, P., Noble, J.E., Marks, H.G. and Funanage, V.L. (1993) Absence of myotonic dystrophy protein kinase (DMPK) mRNA as a result of a triplet repeat expansion in myotonic dystrophy. Genomics 18:340-348.

Carlborg, O. (2006) Detection of epistatic QTL. Proc. 8th World Congress Genet. Appl. Livest. Prod., Belo Horizonte, Brazil. CD-ROM Commun. No. 20-01.

Carlborg, O. and Haley, C.S. (2004) Epistasis: Too often neglected in complex trait studies? Nature Reviews Genetics 5:618–625.

Carlborg, O., Jacobsson, L., Ahgren, P., Siegel, P. and Andersson, L. (2006) Epistasis and the release of genetic variation during long-term selection. Nature Genetics 38:418–420.

Carlborg, Ö., Kerje, S., Schütz, K., Jacobsson, L., Jensen, P. and Andersson, L. (2003) A global search reveals epistatic interaction between QTLs for early growth in the chicken. Genome Research 13:413–421.

Carlsson, E., Poulsen, P., Storgaard, H., Almgren, P., Ling, C., Jensen, C.B., Madsbad, S., Groop, L., Vaag, A. and Ridderstråle, M. (2005) Genetic and nongenetic regulation of CAPN10 mRNA expression in skeletal muscle. Diabetes 54:3015-3020.

254

Carpenter, C.E., Rico O.D., Cockett, N.E. and Snowder, G.D. (1996) Histology and composition of muscles from normal and callipyge lambs. Journal of Animal Science 4:388-393.

Casas, E., Shackelford, S. D., Keele, J. W., Koohmaraie, M., Smith, T. P. L. and Stone, R.T. (2003) Detection of quantitative trait loci for growth and carcass composition in cattle. Journal of Animal Science 81:2976–2983.

Casas, E., Shackelford, S.D., Keele, J. W., Stone, R.T., Kappes, S.M. and Koohmaraie, M. (2000) Quantitative trait loci affecting growth and carcass composition of cattle segregating alternate forms of myostatin. Journal of Animal Science 78: 560-569.

Casas, E., Stone, R. T., Keele, J. W., Shackelford, S. D., Kappes, S. M. and Koohmaraie, M. (2001) A comprehensive search for quantitative trait loci affecting growth and carcass composition of cattle segregating alternative forms of the myostatin gene. Journal of Animal Science 79:854-860.

Casas, E., White, S.N., Riley, D.G., Smith, T.P.L., Brenneman, R.A., Olson, T.A., Johnson, D.D., Coleman, S.W., Bennett, G.L. and Chase Jr., C.C. (2005) Assessment of single nucleotide polymorphisms in genes residing on chromosomes 14 and 29 for association with carcass composition traits in Bos indicus cattle. Journal of Animal Science 83:13−19.

Casas, E., White, S.N., Wheeler, T.L., Shackelford, S.D., Koohmaraie, M., Riley, D.G., Chase Jr., C.C. Johnson, D.D. and Smith, T.P.L. (2006a) Effects of calpastatin and μ-calpain markers in beef cattle on tenderness traits. Journal Animal Science 84:520-525.

Casas, E., White, S., Wheeler, T., Shackelford, S., Koohmaraie, M., Riley, D., Chase, C. Jr., Johnson, D. and Smith, T. (2006b) Testing for interactions between calpastatin and m-calpain markers in beef cattle on tenderness traits. Journal of Animal Science 84:135-136.

Charge, S.B. and Rudnicki M.A. (2004) Cellular and molecular regulation of muscle regeneration. Physiological Reviews 84: 209–238.

Chen, R., Davydov, E.V., Sirota, M. and Butte, A.J. (2010) Non-synonymous and synonymous coding SNPs show similar likelihood and effect size of human disease association. PLoS ONE 5: e13574.

Chiang, H.R., Schoenfeld, L.W., Ruby, J.G., Auyeung, V.C., Spies, N., Baek, D., Johnston, W.K., Russ, C., Luo, S., Babiarz, J.E., Blelloch, R., Schroth, G.P., Nusbaum, C. and Bartel, D.P. (2010) Mammalian microRNAs: experimental evaluation of novel and previously annotated genes. Genes & Development 24:992-1009.

255

Childers, C.P., Reese, J.T., Sundaram, J.P., Vile, D.C., Dickens, C.M., Childs, K.L., Salih, H., Bennett, A.K., Hagen, D.E. Adelson, D.L. and Elsik, C.G. (2010) Bovine Genome Database: integrated tools for genome annotation and discovery. Nucleic Acids Research 39(suppl. 1):D830-D834.

Choi, B.H., Lee, Y.-M., Alam, M., Lee, J.-H., Kim, T.H., Kim, K.-S. and Kim, J.-J. (2011) Detection of Mendelian and parent-of-origin quantitative trait loci for meat quality in a cross between Korean Native Pig and Landrace. Australasian Journal of Animal Sciences 24:1644-1650.

Christensen, M., Ertbjerg, P., Failla, S., Sañudo, C., Richardson, R.I., Nute, G.R., Olleta, J.L., Panea, B., Albertí, P., Juárez, M., Hocquette, J.F. and Williams, J.L. (2011) Relationship between collagen characteristics, lipid content and raw and cooked texture of meat from young bulls of fifteen European breeds. Meat Science 87:61–65.

Christensen, M., Kok, C. and Ertbjerg, P. (2006) Mechanical properties of type I and type IIB single porcine muscle fibres. Meat Science 73:422–425.

Chrystall, B.B. and Devine, C.E. (1978) Electrical stimulation, muscle tension and glycolysis in bovine sternomandibularis. Meat Science 2:49–58.

Churchill, G.A. and Doerge, R.W. (1994) Empirical Threshold Values for Quantitative Trait Mapping. Genetics 138:936-971.

Ciobanu, D.C. Bastiaansen, J.W.M., Lonergan, S.M., Thomsen, H., Dekkers, J.C.M., Plastow, G.S. and Rothschild, M.F. (2004) New alleles in calpastatin gene are associated with meat quality traits in pigs. Journal of Animal Science 82:2829- 2839.

Claeys, E., Uytterhaegen, L., Demeyer, D. and De Smet, S. (1994) Beef myofibrillar protein salt solubility in relation to tenderness and proteolysis. Proceedings of 40th International Congress of Meat Science and Technology (S-IVB. 09). The Hague, The Netherlands. 40, SIVB.09.

Clop, A., Marcq, F., Takeda, H., Pirottin, D., Tordoir, X., Bibé, B., Bouix, J., Caiment, F., Elsen, J. M., Eychenne, F., Larzul, C., Laville, E., Meish, F., Milenkovic, D., Tobin, J., Charlier, C. and Georges, M. (2006) A mutation creating a potential illegitimate microRNA target site in the myostatin gene affects muscularity in sheep. Nature Genetics 38: 813–818.

Colgravea, M.L., Allinghama, P.G. and Jones, A. (2008) Hydroxyproline quantification for the estimation of collagen in tissue using multiple reaction monitoring mass spectrometry. Journal of Chromatography A 1212:150-153.

Cooper, C.C., Breidenstein, B.B., Cassens, R.G., Evans, G. and Bray, R.W. (1968) Influence of marbling and maturity on the palatability of beef muscle II. Histological considerations. Journal of Animal Science 27:1542-1546.

256

Cooper, D.N., Bacolla, A., F´erec, C., Vasquez, K.M. Kehrer-Sawatzki, H. and Chen, J.M. (2011) On the Sequence-Directed Nature of Human Gene Mutation: The Role of Genomic Architecture and the Local DNA Sequence Environment in Mediating Gene Mutations Underlying Human Inherited Disease. Human Mutation 32:1075–1099.

Cordell, H.J. (2009) Detecting gene-gene interactions that underlie human diseases. Nature Reviews Genetics 10: 392–404.

Cossu, G. and Biressi S. (2005) Satellite cells, myoblasts and other occasional myogenic progenitors: Possible origin, phenotypic features and role in muscle regeneration. Seminars in Cell and Developmental Biology 16:623–631.

Cross, H.R., Carpenter, Z.L. and Smith, G.C. (1973) Effects of intramuscular collagen and elastin on bovine muscle tenderness. Journal of Food Science 38: 998–1003.

Crouse, J.D., Koohmaraie, M. and Seideman, S.D. (1991) The relationship of muscle fibre size to tenderness of beef. Meat Science 30:295-302.

Crouse, J.D., Smith, G.M. and Mandigo, R.W. (1978) Relationship of selected beef carcass traits with meat palatability. Journal of Food Science 43:152–157.

Culler, R.D., Parrish J.R., Smith, G.C. and Cross, H.R. (1978) Relationship of myofibril fragmentation index to certain chemical physical and sensory characteristics of bovine longissimus muscle. Journal of Food Science 43:177- 1180.

Curi, R.A., Chardulo, L.A.L., Giusti, J., Silveira, A.C., Martins, C.L. and de Oliveira, H.N. (2010) Assessment of GH1, CAPN1 and CAST polymorphisms as markers of carcass and meat traits in Bos indicus and Bos taurus–Bos indicus cross beef cattle. Meat Science 86:915-920.

Curi, R.A., Chardulo, L.A.L., Mason, M.C., Arrigoni, M.D.B., Silveira, A.C. and de Oliveira, H.N. (2009) Effect of single nucleotide polymorphisms of CAPN1 and CAST genes on meat traits in Nellore beef cattle (Bos indicus) and in their crosses with Bos taurus. Animal Genetics 40:456–462.

D'Alessandroa, A., Rinalduccia S., Marroccoa C., Zollaa V., Napolitanob F., and Zollaa L. (2012) Love me tender: An Omics window on the bovine meat tenderness network. Journal of Proteomics doi:10.1016/j.jprot.2012.02.013.

Darvasi, A., Weinreb, A., Minke, V., Weller, J.I. and Soller, M. (1993) Detecting Marker-QTL Linkage and Estimating QTL Gene Effect and Map Location Using a Saturated Genetic Map. Genetics 134:944-951.

Davis, G.P., Moore, S.S., Drinkwater, R.D., Shorthose, W.R., Loxton, I.D., Barendse, W. and Hetzel, D.J.S. (2007) QTL for meat tenderness in the M. longissimus lumborum of cattle. Animal Genetics 39:40-45.

257

Davis, M.E. and Simmen, R.C.M. (2006) Genetic parameter estimates for serum insulin-like growth factor I concentrations, and body weight and weight gains in Angus beef cattle divergently selected for serum insulin-like growth factor I concentration. Journal of Animal Science 84:2299–2308.

Davis, T. L., Walker, J. R., Finerty, P. J. Jr, Mackenzie, F., Newman, E. M. and DePaganon, S. (2007) The crystal structures of human calpains 1 and 9 imply diverse mechanisms of action and auto-inhibition. Journal of Molecular Biology 366:216-229. de Roos, A.P.W., Hayes, B.J., Spelman, R.J. and Goddard, M.E. (2008) Linkage Disequilibrium and Persistence of Phase in Holstein-Friesian, Jersey and Angus Cattle. Genetics 179:1503–1512.

Dear, T.N. and Boehm, T. (1999) Diverse mRNA expression patterns of the mouse calpain genes Capn5, Capn6 and Capn11 during development. Mechanisms of Development 89:201-209.

Dear, T.N., Meier, N.T., Hunn, M. and Boehm, T. (2000) Gene structure, chromosomal localization, and expression pattern of Capn12, a new member of the calpain large subunit gene family. Genomics 68:152-160.

Deng, H.W., Chen, W.M. and Recker, R.R. (2000) QTL fine mapping by measuring and testing for Hardy-Weinberg and linkage disequilibrium at a series of linked marker loci in extreme samples of populations. American Journal of Human Genetics 66:1027-1045.

Destefanis, G., Brugiapaglia, A., Barge, M.T. and Dal Molin, E. (2008) Relationship between beef consumer tenderness perception and Warner–Bratzler shear force. Meat Science 78:153–156.

Devine, C.E., Payne, S.R., Peachey, B.M., Lowe, T.E., Ingram, J.R. and Cook, C.J. (2002) High and low rigour temperature effects on sheep meat tenderness and ageing. Meat Science 60:141-146.

Devine, C.E., Wahlgren, N.M. and Tornberg, E. (1999) Effect of rigour temperature on muscle shortening and tenderisation of restrained and unrestrained beef M. longissimus thoracius et lumborum. Meat Science 51:61– 72.

Dikeman, M.E. (1987) Fat reduction in animals and the effects on palatability and consumer acceptance of meat products. Reciprocal Meat Conference 40:93-103.

Dikeman, M. E. (2007). Effects of metabolic modifiers on carcass traits and meat quality. Meat Science 77:121−135.

Dominique, J.-E. and Gérard C. (2006) Myostatin regulation of muscle development: Molecular basis, natural mutations, physiopathological aspects. Experimental Cell Research 312: 2401-2414.

258

Dransfield, E., Martin, J.F., Bauchart, D., Abouelkaram, S., Lepetit, J., Culioli, J., Jurie, C. and Picard, B. (2003). Meat quality and composition of three muscles from French cull cows and young bulls. Animal Science 76:387–399.

Dreyer, J.H., Naude, R.T., Henning, J.W.N. and Roussouw, E. (1977) The influence of breed, castration and age on muscle fibre type and diameter in Friesland and Afrikaner cattle. South African Journal of Animal Science 7:171-180.

Drinkwater, R.D., Li, Y., Lenane, I., Davis, G. P., Shorthose, R.P., Harrison, B.E., Richardson, K., Ferguson, D., Stevenson, R., Renaud, J., Loxton, I., Hawken, R.J., Thomas, M.B., Newman, S., Hetzel, D.J.S. and Barendse, W. (2006) Detecting quantitative trait loci affecting beef tenderness on bovine chromosome 7 near calpastatin and lysyl oxidase. Australian Journal of Experimental Agriculture 46:159–164.

Dutaud, D., Aubry, L., Sentandreu, M.A. and Ouali, A. (2006) Bovine muscle 20S proteasome: I. Simple purification procedure and enzymatic characterization in relation with postmortem conditions. Meat Science 74:327-336.

Duthie, C., Simm, G., Doeschl-Wilson, A., Kalm, E., Knap, P.W. and Roehe, R. (2010) Epistatic analysis of carcass characteristics in pigs reveals genomic interactions between quantitative trait loci attributable to additive and dominance genetic effects. Journal of Animal Science 88:2219-2234.

Dutson, T.R, Smith, G.C. and Carpenter, Z.L. (1980) Lysosomal enzyme distribution in electrically stimulated ovine muscle. Journal of Food Science 45:1097–1098.

Earnshaw, W. C., Martins, L. M., and Kaufmann, S. H. (1999) Mammalian caspases: Structure, activation, substrates, and functions during apoptosis. Annual Review of Biochemistry 68:383–424.

Edwards, D.B., Ernst, C.W., Raney, N.E., Doumit, M.E., Hoge, M.D. and Bates, R.O. (2008) Quantitative trait locus mapping in an F-2 Duroc x Pietrain resource population: II. Carcass and meat quality traits. Journal of Animal Science 86(2):254-266.

Erman, B. and Mark, J.E. (1997) Bioelastomers. In “Structures and properties of rubberlike networks” New York, Oxford: Oxford University Press.

Esmailizadeh, A.K. (2006) Multiple trait analysis for genetic mapping of quantitative trait loci for carcass and beef quality. Thesis submitted for Doctor of Philosophy at the University of Adelaide.

Esmailizadeh, A.K., Bottema, C.D.K., Sellick, G.S., Verbyla, A.P., Morris, C.A., Cullen, N.G. and Pitchford W.S. (2008) Effects of the myostatin F94L substitution on beef traits. Journal of Animal Science 86:1038-1046.

259

Esmailizadeh, A.K., Morris, C.A., Cullen, N.G. Kruk, Z.A., Lines, D.S., Hickey, P.M., Dobbie, P.M., Bottema, C.D.K. and Pitchford W.S. (2011) Genetic mapping of quantitative trait loci for meat quality and muscle metabolic traits in cattle. Journal of Animal Genetics 42:592-599.

Estelle, J., Gil, F., Vazquez, J.M., Latorre, R., Ramirez, G., Barragan, M.C., Folch, J.M., Noguera, J.L., Toro, M.A. and Perez-Enciso, M. (2008) A quantitative trait locus genome scan for porcine muscle fiber traits reveals overdominance and epistasis. Journal of Animal Science 86:3290-3299.

Eyre, D.R., Wu, J.J. and Woods, P. (1992) Cartilage-specific collagens: Structural studies. In Kuettner KE, Schleyrenbach R, Peyron JG, Hascall VC (eds): ‘‘Articular Cartilage and Osteoarthritis’’. New York, Raven Press, p119–131.

Fan, R. and Xiong, M. (2002) High resolution mapping of quantitative trait loci by linkage disequilibrium analysis. European Journal of Human Genetics 10:607- 615.

Fan, Y.Y., Zan, L.S., Fu, C.Z., Tian, W.Q., Wang, H.B., Liu, Y.Y. and Xin, Y.P. (2011) Three novel SNPs in the coding region of PPARc gene and their associations with meat quality traits in cattle. Molecular Biology Reports 38:131–137.

Fang, S.H., Nishimura, T. and Takahashi, K. (1999) Relationship between development of intramuscular connective tissue and toughness of pork during growth of pigs. Journal of Animal Science 77:120–130.

Farnir, F., Coppieters, W., Arranz, J.J., Berzi, P., Cambisano, N., Grisart, B., Karim, L., Marcq, F., Moreau, L., Mni, M., Nezer, C., Simon, P., Vanmanshoven, P., Wagenaar, D. and Georges, M. (2000) Extensive genome-wide linkage disequilibrium in cattle. Genome Research 10:220-227.

Ferguson, D.M., Jiang, S., Hearnshaw, H., Rymill, S.R. and Thompson, J.M. (2000) Effect of electrical stimulation on protease activity and tenderness of M. longissimus from cattle with different proportions of Bos indicus content. Meat Science 55:265–272.

Fernandez, X., Lefaucheur, L. and Candek, M. (1995) Comparative study of two classifications of muscle fibres: Consequences for the photometric determination of glycogen according to fibre type in red and white muscle of pig. Meat Science 41:225-235.

Findlay, C.J., Stanley, D.W. and Gellett, E.A. (1986) Thermomechanical properties of beef muscle. Meat Science 16:57-70.

Finsterer, J. (2002) Myotonic dystrophy type 2. European Journal of Neurology 9: 441–447.

Franz, T., Vingron, M., Boehm, T. and Dear, T. N. (1999) Capn7: a highly divergent vertebrate calpain with a novel C-terminal domain. Mammalian Genome 10(3):318-321.

260

Friedman, R.C., Farh, K.K-H., Burge, C.B. and Bartel, D.P. (2009) Most mammalian mRNAs are conserved targets of microRNAs. Genome Research 19:92-105.

Frylinck, L., van Wyka, G.L., Smith, T.P.L., Strydom, P.E., van Marle-Köster, E., Webb, E. C., Koohmaraie, M. and Smith, M. F. (2009) Evaluation of biochemical parameters and genetic markers for association with meat tenderness in South African feedlot cattle. Meat Science 83:657-665.

Funkenstein, B., Rebhan, Y. and Skopal, T. (2009) Molecular cloning and characterization of follistatin in the gilthead sea bream, Sparus aurata. Molecular Biology Reports 36:510-511.

Gariepy, C., Jones, S.D.M. and Robertson, W.M. (1990) Variation in meat quality at three sites along the length of the beef longissimus muscle. Canadian Journal of Animal Science 70:707-710.

Gault, N.F.S. (1992) In “The Chemistry of Muscle-Based Foods” Ed. by Johnston, D.E., Knight, M.K. and Ledward, D.A. The Royal Society of Chemistry, Cambridge.

Gauthier, G.F. (1969) On the relationship of ultrastructural and cytochemical features to color in mammalian skeletal muscle. Zeitschrift fur Mikroskopisch- Anatomische Forschung 95:462-482.

Geesink, G. H. and Koohmaraie, M. (1999) Effect of calpastatin on degradation of myofibrillar proteins by mu-calpain under mortem conditions. Journal of Animal Science 77:2685-2692.

Geesink, G.H., Kuchay, S., Chishti, A.H. and Koohmaraie, M (2006) mu-Calpain is essential for postmortem proteolysis of muscle proteins. Journal of Animal Science 84:2834-2840.

Gerrard, D.E., Jones, S.J., Aberle, E.D., Lemenager, R.P., Diekman, M.A. and Judge, M.D. (1987) Collagen, testosterone and meat tenderness in bulls. Journal of Animal Science 65:1236-1242.

Gilmour, A.R., Gogel, B.J., Cullis, B.R. and Thompson, R. (2006) ASReml User Guide Release 2.0. VSN International Ltd, Hemel Hempstead, HP1 1ES, UK.

Glazov, E.A., Kongsuwan, K., Assavalapsakul, W., Horwood, P.F., Mitter, N. and Mahony, T.J. (2009) Repertoire of Bovine miRNA and miRNA-Like Small Regulatory RNAs Expressed upon Viral Infection. PLoS ONE 4: e6349.

Goll, D.E. (1991) Role of proteinases and protein turnover in muscle growth and meat quality. Recip. Meat Conference (American Meat Science Association), 44:25-36.

Goll, D.E., Geesink, G.H., Taylor, R.G. and Thompson, V.F. (1995) Proc. Int. Cong. Meat Science and Technology 41:537.

261

Goll, D.E., Kleese, W.C., Okitant, A. Kumamoto, T. Cong, J. and Kapprell, H.P. (1990) Historical background and current status of the Ca2-dependent proteinase system. In: Intracellular Calcium-Dependent Proteolysis edited by Mellgren RL and Murachi T. Boca.

Goll, D.E., Kleese, W.C. and Szpacenko, A. (1989) Skeletal muscle proteases and protein turnover. In “Animal growth regulation” (Eds OR Campion, GJ Hausman, RJ Martin) Pleum Press, New York.

Goll, D.E., Neti, G., Mares, S.W. and Thompson, V.F. (2008) Myofibrillar protein turnover: The proteasome and the calpains. Journal of Animal Science 86:E19- E35.

Goll, D.E., Otuka, Y., Nagainis, P.A., Shannon, J.D., Sathe, S.K. and Muguruma, M. (1983) Role of muscle proteinases in maintenance of muscle integrity and mass. Journal of Food Biochemistry 7:137-145.

Goll, D.E., Robson R.M., and Stromer, M.H. (1984) Skeletal muscle. In Duke’s physiology of domestic animals (10th edition). Swensen, M.J. (ed.), Cornell University Press, Ithaca, New York, pp. 548–580.

Goll, D.E., Thompson, V.F., Li, H., Wei, W. and Cong, J. (2003) The Calpain System. Physiological Reviews 83: 731-801.

Goll, D.E., Thompson, V.F., Taylor, R.G. and Ouali, A. (1998) The calpain system and skeletal muscle growth. Canadian Journal of Animal Science 78:503–512.

Goodman, R. H. and Smolik. S. (2000) CBP/p300 in cell growth, transformation, and development. Genes and Development 14:1553-1577.

Gorlov, I.P., Gorlova, O.Y., Sunyaev, S.R., Spitz, M.R. and Amos, C.I. (2008) Shifting paradigm of association studies: value of rare single-nucleotide polymorphisms. American Journal of Human Genetics 82: 100-112.

Gorza, L. (1990) Identification of a novel type 2 fibre population in mammalian skeletal muscle by combined use of histochemical myosin ATPase and anti- myosin monoclonal antibodies. Journal of Histochemistry and Cytochemistry 38:257-265.

Grapes, L., Firat, M.Z., Dekkers, J.C.M., Rothschild, M.F. and Fernando, R.L. (2006) Optimal Haplotype Structure for Linkage Disequilibrium-Based Fine Mapping of Quantitative Trait Loci Using Identity by Descent. Genetics 172:1955–1965.

Griffin, T.A., Slack, J.P., McCluskey, T.S., Monaco, J.J. and Colbert, R.A. (2000) Identification of proteassemblin, a mammalian homologue of the yeast protein, ump1p, that is required for normal proteasome assembly. Molecular Cell Biology Research Communications 3:212-217.

262

Griffiths-Jones, S., Saini, H.K., Dongen, S.V. and Enright, A.J. (2008) miRBase: tools for microRNA genomics. Nucleic Acids Res. 36: D154–D158.

Grobet, L., Martin, L.J.R., Poncelet, D., Pirottin, D., Brouwers, B., Riquet, J., Schoeberlein, A., Dunner, S., Menissier, F., Massabanda, J., Fries, R., Hanset, R. and Georges, M. (1997) A deletion in the bovine myostatin gene causes the double-muscled phenotype in cattle. Nature Genetics 17:71-74.

Gruber, S. L., Tatum, J.D., Scanga, J.A., Chapman, P.L., Smith G.C. and Belk K.E. (2006) Effects of postmortem ageing and USDA quality grade on Warner- Bratzler shear force values of seventeen individual beef muscles. Journal of Animal Science 84:3387-3396.

Gruendler, C., Lin, Y., Farley, J. and Wang, T. (2001) Proteasomal degradation of Smad1 induced by bone morphogenetic proteins. Journal of Biological Chemistry 276:46533-46543.

Guillemin N., Bonnet, M, Jurie C. and Picard, B. (2011) Functional analysis of beef tenderness. Journal of Proteomics 75: 352-365.

Guo, S., Lud, J., Schlanger, R., Zhange, H., Wange, J.Y., Fox, M.C., Purton, L.E., Fleming, H.H., Cobb, B., Merkenschlager, M., Golub, T.R. and Scadden, D.T. (2010) MicroRNA miR-125a controls hematopoietic stem cell number. PNAS 107:14229-14234.

Gutie´rrez-Gil, B., Wiener, P., Nute, G.R., Burton, D., Gill, J.L., Wood, J.D. and Williams, J. (2007) Detection of quantitative trait loci for meat quality traits in cattle. Animal Genetics 39:51–61.

Hailstones, D., Barton, P., Chan-Thomas, P., Sasse, S., Sutherland, C. Hardeman, E. and Gunning, P. (1992) Differential regulation of the atrial isoforms of the myosin light chains during striated muscle development. Journal of Biological Chemistry 267:23295-23300.

Haley, C.S. and Knott S.A. (1992) A simple regression method for mapping quantitative trait loci in line crosses using flanking markers. Heredity 69:315- 324.

Hammell, M., Long, D., Zhang, L., Lee, A., Carmack, C.S., Han, M., Ding, Y. and Ambros, V. (2008) mirWIP: microRNA target prediction based on microRNA- containing ribonucleoprotein-enriched transcripts. Nature Methods 9:813–819.

Harmegnies, N., Davin, F., De Smet, S., Buys, N., Georges, M. and Coppieters, W. (2006) Results of a whole-genome quantitative trait locus scan for growth, carcass composition and meat quality in a porcine four-way cross. Animal Genetics 37:543-553.

Harper, G.S. (1999) Trends in skeletal muscle biology and the understanding of toughness in beef. Australian Journal of Agricultural Research 50:1105–1129.

263

Harrell, R.A., Bidner, T.D. and Icaza, E.A. (1978) Effect of Altered Muscle pH on Beef Tenderness. Journal of Animal Science 46:1592-1596.

Hayes, B.J., Chamberlain, A.J., McPartlan, H., Macleod, I., Sethuraman, L. and Goddard, M.E. (2007) Accuracy of marker-assisted selection with single markers and marker haplotypes in cattle. Genetical Research 89:215-220.

Haynes, F.E.M., Greenwood, P.L., McDonagh, M.B. and Oddy, V.H. (2012) Myostatin allelic status interacts with level of nutrition to affect growth, composition, and myofiber characteristics of lambs. Journal of Animal Science 90(2):456-465.

Herring, H.K., Cassens, R.G. and Briskey, E.J. (1965) Further studies on bovine muscle tenderness as influenced by carcass position, sarcomere length and fiber diameter. Journal of Food Science 30:1049-1054.

Hertzman, C., Olsson, U. and Tornberg, E. (1993) The influence of high temperature, type of muscle and electrical stimulation on the course of rigour, ageing and tenderness of beef muscles. Meat Science 35:119-141.

Hess, A. and Rosner, S., (1970) The satellite cell bud and myoblast in denervated mammalian muscle fibers. American Journal of Anatomy 1:21-39.

Hill, F. (1966) The solubility of intramuscular collagen in meat animals of various ages. Journal of Food Science 31:161-166.

Hiner, R. L., Hanxins, O.G., Sloane, H.S., Fellers, R.L. and Anderson E.E. (1953) Fiber diameter in relation to tenderness of beef muscle. Food Research 18:364- 376.

Hocquette, J.-F., Renand, G., Leveziel, H. Picard, B. and Cassar-Malek, I. (2006) The potential benefits of genetics and genomics to improve beef quality- a review. Animal Science Papers and Reports 24:173-189.

Honikel, K.O., Hamid, A., Fischer, C. and Hamm, R. (1981) Influence of postmortem changes in bovine muscle on the water-holding capacity of beef post-mortem storage of muscle at various temperatures between 0 and 30℃. Journal of Food Science 46:23-24.

Hope-Jones, M., Strydom, P.E., Frylinck, L. and Webb, E.C. (2010) The efficiency of electrical stimulation to counteract the negative effects of β-agonists. Meat Science 86:699–705.

Horgan, D.J., Jones, P.N., King, N.L., Kurth, L.B. and Kuypers, R. (1991) The relationship between animal age and the thermal stability and cross-link content of collagen from five goat muscles. Meat Science 29:251–262.

Horgan, D.J., King, N.L., Kurth, L.B. and Kuypers, R. (1990) Collagen crosslinks and their relationship to the thermal properties of calf tendons. Archives of Biochemistry and Biophysics 281:21–26.

264

Hosteler, R.L., Link, B.A., Landmann, W.A. and Fitzhugh, H.A. Jr. (1972) Effect of carcass suspension on sarcomere length and shear force of some major bovine muscles. Journal of Food Science 37:132-135.

Huang, Y. and Wang, K. (2001) The calpain family and human disease. Trends in Molecular Medicine 7:355–362.

Huff-Lonergan, E., Lonergan, S. T. and Vaske, L. (2000) pH relationships to quality attributes: tenderness. Proc. symposia: pH and pork quality. In Proceedings reciprocal meat conference 53:1–4.

Hunter, R.A., Burrow, H.M. and McCrabb, G.J. (2001b) Sustained growth promotion, carcass and meat quality of steers slaughtered at three liveweights. Australian Journal of Experimental Agriculture 41:1033-1040.

Hunter, R.A., McCrabb, G.J. and O’Neill, C.J. (2001a) The influence of hormonal growth promotants on marbling. Proceedings CRC Marbling Symposium pp.129-133.

Hwang, I.H., Park, B.Y., Cho, S.H. and Lee, J.M. (2004) Effects of muscle shortening and proteolysis on Warner–Bratzler shear force in beef longissimus and semitendinosus. Meat Science 68:497–505.

Hwang, I.H. and Thompson, J.M. (2001a). The eﬀect of time and type of electrical stimulation on the calpain system and meat tenderness in beef longissimus dorsi muscle. Meat Science 58:135–144.

Hwang, I.H. and Thompson, J.M. (2001b). The interaction between pH and temperature decline early postmortem on the calpain system and objective tenderness in electrically stimulated beef longissimus dorsi muscle. Meat Science 58:167–174.

Ilian, M.A., Bickerstaffe, R. and Greaser, M.L. (2003) Postmortem changes in myofibrillar-bound calpain 3 revealed by immunofluorescence microscopy. Meat Science 66:231–240.

Isobe, S., Nakaya, A. and Tabata, S. (2007) Genotype Matrix Mapping: Searching for Quantitative Trait Loci Interactions in Genetic Variation in Complex Trait. DNA Research 14(5):217-225.

Issanchou, S. (1996) Consumer expectations and perceptions of meat and meat products quality. Meat Science 43:S5–S19.

Iwanowska, A., Grzes, B., Mikolajczak, B., Iwanska, E., Juszczuk-Kubiak, E., Rosochacki, S.J. and Pospiech, E. (2011) Impact of polymorphism of the regulatory subunit of the mu-calpain (CAPN1S) on the proteolysis process and meat tenderness of young cattle. Molecular Biology Reports 38:1295-1300.

265

Jansen, R.C. (1994) Controlling the Type I and Type I1 Errors in Mapping Quantitative Trait Loci. Genetics 138:871-881.

Jeremiah, L.E. and Phillips, D.M. (2000) Evaluation of a probe for prediction beef tenderness. Meat Science 55:493-502.

Jiang, S.T., Lee, J.J. and Chen, H.C. (1994) Purification and characterization of cathepsin B from ordinary muscle of mackerel (Scomber australasicus). Journal of Agriculture Food Chemistry 42:1073-1079.

Jiang, S.T., Lee, J.J. and Chen, H.C. (1996) Proteolysis of Actomyosin by Cathepsins B, L, L-like and X from mackerel (Scomber australasicus). Journal of Agriculture Food Chemistry 44:769-773.

Jiang, S.T., Wang, Y.T. and Chen, C.S. (1992) Lysosomal enzyme effect on the postmortem changes in tilapia (Tilapia nilotica X T. aurea). Journal of Food Science 57:277-279.

John, B., Enright, A.J., Aravin, A., Tuschl, T., Sander, C. and Marks, D.S. (2004) Human microRNA targets. PLoS Biology 2:e363.

Johnson, M.H., Calkins, C.R., Huffman, R.D., Johnson, D.D. and Hargrove, D.D. (1990) Differences in cathepsin B+L and calcium-dependent protease activities among breed type and their relationship to beef tenderness. Journal of Animal Science 68:2371–2379.

Johnson, P.L., McEwan, J.C., Dodds, K.G., Purchas, R.W. and Blair, H.T. (2005) Meat quality traits were unaffected by a quantitative trait locus affecting leg composition traits in Texel sheep. Journal of Animal Science 83:2729-2735.

Johnston, D.J. and Graser, H.-U. (2010) Estimated gene frequencies of GeneSTAR markers and their size of effects on meat tenderness, marbling, and feed efficiency in temperate and tropical beef cattle breeds across a range of production systems. Journal of Animal Science 88: 1971-1935.

Jurie, C., Robeline, J., Picard, B. and Geay, Y. (1995) Inter-animal variation in the biological characteristics of muscle tissue in male Limousin cattle. Meat Science 39:415-425.

Juszczuk-Kubiak, E., Sakowski, T., Flisikowski, K., Wiciñska, K., Oprzadek, J. and Rosochacki, S.J. (2004) Bovine-calpain (CAPN1) gene: new SNP within intron 14. Journal of Applied Genetics 45:457-460.

Kambadur, R., Sharma, M., Smith, T.P.L. and Bass, J.J. (1997) Mutations in myostatin (GDF8) in double-muscled Belgian Blue and Piedmontese cattle. Genome research 7:910-915.

Kanadia, R.N., Johnstone, K., Mankodi, A., Lungu, A.C., Thornton, C.A., Esson, D., Timmers, A.M., Hauswirth, W.W. and Swanson, M.S. (2003) A Muscleblind Knockout Model for Myotonic Dystrophy. Science 302:1978-1980.

266

Kanawa, Ji, J.-R.. and Takahashi, K. (2002) Inactivity of µ-calpain throughout postmortem ageing of meat. Journal of Food Science 67:635-638.

Kao, C.H. and Zeng, Z.B. (2002) Modeling epistasis of quantitative trait loci using Cockerham's model. Genetics 160:1243-1261.

Karamichou, E., Richardson, R.I., Nute, G.R., McLean, K.A. and Bishop, S.C. (2006) A partial genome scan to map quantitative trait loci for carcass composition, as assessed by X-ray computer tomography, and meat quality traits in Scottish Blackface Sheep. Animal Science 82:301-309.

Karim, L., Coppieters, W., Grobet, L., Valentini, A. and Georges, M. (2002) Convenient genotyping of six myostatin mutations causing double-muscling in cattle using a multiplex oligonucleotide ligation assay. Animal Genetics 31:396– 399.

Kawase, T., Akatsuka, Y., Torikai, H., Morishima, S., Oka, A., Tsujimura, A., Miyazaki, M., Tsujimura, K., Miyamura, K., Ogawa, S., Inoko, H., Morishima, Y., Kodera, Y., Kuzushima, K. and Takahashi, T. (2007) Alternative splicing due to an intronic SNP in HMSD generates a novel minor histocompatibility antigen. Blood 110:1055-1063.

Keele, J. W., Shackelford, S.D., Kappes, S.M., Koohmaraie, M. and Stone, R.T. (1999) A region on bovine chromosome 15 influences beef longissimus tenderness in steers. Journal of Animal Science 77:1364–1671.

Kemp, C. M., Parr, T., Bardsley, R. G. and Buttery, P. J. (2006) Comparison of them relative expression of caspase isoforms in different porcine skeletal muscles. Meat Science 73: 426–431.

Kemp, C. M. and Parr, T. (2008). The effect of recombinant caspase 3 on myofibrillar proteins in porcine skeletal muscle. Animal 2:1254–1264.

Kertesz, M., Iovino, N., Unnerstall, U., Gaul, U. and Segal, E. (2007) The role of site accessibility in microRNA target recognition. Nature Genetics 39:1278–1284.

Khatkar, M.S., Thomson, P. C., Tammen, I. and Raadsma, W.R. (2004) Quantitative trait loci mapping in dairy cattle: review and meta-analysis. Genetics Selection Evolution 36:163–190.

Khatkar, M.S., Zenger, K.R., Hobbs, M., Hawken, R.J., Cavanagh, J.A.L., Barris, W., McClintock, A.E., McClintock, S., Thomson, P.C., Tier, B., Nicholas, F.W. and Raadsma, H.W. (2007) A Primary Assembly of a Bovine Haplotype Block Map Based on a 15,036-Single-Nucleotide Polymorphism Panel Genotyped in Holstein–Friesian Cattle. Genetics 176:763–772.

Kim, R.H., Flanders, K.C., Reffey, S.B. Anderson, L.A., Duckett, C.S., Perkins, N.D. and Roberts, A.B. (2001) SNIP1 inhibits NF-{kappa}B signaling by competing for its binding to the C/H1 domain of CBP/p300 transcriptional co-activators. Journal of Biological Chemistry 276:46297-46230.

267

Kim, R.H., Wang, D., Tsang, M., Martin, J., Huff, C., de Caestecker, M.P. Parks, W.T., Meng, X., Lechleider, R.J., Wang, T. and Roberts, A.B. (2000) A novel Smad nuclear interacting protein, SNIP1, suppresses p300-dependent TGF- signal transduction. Genetical Research 14:1605-1616.

Kim, S.W., Li, X.P., Lee, Y.M., Choi, Y.I., Cho, B.W., Choi, B.H., Kim, T.H., Kim, J.J. and Kim, K.S. (2011) QTL Scan for Meat Quality Traits Using High-density SNP Chip Analysis in Cross between Korean Native Pig and Yorkshire. Asian- Australasian Journal of Animal Sciences 24:1184-1191.

King, N.L. (1984) Breakdown of connectin during cooking of meat. Meat Science 11:27-43.

Kobolak, J. and Gocza, E. (2002) The role of the myostatin protein in meat quality: A review. Archiv fuer Tierzucht 45:159-170.

Koohmaraie, M. (1992) Ovine skeletal muscle multicatalytic proteinase complex (proteasome): purification, characterization, and comparison of its effects on myofibrils with m-calpain. Journal of Animal Science 70:3697–3708.

Koohmaraie, M. (1994) Muscle proteinases and meat ageing. Meat Science 36:93- 104.

Koohmaraie, M. (1996) Biochemical factors regulating the toughening and tenderisation process of meat. Meat Science 43:S193–S201.

Koohmaraie, M., Crouse, J.D. and Mersmann, H.J. (1989) Acceleration of post- mortem tenderisation in lamb and Brahman-Cross beef carcasses through infusion of calcium chloride: effect of concentration and ionic strength. Journal of Animal Science 67:934-942.

Koohmaraie, M. and Geesink, G. H. (2006) Contribution of postmortem muscle biochemistry to the delivery of consistent meat quality with particular focus on the calpain system. Meat Science 74:34–43.

Koohmaraie, M., Kent, M.P., Shackelford, S.D., Veiseth, E. and Wheeler, T.L. (2002) Meat tenderness and muscle growth: is there any relationship? Meat Science 62:345-352.

Koohmaraie, M., Whipple, G., Kretchmar, D.H., Crouse, J.D. and Mersmann, H.J. (1991) Postmortem proteolysis in longissimus muscle from beef, lamb and pork carcasses. Journal of Animal Science 69:617-624.

Krahe, R., Ashizawa, T., Abbruzzese, C., Roeder, E., Carango, P., Giacanelli, M., Funanage, V. L. and Siciliano, M. J. (1995) Effect of myotonic dystrophy trinucleotide repeat expansion on DMPK transcription and processing. Genomics 28:1-14.

268

Krendel, M. and Mooseker, M.S. (2005) Myosins: tails (and heads) of functional diversity. Physiology 20: 239–251.

Kroger, C., Bartle, C.M., West, J.G., Purchas, R.W. and Devine, C.E. (2006). Meat tenderness evaluation using dual energy X-ray absorptiometry (DEXA). Computers and Electronics in Agriculture 54:93-100.

Kuhn, K. (1986) The collagen family-variations in the molecular and supermolecular structure, Rheumatology 10:29–69.

Kuypers, R. and Kurth, L.B. (1995) Collagen’s contribution to meat texture. In ‘Meat ’95: proceedings of the Australian meat industry research conference’ Gold Coast, September 1995. pp. 11B 1–8. (McBerry Pty Ltd: Queensland)

Lall, S., Grun, D., Krek, A., Chen, K., Wang, Y.L., Dewey, C.N., Sood, P., Colombo, T., Bray, N., MacMenamin, P., Kao, H.L., Gunsalus, K.C., Pachter, L. and Piano, F. (2006) A genome-wide map of conserved microRNA targets in C. elegans. Current Biology 16:460–471.

Lander, E.S. and Botstein, D. (1989) Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-199.

Lawes Agricultural Trust (2005) GenStat Release "8.1 ( PC/Windows 2000) 09 April 2009 15:10:55 Copyright". (Rothamsted Experimental Station).

Lawrie, R.A. (1985) In “Meat Science” 4th Ed. Pergamon Press, Oxford, UK.

Lawrie, R.A. (1992) In “The Chemistry of Muscle-Based Foods”, p.43, Ed. by Johnston, D. E., Knight, M. K. and Ledward, D. A.; The Royal Society of Chemistry, Cambridge.

Layman, D.L., McGoodwin, E.B. and Martin, G.R. (1971) The Nature of the Collagen Synthesized by Cultured Human Fibroblasts. PNAS 68:454-458.

Lee, J.J., Chen, H.C. and Jiang, S.T. (1993) Purification and characterization of proteinases identified as cathepsin L and L-like (58kDa) from mackerel (Scomber australasics). Bioscience Biotechnology and Biochemistry 57:1470- 1476.

Lee, S., Stevenson-Barry, J.M., Kauffman, R. G. and Kim, B.C. (2000) Effect of ion fluid injection on beef tenderness in association with calpain activity. Meat Science 56:301-310.

Lefaucheur, L. (2010) A second look into fibre typing--relation to meat quality. Meat Science 84:257-270.

Lehnert, S.A., Byrne, K.A., Reverter, A., Nattrass, G.S., Greenwood, P.L., Wang, Y.H., Hudson, N.J. and Harper, G.S. (2006) Gene expression profiling of bovine skeletal muscle in response to and during recovery from chronic and severe undernutrition. Journal of Animal Science 84:3239-3250.

269

Leloup, L., Daury, L., Maz`eres, G., Cottin, P. and Brustis, JJ. (2007) Involvement of the ERK/MAP kinase signalling pathway in milli-calpain activation and myogenic cell migration. The International Journal of Biochemistry and Cell Biology 39:1177–1189.

Lepetit, J., (2007) A theoretical approach of the relationships between collagen content, collagen cross-links and meat tenderness. Meat Science 76:147-159.

Lepetit, J. and Culioli, J. (1994) Mechanical properties of meat. Meat Science 36:203-237.

Li, C.F., MacDonald, J.R., Wei, R.Y., Ray, J., Lau, K., Kandel, C., Koffman, R., Bell, S., Scherer, S.W. and Alman, B. A. (2007a) Human sterile alpha motif domain 9, a novel gene identified as down-regulated in aggressive fibromatosis, is absent in the mouse. BMC Genomics 8:92.

Li, C.B., Xu, X.L., Zhou, G.H., Xu, S.Q. and Zhang, J.B. (2007b) Effects of carcass maturity on meat quality characteristics of beef semitendinosus muscle for Chinese native yellow steers. Animal 1:780-786.

Light, N., Champion, A.E., Voyle, C. and Bailey, A.J. (1985) The role of epimysial, perimysial and endomysial collagen in determining texture in six bovine muscles. Meat Science 13:137–149.

Lin, Y., Martin, J., Gruendler, C., Farley, J., Meng, X., Li, B.Y., Lechleider, R., Huff, C., Kim, R. H., Grasser, W., Paralkar, V. and Wang, T. (2002) A novel link between the proteasome pathway and the signal transduction pathway of the Bone Morphogenetic Proteins (BMPs). BMC Cell Biology 3:15.

Linder, E., Anders, R.F. and Natvig, J.B. (1976) Connective tissue origin of the amyloid-related protein SAA. The Journal of Experimental Medicine 144:1336:1346.

Lines, D. (2006) Meat quality in beef cattle: Factors affecting tenderness. Thesis submitted for Bachelor of Science (Agricultural Science) with Honours at the University of Adelaide.

Lines, D. S., Pitchford, W.S., Kruk, Z.A. and Bottema, C.D.K. (2009) Limousin myostatin F94L variant affects semitendinosus tenderness. Meat Science 81:126–31.

Lindholm-Perry, A.K., Rohrer, G.A., Holl, J.W., Shackelford, S.D., Wheeler, T.L., Koohmaraie, M. and Nonneman, D. (2009) Relationships among calpastatin single nucleotide polymorphisms, calpastatin expression and tenderness in pork longissimus. Animal Genetics 40:713-721.

Listrat, A., Lethias, C., Hocquette, J.F., Renand, G., Me´nissier, F., Geay, Y. and Picard, B. (2000) Age-related changes and location of types I, III, XII and XIV collagen during development of skeletal muscles from genetically different animals. Histochemical Journal 32:349–356.

270

Listrat, A., Rakadjiyski, N., Jurie, C., Picard, B., Touraille, C. and Geay, Y. (1999) Effect of the type of diet on muscle characteristics and meat palatability of growing Salers bulls. Meat Science 53:115-124.

Liu D.J. and Leal S. M. (2010) A novel adaptive method for the analysis of next- generation sequencing data to detect complex trait associations with rare variants due to gene main effects and interactions. PLoS Genetics 6: e1001156.

Liu, H.C., Hicks, J.A., Trakooljul, N. and Zhao, S.H. (2010) Current knowledge of microRNA characterization in agricultural animals. Animal Genetics 41:225- 231.

Liu, X., Zhao, Y., Gao, J., Pawlyk, B., Starcher, B., Spencer, J. A., Yanagisawa, H., Zuo, J. and Li, T. (2004) Elastic fiber homeostasis requires lysyl oxidase−like 1 protein. Nature Genetics 36:178 - 182.

Locker, R.H. and Hagyard, C.J. (1963) A cold shortening eﬀect in beef muscle. Journal of the Science of Food 14:787–793.

Lucero, H.A. and Kagan, H.M. (2006) Lysyl oxidase: an oxidative enzyme and effector of cell function (Review). Cellular and Molecular Life Sciences 63:2304-2316.

Mackay, T.F.C (2001). The genetic architecture of quantitative traits. Annual Review of Genetics 34:303-339.

Maltin, C.A., Delday, M. I., Hay, S.M., Smith, F.G., Lobley, G.E. and Reeds, P.J. (1987) The effect of the anabolic agent, clenbuterol on overloaded rat skeletal muscle. Bioscience Report 7:143-149.

Maltin, C.A., Warkup, C.C., Matthews, K.R., Grant, C.M., Porter, A.D. and Delday, M.I. (1997) Pig muscle fibre characteristics as a source of variation in eating quality. Meat Science 47:237-248.

Marian, A.J. (2012) Molecular genetic studies of complex phenotypes. Translational Research 159:65-79.

Marth, G.T., Korf, I., Yandell, M.D., Yeh, R.T., Gu, z., Zakeri, H., Stitziel, N.O., Hillier, L., Kwok, P.Y. and Gish, W.R. (1999) A general approach to single- nucleotide polymorphism discovery. Nature Genetics 23:452-456.

Martyn, J.K., Bass, J.J. and Oldham, J.M. (2004) Skeletal muscle development in normal and double-muscled cattle. Anatomical Record Part A: Discoveries in Molecular Cellular and Evolutionary Biology 281A:1363-1371.

Matena, K., Boehm, T. and Dear T.N., (1998) Genomic Organization of Mouse Capn5 and Capn6 Genes Confirms That They Are a Distinct Calpain Subfamily. Genomics 48:117–120.

271

Matukumalli, L. K., Lawley, C.T., Schnabel, R.D., Taylor, J.F., Allan, M.F., Heaton, M.P., O’Connell, J., Moore, S.S., Smith, T.P.L., Sonstegard, T.S. and Van Tassell C.P. (2009) Development and Characterization of a High Density SNP Genotyping Assay for Cattle. PLoS ONE 4: e5350.

McDonagh, M.B., Fernandez, C. and Oddy, V.H. (1999) Hind-limb protein metabolism and calpain system activity influence post-mortem change in meat quality in lamb. Meat Science 52:9–18.

McKay, S.D., Schnabel, R.D., Murdoch, B.M., Matukumalli, L.K., Aerts, J., Coppieters, W., Crews, D., Neto, E.D., Gill, C.A., Gao, C., Mannen, H., Stothard, P., Wang, Z., Tassell, C.P.V., Williams, J.L., Taylor, J.F. and Moore, S.S. (2007) Whole genome linkage disequilibrium maps in cattle. BMC Genetics 8:74.

McPherron, A.C., Lawler, A.M. and Lee, S.J. (1997) Regulation of skeletal muscle mass in mice by a new TGF-β superfamily member. Nature 387:83–90.

Meuwissen, T.H.E. and Goddard, M.E. (2000) Fine mapping of quantitative trait loci using linkage disequilibria with closely linked marker loci. Genetics 155:421– 430.

Meyers, S.N. and Beever, J.E. (2008) Investigating the genetic basis of pork tenderness: genomic analysis of porcine CAST. Animal Genetics 39:531-543.

Miller, J.W., Urbinati, C.R., Teng-umnuay, P., Stenberg, M.G., Byrne, B.J., Thornton, C.A. and Swanson, M.S. (2000) Recruitment of human muscleblind proteins to (CUG)n expansions associated with myotonic dystrophy. EMBO Journal 19:4439-4448.

Miller, M.F., Carr, M.A., Ramsey, C.B., Crockett, K.L., and Hoover, L.C. (2001) Consumer thresholds for establishing the value of beef tenderness. Journal of Animal Science 79:3062–3068.

Miller, R. D. and Kwok, P.Y. (2001) The birth and death of human single-nucleotide polymorphisms: new experimental evidence and implications for human history and medicine. Human Molecular Genetics 10(20):2195-2198.

Miquel, M.C., Villarreal, E., Mezzadra, C., Melucci, L., Soria, L., Corva, P. and Schor, A. (2009) The association of CAPN1 316 marker genotypes with growth and meat quality traits of steers finished on pasture. Genetics and Molecular Biology 32(3):491-496.

Moody, W.G., Kemp, J.D., Mahyuddin, M., Johnston, D.M. and Ely, D.G. (1980) Effect of Feeding Systems, Slaughter Weight and Sex on Histological Properties of Lamb Carcasses. Journal Animal Science 50:249-256.

Morgan, J.B., Savell, J.W., Hale, D.S., Miller, R.K., Griffin, D.B., Cross, H.R. and Shackelford, S.D. (1991) National beef tenderness survey. Journal of Animal Science 69:3274-3283.

272

Morris, C.A., Cullen, N.G. Hickey, S.M., Dobbie, P.M., Veenvliet B.A., Manley, T.R., Pitchford, W.S. Kruk, Z.A., Bottema, C.D.K. and Wilson, T. (2006) Genotypic effects of calpain 1 and calpastatin on the tenderness of cooked m. longissimus dorsi steaks from Jersey x Limousin, Angus and Hereford-cross cattle. Animal Genetics 37: 411-414.

Morris, C.A., Pitchford, W.S., Cullen, N.G., Esmailizadeh, A.K., Hickey, S.M., Hyndman, D., Dodds, K.G., Afolayan, R.A., Crawford, A.M. and Bottema, C.D.K. (2009) Quantitative trait loci for live animal and carcass composition traits in Jersey and Limousin back-cross cattle finished on pasture or feedlot. Animal Genetics 40:648–654.

Muir, P.D., Deaker, J.M., and Bown, M.D. (1998) Effects of forage- and grain-based feeding systems on beef quality: A review. New Zealand Journal of Agricultural Research 41:623−635.

Mulders, S.A.M., van Engelen, B.G.M., Be´ Wieringa and Wansink, D.G. (2010) Molecular therapy in myotonic dystrophy: focus on RNA gain-of-function. Human Molecular Genetics 19:R90-R97.

Muroya, S. Ertbjerg, P. Pomponio, L. and Christensen, M. (2010) Desmin and troponin T are degraded faster in type IIb muscle fibers than in type I fibers during postmortem aging of porcine muscle. Meat Science 86:764–769.

Musarò, A., McCullagh, K., Paul, A., Houghton, L., Dobrowolny, G., Molinaro, M., Barton, E.R., Sweeney, H.L. and Rosenthal, N. (2001) Localized Igf-1 transgene expression sustains hypertrophy and regeneration in senescent skeletal muscle. Nature Genetics 27:195-200.

Naik, M. B. (2008). Identification and characterization of genetic markers and metabolic pathways controlling net feed efficiency in beef cattle. Adelaide, The University of Adelaide. Doctor of Philosophy.

Neath, K.E., Del Barrio, A.N., Lapitan, R.M., Herrera, J.R.V., Cruz, L.C., Fujihara, T., Muroya, S., Chikuni, K., Hirabayashi, M. and Kanai, Y. (2007a) Difference in tenderness and pH decline between water buffalo meat and beef during postmortem aging. Meat Science 75(3):499–505.

Neath, K.E., Del Barrio, A.N., Lapitan, R.M., Herrera J.R.V., Cruz, L.C., Fujihara, T., Muroya, s., Chikuni, K., Hirabayashi, M. and Kanai, Y. (2007b) Relation of postmortem protease activity to tenderness in buffalo meat and Brahman beef. Italian Journal of Animal Science 6(Suppl. 2):1175-1177.

Ngapo, T. M., Berge, P., Culioli, J., Dransfield, E., De Smet, S. and Claeys, E. (2002) Perimysial collagen crosslinking and meat tenderness in Belgian Blue double muscled cattle. Meat Science 61:91−102. Nishi, M., Yasue, A., Nishimatu, S., Nohno, T., Yamaoka, T., Itakura, M., Moriyama, K., Ohuchi, H. and Noji, S. (2002) A missense mutant myostatin causes hyperplasia without hypertrophy in the mouse muscle. Biochemical and Biophysical Research Communications 293:247-251.

273

Nishimura, T., Hattori, A., and Takahashi, K. (1999) Structural changes in intramuscular connective tissue during the fattening of Japanese black cattle: effect of marbling on beef tenderisation. Journal of Animal Science 77: 93–104.

Nonneman, D., Lindholm-Perry, A.K., Shackelford, S.D., King, D.A., Wheeler, T.L., Rohrer, G.A., Bierman, C.D., Schneider, J.F., Miller, R.K., Zerby, H. and Moeller, S.J. (2011) Predictive markers in calpastatin for tenderness in commercial pig populations. Journal of Animal Science 89:2663-2672.

Norman, G. A. (1982). Effect of breed and nutrition on the productive traits of beef cattle in South-east Brazil: part 3 – meat quality. Meat Science 6:79–96.

Novianti, I. (2009) Molecular genetics of cattle muscularity. Animal Science. Adelaide, The University of Adelaide. Master of Agricultural Science.

Nuernberg, K., Dannenberger, D., Nuernberg, G., Ender, K., Voigt, J., Scollan, N. D., Wood, J.D., Nutec, G.R. and Richardson, R.I. (2005) Effect of a grass-based and a concentrate feeding system on meat quality characteristics and fatty acid composition of longissimus muscle in different cattle. Livestock Production Science 94:137−147.

O'Connor, S.F., Tatum, J.D. Wulf, D.M., Green, R.D. and Smith, G.C. (1997) Genetic effects on beef tenderness in Bos indicus composite and Bos taurus cattle. Journal of Animal Science 75:1822-1830.

Oddy, V.H., Harper, G.S., Greenwood, P.L. and McDonagh, M.B. (2001) Nutritional and developmental effects on the intrinsic properties of muscles as they relate to the eating quality of beef. Australian Journal of Experimental Agriculture 41:921–942.

O'Halloran, G.R., Troy, D.J. Buckley, D.J. and Reville, W.J. (1997) The role of endogenous proteases in the tenderisation of fast glycolysing muscle. Meat Science 47:187-210.

Ohno, S., Minoshima, S., Kudoh, J., Fukuyama, R., Shimuzu, Y., Ohmi-Imajoh, S., Shimizu, N. and Suzuki, K. (1990) Four genes for the calpain family locate on four distinct human chromosomes. Cytogenetics Cell Genetics 33:225–229.

Olety, B., Waite, M., Honnert, U., Schillers, H. and Bahler, M. (2010) Myosin 1G (Myo1G) is a haematopoietic specific myosin that localises to the plasma membrane and regulates cell elasticity. FEBS Letters 584: 493–499.

Olsson, U., Hertzman, C. and Tornberg, E. (1994) The influence of low temperature, type of muscle and electrical stimulation on the course of rigor mortis, ageing and tenderness of beef muscles. Meat Science 37:115–131.

Ouali, A. (1992) Proteolytic and physicochemical mechanisms involved in meat texture development. Biochimie 74:251–265.

274

Ouali, A. and Talmant, A. (1990) Calpains and calpastatin distribution in bovine, porcine and ovine skeletal muscles. Meat Science 28:331-348.

Ouchi, N., Oshima, Y., Ohashi, K., Higuchi, A., Ikegami, C., Izumiya, Y. and Walsh, K. (2008) Follistatin-like 1, a Secreted Muscle Protein, Promotes Endothelial Cell Function and Revascularization in Ischemic Tissue through a Nitric-oxide Synthase-dependent Mechanism. The Journal of Biological Chemistry 283:32802–32811.

Oury, M.P., Picard, B., Briand, M., Blanquet, J.P. and Dumont, R. (2009) Interrelationships between meat quality traits, texture measurements and physicochemical characteristics of M. rectus abdominis from Charolais heifers. Meat Science 83:293–230.

Page, B.T., Casas, E., Heaton, M.P., Cullen, N.G., Hyndman, D.L., Morris, C.A., Crawford, A. M., Wheeler, T.L., Koohmaraie, M., Keele, J.W. and Smith, T.P.L. (2002) Evaluation of single-nucleotide polymorphisms in CAPN1 for association with meat tenderness in cattle. Journal of Animal Science 80:3077- 3085.

Page, B.T., Casas, E., Quaas, R.L., Thallman, R.M., Wheeler, T.L., Shackelford, S.D., Kochmaraie, M., White, S.N., Bennett, G.L., Keele, J.W., Dikeman, M.E. and Smith, T.P.L. (2004) Association of markers in the bovine CAPN1 gene with meat tenderness in large crossbred populations that sample influential industry sires. Journal of Animal Science 82:3474-3481.

Park, K.W., Halperin, D.S. and Tontonoz, P. (2008) Before they were fat: adipocyte progenitors. Cell Metabolism 8(6):454-457.

Parr, T., Bardsley, R.G., Gilmour, R.S. and Buttery, P.J. (1992) Changes in calpain and calpastatin mRNA induced by beta-adrenergic stimulation of bovine skeletal muscle. European Journal of Biochemistry 208:333–339.

Paul, A.C. and Rosenthal, R. (2002) Different modes of hypertrophy in skeletal muscle fibers. The Journal of Cell Biology 156:751-760.

Pearson, A.M. and Young, R.B. (1989) Muscle and meat biochemistry. San Diego: Academic Press.

Penny, I.F., Etherington, D.J., Reeves, J.L. and Taylor, M.A.J. (1984) Proc. 30th Europ. Meet. Meat Res. Workers Bristol. 133.

Peter, J.B., Barnard, R.J., Edgerton, V.R., Gillespie, C.A. and Stemple, K.E. (1972) Metabolic profiles of three fiber types of skeletal muscle in guinea pigs and rabbits. Biochemistry 11:2627-2633.

Picard, B., Gagniere, H., Geay, Y., Hocquette, J.F. and Robelin, J. (1994) Study of the influence and age of weaning on the contractile and metabolic characteristics of bovine muscle. Reproduction Nutrition and Development 35:71-84.

275

Picard, B., Lefaucheur L., Berri C. and Duclos J.M. (2002) Muscle fibre ontogenesis in farm animal species. Reproduction Nutrition Development 42: 415–431.

Platter, W. J., Tatum, J.D., Belk, K.E., Chapman, P.L., Scanga, J.A. and Smith, G.C. (2003) Relationships of consumer sensory ratings, marbling score, and shear force value to consumer acceptance of beef strip loin steaks. Journal of Animal Science 81:2741–2750.

Powell, T.H., Dikeman, M.E. and Hunt, M.C. (2000) Tenderness and collagen composition of beef semitendinosus roasts cooked by conventional convective cooking and modeled, multi-stage, convective cooking. Meat Science 55:421- 425.

Powell, W., Morgante, M., Andre, C., Hanafey, M., Vogel, J., Tingey, S. and Rafalski, A. (1996) The comparison of RFLP, RAPD, AFLP and SSR (microsatellite) markers for germplasm analysis. Molecular Breeding 2:225-238.

Purchas, R.W. (1990) An assessment of the role of pH differences in determining the relative tenderness of meat from bulls and steers. Meat Science 27:129-140.

Purchas, R.W. and Aungsupakorn, R. (1993) Further investigations into the relationship between ultimate pH and tenderness for beef samples from bulls and steers. Meat Science 34:163-178.

Purchas, R.W., Burnhamd, L. and Morris, S.T. (2002) Effects of growth potential and growth path on tenderness of beef longissimus muscle from bulls and steers. Journal of Animal Science 80:3211-3221.

Purslow, P.P. (1999) The intramuscular connective tissue matrix and cell–matrix interactions in relation to meat toughness. In: Proceedings of the 45th international congress of meat science and technology Yokohama, Japan. 210- 219.

Reagan, J. O., Carpenter, Z. L. and Smith, G. C. (1976) Age-Related Traits Affecting the Tenderness of the Bovine. Journal of Animal Science 43:1198-1205.

Redowicz, M. J. (2007) Unconventional myosins in muscle. European Journal of Cell Biology 86: 549–558.

Reed, G.H. and Wittwer, C.T. (2004) Sensitivity and specificity of single-nucleotide polymorphism scanning by high-resolution melting analysis. Clinical Chemistry 50:1748-1754.

Rehfeldt, C., Fiedler I. and Stickland N.C. (2004) Number and size of muscle fibres in relation to meat production. In Muscle development of livestock animals – physiology, genetics, and meat quality. Te Pas M.F.W., Everts, M.E. and Haagsman, H.P. (eds.), CABI Publishing, Wallingford, UK, pp. 1–38.

Relaix, F. and Marcelle, C. (2009) Muscle stem cells. Current Opinion in Cell Biology 21:748–753.

276

Renand, G., Picard, B., Touraille, C., Berge, P., and Lepetit, J. (2001) Relationships between muscle characteristics and meat quality traits of young Charolais bulls. Meat Science 59:49-60.

Rexroad, C.E., Bennett, G.L., Stone, R.T., Keele, J.W., Fahrenkrug, S.C., Freking, B. A., Kappes, S.M. and Smith T.P.L., (2001) Comparative mapping of BTA15 and HSA11 including a region containing a QTL for meat tenderness. Mammalian Genome 12: 561–565.

Riley, D.G., Johnson, D.D., Chase Jr., C.C., West, R. L., Coleman, S.W., Olson, T.A. and Hammond, A.C. (2005) Factors influencing tenderness in steaks from Brahman cattle. Meat Science 70:347-356.

Riley, D.G., Chase, C.C. Jr., Hammond, A.C., West, R.L., Johnson, D.D., Olson, T.A. and Coleman, S.W. (2003) Estimated genetic parameters for palatability traits of steaks from Brahman cattle. Journal of Animal Science 81:54-60.

Robert, N., Briand, M., Taylor, R. and Briand Y. (1999) The effect of proteasome on myofibrillar structures in bovine skeletal muscle. Meat Science 51(2):149-153.

Robinson, D.L., Ferguson, D.M., Oddy, V.H., Perry, D. and Thompson, J.M. (2001) Genetic and environmental influences on beef tenderness. Australian Journal of Experimental Agriculture 41:997–1003.

Robson, RM (1995) Myofibrillar and cytoskeletal structures and proteins in mature skeletal muscle cells. In Expression of tissueproteinases and regulation of protein degradation as related to meat quality. Ouali, A., Demeyer D.I, and Smulders, F.J.M. (eds.), ECCEAMST, Utrecht, pp. 267–288.

Robson, R.M., Huiatt T.W., and Parrish F.C. (1991) Biochemical and structural properties of titin, nebulin and intermediate filaments in muscle. Proceedings of the Annual Reciprocal Meat Conference 44: 7–20.

Rodas-González, A., Huerta-Leidenz, N., Jerez-Timaure, N. and Miller, M.F. (2009) Establishing tenderness thresholds of Venezuelan beef steaks using consumer and trained sensory panels. Meat Science 83:218–223.

Roeber, D.L., Cannell, R.C., Belk, K.E., Millar, R.K., Tatum, J.D. and Smith, G.C. (2000) Implant strategies during feeding: impact on carcass grades and consumer acceptability. Journal of Animal Science 78:1867–1874.

Roh, S.G., He, M.L., Matsunaga, N., Hidaka, S. and Hidari, H. (1997) Mechanisms of action of growth hormone-releasing peptide-2 in bovine pituitary cells. Journal of Animal Science 75:2744-2748.

Rohrer, G.A., Thallman, R.M., Shackelford, S., Wheeler, T. and Koohmaraie, M. (2006) A genome scan for loci affecting pork quality in a Duroc-Landrace F-2 population. Animal Genetics 37:17-27.

277

Rosser, B.W.C., Norris, B.J. and Nemeth, P.M. (1992) Metabolic capacity of individual fibres from different anatomic locations. Journal of Histochemistry and Cytochemistry 40: 819-825.

Rowe, R.W.D. (1981) Morphology of perimysial and endomysial connective tissue in skeletal muscle. Tissue and Cell 1:681-690.

Rowe, R.W.D. (1974) Collagen fibre arrangement in intramuscular connective tissue. Changes associated with muscle shortening and their possible relevance to raw meat toughness measurements. Journal of Food Technology 9:501–508.

Ruiz de Huidobro, F., Miguel, E., Bla´zquez, B. and Onega, E. (2005) A comparison between two methods (Warner–Bratzler and texture profile analysis) for testing either raw meat or cooked meat. Meat Science 69:527–536.

Sargolzaei, M., Schenkel, F.S., Jansen, G.B. and Schaeffer, L.R. (2008) Extent of Linkage Disequilibrium in Holstein Cattle in North America. Journal of Dairy Science 91:2106–2117.

Sato, K., Yamogida, k., Wada, T., Yorihuzi, T., Nishmune, Y., Hosokawa, N. and Nagata, K. (2002). Type VVXI collagen, a new member of the collagen family, is specially expressed in the testis and ovary. Journal of Biological Chemistry 277:37678-37684.

Sazili, A.Q., Lee, G.K., Parr, T., Sensky, P.L., Bardsley, R.G. and Buttery, P.J. (2003) The effect of altered growth rates on the calpain proteolytic system and meat tenderness in cattle. Meat Science 66:195–201.

Schanbarche, B. D. (1984) Manipulation of Endogenous and Exogenous Hormones for Red Meat Production. Journal of Animal Science 59:1621-1630.

Schenkel, F.S., Miller, S.P., Jiang, Z., Mandell, I.B., Ye, X., Li, H. and Wilton, J.W. (2006) Association of a single nucleotide polymorphism in the calpastatin gene with carcass and meat quality traits of beef cattle. Journal of Animal Science 84:291–299.

Schiaffino, S., Gorza, L., Sartore, S., Saggin, L., Ausoni, S., Vianello, M., Gundersen, K. and Lomo, T. (1989) Three myosin heavy chain isoforms in type 2 skeletal muscle fibres. Journal of Muscle Research and Cell Motility 10:191- 205.

Schiaffino, S. and Reggiani, C. (1996) Molecular diversity of myofibrillar proteins: Gene regulation and functional significance. Physiological Reviews 76:371–423.

Schmierer, B. and Hill, C.S. (2007) TGF beta-SMAD signal transduction: molecular specificity and functional flexibility. Nature Reviews Molecular Cell Biology 8: 970-982.

Seideman S.C. Koohmaraie, M. and Crouse, J.D. (1988) The influence of muscle fiber size on tenderness in A-maturity heifers. Journal of Food Quality 11:27-34. 278

Sellick, G.S., Pitchford, W.S., Morris, C.A., Cullen, N.G., Crawford, A.M., Raadsma, H.W. and Bottema, C.D.K. (2007) Effect of myostatin F94L on carcass yield in cattle. Animal Genetics 38:440-446.

Serrano, E., Pradel, P., Jailler, R., Dubroeucq, H., Bauchart, D. Hocquette, J.-F., Listrat, A., Agabriel, J. and Micol, D. (2007) Young Salers suckled bull production: effect of diet on performance, carcass and muscle characteristics and meat quality. Animal 1:1068–1079.

Shackelford, S.D., Koohmaraie, M., Cundiff, L.V., Gregory, K.E., Rohrer G.A. and Savell, W. (1994) Heritabilities and phenotypic and genetic correlations for bovine post rigor calpastatin activity, intramuscular fat content, Warner-Bratzler shear force, retail product yield, and growth rate. Journal of Animal Science 72:857-863.

Shackelford, S.D., Wheeler, T.L. and Koohmaraie, M. (1995) Relationship between shear force and trained sensory panel tenderness ratings of 10 major muscles from Bos indicus and Bos taurus cattle. Journal of Animal Science 73:3333- 3340.

Shackelford, S.D, Wheeler, T.L. and Koohmaraie, M. (1997) Repeatability of tenderness measurement in beef round muscles. Journal of Animal Science 75: 2411-2416.

Shackelford, S.D., Wheeler, T.L. and Koohmaraie, M. (1998) Coupling of image analysis and tenderness classification to simultaneously evaluate carcass cutability, longissimus area, subprimal cut weights, and tenderness of beef. Journal of Animal Science 76:2361-2640.

Shanks, B.C., Wulf, D.M. and Maddock R. J. (2002) Technical note: The effect of freezing on Warner-Bratzler shear force values of beef longissimus steaks across several postmortem ageing periods. Journal of Animal Science 80:2122-2125.

Sidis, Y., Mukherjee, A., Keutmann, H., Delbaere, A., Sadatsuki, M. and Schneyer, A. (2006) Biological activity of follistatin isoforms and follistatin-like-3 is dependent on differential cell surface binding and specificity for activin, myostatin, and bone morphogenetic proteins. Endocrinology 147:3586-3579.

Silva K.M., Bastiaansen J.W.M., Knol E.F., Merks, W.M., Lopes, P.S., Guimara˜es, S.E.F. and van Arendonk, J.A.M. (2011) Meta-analysis of results from quantitative trait loci mapping studies on pig chromosome 4. Animal Genetics 42: 280-292.

Smith, M.E., Kastner, C.L., Hunt, M.C., Kropf, D.H. and Allen, D.M. (1979) Elevated conditioning temperature effects on beef carcasses from four nutritional regimes. Journal of Food Science 44:158-163.

Smith, G.C., Savell, J.W., Dolezal, H.G., Field, T. G., Gill, D.R., Griffin, D.B., Hale, D.S., Morgan, J.B., Northcutt, S.L. and Tatum, J.D. (1995) The National Beef Quality Audit. Colorado State University, Texas A&M University, and

279

Oklahoma State University for the National Cattleman’s Beef Association.

Smulders, F.J.M., Marsh, B.B., Swartz, D.R., Russell, R.L. and Hoenecke, M.E. (1990) Beef tenderness and sarcomere length. Meat Science 28:349–363.

Soller, M., Brody, T. and Genizi, A. (1976) On the power of experimental design for the detection of linkage between marker loci and quantitative loci in crosses between inbred lines. Theoretical and Applied Genetics 47:35-39.

Solomon, V. and Goldberg, A.L. (1996) Importance of the ATP-ubiquitin-proteasome pathway in the degradation of soluble and myofibrillar proteins in rabbit muscle extracts. Journal of Biological Chemistry 271:26690–26697.

Sorheim, O., Idland, J., Halvorsen, E.C., Froystein, T., Lea, P. and Hildrum, K.I. (2001) Influence of beef carcass stretching and chilling rate on tenderness of m. longissimus dorsi. Meat Science 57:79-85.

Sorimachi, N., Shimada, H., Tagawa, K., Maruyama, K. and Suzuki, K. (1995) Muscle-specific calpain, p94, responsible for limb-girdle muscular dystrophy type 2 A, associates with connectin through IS2, a p94-specific sequence. Journal of Biological Chemistry 270:31158–31162.

Squire, J.M. and Morris, E.P. (1998) A new look at thin filament regulation in vertebrate skeletal muscle. The FASEB Journal 12:761-777.

Stienen, G.J.M., Kiers, J.L., Bottinelli, R. and Reggiani, C. (1996) Myofibrillar ATPase activity in skinned human skeletal muscle fibres: fibre type and temperature dependence. Journal of Physiology 493:299-307.

Stolowski, G.D., Baird, B.E., Miller, R.K., Savell, J.W., Sams, A.R., Taylor, J.F., Sanders, J.O. and Smith, S.B. (2006) Factors influencing the variation in tenderness of seven major beef muscles from three Angus and Brahman breed crosses. Meat Science 73:475–483.

Straadt, I.K., Rasmussen, M., Andersen, H.J. and Bertram, H.C. (2006) Aging- induced changes in microstructure and water distribution in fresh and cooked pork in relation to water-holding capacity and cooking loss – A combined confocal laser scanning microscopy (CLSM) and low-field nuclear magnetic resonance relaxation study. Meat Science 75:687-695.

Stranger, B.E, Eli A. Stahl, E.A. and Towfique R. (2011) Progress and promise of genome-wide association studies for human complex trait genetics. Genetics 187: 367–383.

Suyama, M. and Konosu, A. (1987) In “Suisan Shokuhin Gaku” Marine Food Science Koseisha, Koseikaku, Tokyo.

Swatland, H.J. (1992) Bidirectional operation of a UV fluorescence probe for beef carcass connective tissues. Computers and Electronics in Agriculture 7:285-300.

280

Swatland, H.J. and Kieffer, N.M. (1974) Fetal development of the double muscled condition in cattle. Journal of Animal Science 38:752–757.

Taillon-Miller, P., Gu, Z.J., Li, Q., Hillier, L. and Kwok, P.Y. (1998) Overlapping genomic sequences - a treasure trove of single-nucleotide polymorphisms. Genome Reserve 8:748-754.

Takahashi, G, Wang, S.M., Lochner, J.V. and Marsh, B.B. (1987) Effects of 2-Hz and 60-Hz electrical stimulation on the microstructure of beef. Meat Science 19:65–76.

Tatum, J.D., Smith, G.C., Bery, B.W., Murphey, C.E., Williams, F.L. and Carpenter, Z.L. (1980) Carcass characteristics, time on feed and cooked beef palatability attributes. Journal of Animal Science 50:833–840.

Taylor, R.G., Geesink, G.H., Thompson, V.F., Koohmaraie, M. and Goll, D.E. (1995) Is Z-disk degradation responsible for postmortem tenderization? Journal of Animal Science 73:1351–1367.

Therkildsen, M., Houbak, M. and Byrne, D.V. (2008) Feeding strategy for improving tenderness has opposite effects in two different muscles. Meat Science 80:1037- 1045.

Therkildsen, L., Larsen, M., Bang, H.G. and Vestergaard, M. (2002) Effect of growth rate on tenderness development and final tenderness of meat from Friesian calves. Animal Science 74:253–264.

Thomas, M., Langley, B., Berry, C., Sharma, M., Kirk, S., Bass, J. and Kambadur R. (2000) Myostatin, a Negative Regulator of Muscle Growth, Functions by Inhibiting Myoblast Proliferation. Journal of Biological Chemistry 275:40235- 40243.

Thomassin, L., Werneck, C.C., Broekelmann, T.J., Gleyzal, C., Hornstra, I.K., Mecham, R. P. and Sommer, P. (2005) The Pro-regions of Lysyl Oxidase and Lysyl Oxidase-like 1 Are Required for Deposition onto Elastic Fibers. The Journal of Biological Chemistry 280:42848-42855.

Thompson, J. (2002) Managing meat tenderness. Meat Science 62:295-308.

Thompson, J.M., Perry, D., Daly, B., Gardner, G.E., Johnston, D.J. and Pethick, D.W. (2006) Genetic and environmental effects on the muscle structure response post- mortem. Meat Science 74:59–65.

Tian R, Pitchford WS, Morris CA, Cullen NG and CDK Bottema (2010) Genetic variation in the beta,beta-carotene-9',10'-dioxygenase gene and association with fat colour in bovine adipose tissue and milk. Animal Genetics 41:253-259.

Tokuhiro, S., Yamada, R., Chang, X., Suzuki, A., Kochi, Y., Sawada, T., Suzuki, M., Nagasaki, M., Ohtsuki, M., Ono, M., Furukawa, H., Nagashima, M., Yoshino, S., Mabuchi, A., Sekine, A., Saito, S., Takahashi, A., Tsunoda, T., Nakamura, Y. and

281

Yamamoto, K. (2003) An intronic SNP in a RUNX1 binding site of SLC22A4, encoding an organic cation transporter, is associated with rheumatoid arthritis. Nature Genetics 35:341 – 348.

Torrescano, C., Sanchez-Escalante, A., Gimnez, B., Roncales, P. and Beltran, J.A. (2003) Shear values of raw samples of 14 bovine muscles and their relation to muscle collagen characteristics. Meat Science 64:85–91.

Totland, G. K., Kryvi, H. and Slinde, E. (1988). Composition of muscle fibre types and connective tissue in bovine M. semitendinosus and its relation to tenderness. Meat Science 23: 303-315.

Tuma, H. J., Venable, J.H., Wuthier, P.R. and Henrickson, R. L. (1962) Relationship of Fiber diameter to tenderness and meatiness as influenced by bovine age. Journal of Animal Science 21:22-26.

Ueda, Y., Watanabe, A., Higuchi, M., Shingu, H., Kushibiki, S. and Shinoda, M. (2007) Effects of intramuscular fat deposition on the beef traits of Japanese Black steers (Wagyu) Animal Science Journal 78:2 189-194.

Ujan, J.A., Zan, L.S., Ujan, S.A., Adoligbe, C. and Wang, H.B. (2011a) Back fat thickness and meat tenderness are associated with a 526 T -> A mutation in the exon 1 promoter region of the MyF-5 gene in Chinese Bos taurus. Genetics and Molecular Research 10:3070-3079.

Ujan, J.A., Zan, L.S., Wei, S., Adoligbe, C. and Wang, H.B. (2011b) Meat tenderness and water holding capacity are associated with a 959 A-> G mutation in the MyoG gene of Chinese indigenous cattle. African Journal of Biotechnology 10(29):5654-5660.

Underwood, K.R., Means, W,J. and Du, M. (2008) Caspase 3 is not likely involved in the postmortem tenderization of beef muscle. Journal of Animal Science 86:960-966.

Valin, C., Tassy, C., Farias, G., Geesink, G. and Ouali, A. (1993) International Workshop on Proteolysis and Meat Quality, Clermont, Ferrand. 24-28.

Van Eenennaam, A.L., Li, J., Thallman, R.M., Quaas, R.L., Dikeman, M.E., Gill, C.A., Franke, E. and Thomas, M.G. (2007) Validation of commercial DNA tests for quantitative beef quality traits. Journal of Animal Science 85:891–900.

Veiseth, E., Shackelford, S.D., Wheeler, T.L. and M. Koohmaraie, M. (2004) Factors regulating lamb longissimus tenderness are affected by age at slaughter. Meat Science 68:635–640.

Vestergaard, M., Therkildsen, M., Henckel, P., Jensen, L.R., Andersen, H.R. and Sejrsen, K. (2000) Influence of feeding intensity, grazing and finishing feeding on meat and eating quality of young bulls and the relationship between muscle fibre characteristics, fibre fragmentation and meat tenderness. Meat Science 54:187-195.

282

Vo, N. and Goodman, R.H. (2001) CREB-binding Protein and p300 in Transcriptional Regulation. Journal of Biological Chemistry 276: 13505-13508.

Wang, G.J., Yang, P. and Xie, H.G. (2006) Gene variants in noncoding regions and their possible consequences. Pharmacogenomics 7:203-209.

Wang, H., Claus, J.R. and Marriott, N.G. (1994) Selected skeletal alterations to improve tenderness of beef round muscles. Journal of Muscle Foods 5:137–147.

Warner, R.D., Greenwood, P.L., Pethick, D.W. and Ferguson, D.M. (2010) Genetic and environmental effects on meat quality. Meat Science 86:171-183.

Watanable, T., Hosoya, H. and Yonemura, S. (2007) Regulation of myosin II dynamics by phosphorylation and dephosporylation of its light chain in epithelial cells. Molecular Biology of the Cell 18:606-616.

Wegner, j., Albrecht, E., Fiedler, F., Teuscher, Papstein H. J. and Ender, K. (2000) Growth- and breed-related changes of muscle fiber characteristics in cattle. Journal of Animal Science 78:1485-1496.

Wheeler, T.L. and Koohmaraie, M. (1994) Pre-rigor and post-rigor changes in tenderness of ovine longissimus muscle. Journal of Animal Science 72:1232- 1238.

Wheeler, T. L. and Koohmaraie, M. (1997) Effects of the adrenergic agonist L644, 969 on muscle protein turnover, endogenous proteinases activities, and meat tenderness in steers. Journal of Animal Science 70:3035−3043.

Wheeler, T.L., Shackelford, S.D. Casas, E., Cundiff, L.V. and Koohmaraie, M. (2001) The effects of Piedmontese inheritance and myostatin genotype on the palatability of longissimus thoracis, gluteus medius, semimembranosus, and biceps femoris. Journal of Animal Science 79:3069-3074.

White, S.N., Casas, E., Wheeler, T.L., Shackelford, S.D., Koohmaraie, M. , Riley, D. G., Chase, C.C. Jr., Johnson, D.D., Keele, J.W. and Smith, T.P.L. (2005) A new single nucleotide polymorphism in CAPN1 extends the current tenderness marker test to include cattle of Bos indicus, Bos taurus, and crossbred descent. Journal of Animal Science 83:2001–2008.

White, A., O’Sullivan, A., Troy, D. J. and O’Neill, E.E. (2006a) Effects of electrical stimulation, chilling temperature and hot-boning on the tenderness of bovine muscles. Meat Science 73:196-203.

White, A., O’Sullivan, A., Troy, D.J. and O’Neill, E.E. (2006b) Manipulation of the pre-rigor glycolytic behaviour of bovine M. longissimus dorsi in order to identify causes of inconsistencies in tenderness. Meat Science 73:151–156.

283

Wong, K.K., Tsang, Y.T.M., Shen, J., Cheng, R.S., Chang, Y.M., Man, T.K. and Lau, C.C. (2004) Allelic imbalance analysis by high‐density single‐nucleotide polymorphic allele (SNP) array with whole genome amplified DNA. Nucleic Acids Research 329:e69.

Xu, S. (1995) A comment on the simple regression method for interval mapping. Genetics 141:1657-1659.

Yan, J.B., Tang, J.H., Meng, Y.J., Ma, X.Q., Teng, W.T., Chander, S., Li L. and Li, J.H. (2006) Improving QTL Mapping Resolution Based on Genotypic Sampling—a Case Using a RIL Population. Acta Genetica Sinica 33(7):617–624.

Yang, J., Benyamin, B., McEvoy, B.P., Gordon, S., Henders, A.K., Nyholt, D.R., Madden, P.A., Heath, A.C., Martin, N.G., Montgomery, G.W., Goddard, M.E. and Visscher, P.M. (2010) Common SNPs explain a large proportion of the heritability for human height. Nature Genetics 42:565-569.

Yoshimura, N., Kikuchi, T., Sasaki, T., Kitahara, A., Hatanaka, M. and Murachi, T. (1983) Two distinct Ca2+ proteases (calpain I and calpain II) purified concurrently by the same method from rat kidney. Journal of Biological Chemistry 258:8883–8889.

Young, O.A. and Bass, J.J. (1984) Effect of castration on bovine muscle composition. Meat Science 11:139-156.

Young, P., Ferguson, C., Bañuelos, S. and Gautel, M. (1998) Molecular structure of the sarcomeric Z-disk: two types of titin interactions lead to an asymmetrical sorting of alpha-actinin. EMBO Journal 17:1614–1624.

Yu, L.P. and Lee, Y.B. (1986) Effects of postmortem pH and temperature on bovine muscle structure and meat tenderness. Journal of Food Science 51:774-780.

Zak, M., Baier, A., Bogdan, M. and Futschik, A. (2007) Locating multiple interacting quantitative trait loci using rank-based model selection. Genetics 176:1845-1854.

Zeng, Z.B. (1993) Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proceedings of the National Academy of Sciences 90:10972-10976.

Zhang, Z.R. Liu, Y.P., Jiang, X. Du, H.R. and Zhu, Q. (2008) Study on association of single nucleotide polymorphism of CAPN1 gene with muscle fibre and carcass traits in quality chicken populations. Journal of Animal Breeding and Genetics 125:258-264.

Zhang, L., Wang, S., Li, H., Deng, Q., Zheng, A., Li, S., Li, P., Li, Z. and Wang, J. (2010) Effects of missing marker and segregation distortion on QTL mapping in F2 populations. Theoretical and Applied Genetics 121:1071–1082.

284

Zhang, JM., Wei, Q., Zhao, X. and Paterson B. M. (1999) Coupling of the cell cycle and myogenesis through the cyclin D1-dependent interaction of MyoD with cdk4. EMBO Journal 18: 926 – 933.

Zhao, J., Boerwinkle, E. and Xiong-AM, M. (2005) An entropy-based statistic for genomewide association studies. American Journal of Human Genetics 77:27-40.

Zhou, G., Dudgeon, C., Li, M., Cao, Y., Zhang, L. and Jin, H. (2010) Molecular cloning of the HGD gene and association of SNPs with meat quality traits in Chinese red cattle. Molecular Biology Reports 37:603–611.

285

Appendices

286

Appendix 1. Sample preparation and measuring tenderness.

Samples of M. longissimus dorsi muscle (LD) and M. semitendinosus muscles (ST) were collected after slaughter, aged and then frozen at -20℃. Prior to the day for the measurement of shear force, the frozen steaks of the M. longissimus dorsi muscle

(LD) and M. semitendinosus muscles (ST) were thawed and trimmed into cubes of the size of 70 mm2 and 2.5 cm height with subcutaneous fat and perimysial connective tissue removed (Esmailizadeh et al., 2008). After the cubic samples were cooked in 70℃ water for 40 minutes and stored at 2℃ overnight, samples were sliced into 15 mm2 squares of 6 mm thickness to measure shear force with a

Lloyd LRX Materials Testing Machine and a 2500N load cell in triplicate (Lines et al., 2009). Samples of M. semitendinosus muscles (ST) were collected and prepared to estimate the total collagen content by quantifying the content of hydroxyproline

(Lines et al., 2009). A 20 g minced meat sample with fat and connective tissue removed was prepared by thawing the frozen samples for 24 hours. After dehydration by freeze drying for 2 to 5 days, 250 mg ground meat samples were hydrolysed in a hydrolysis vial with 10 ml of 6N HCL and flushed with nitrogen at 115℃ for 16 hours. 90 mm Whatman® No.1 filter paper was used to filter the hydrolysate. The filtered hydrolysate was adjusted to pH 6.0 with 6.5 M NaOH in a total volume of 100 ml. The quantity of trans-4-hydroxyproline-L-proline was measured at 558 nm using a Shimadzu UV-16A spectrophotometer.

287

Appendix 2. QTL mapping results for adjusted shear force without CAPN1.

Chromosome Trait Peak F Value 1 wbld_adjusted 36cM 3.71 1 wbst_adjusted 40cM 1.86 2 wbld_adjusted 64cM 2.43 2 wbst_adjusted 40cM 2.24 3 wbld_adjusted 84cM 0.84 3 wbst_adjusted 84cM 3.71 4 wbld_adjusted 32cM 2.29 4 wbst_adjusted 36cM 3.54 5 wbld_adjusted 92cM 4 5 wbst_adjusted 96cM 4.72 6 wbld_adjusted 116cM 1.79 6 wbst_adjusted 0cM 3.13 7 wbld_adjusted 0cM 1.02 7 wbst_adjusted 0cM 2.43 8 wbld_adjusted 72cM 1.59 8 wbst_adjusted 44cM 2.79 9 wbld_adjusted 48cM 2.12 9 wbst_adjusted 104cM 3.78 10 wbld_adjusted 16cM 2.46 10 wbst_adjusted 0cM 1.71 11 wbld_adjusted 60cM 2.26 11 wbst_adjusted 72cM 2.72 12 wbld_adjusted 64cM 0.9 12 wbst_adjusted 64cM 1.18 13 wbld_adjusted 44cM 3.38 13 wbst_adjusted 44cM 2.88 14 wbld_adjusted 68cM 1.68 14 wbst_adjusted 24cM 1.9 15 wbld_adjusted 80cM 1.38 15 wbst_adjusted 0cM 1.61 16 wbld_adjusted 76cM 2.55 16 wbst_adjusted 72cM 2.23 17 wbld_adjusted 84cM 1.61 17 wbst_adjusted 88cM 2.24 18 wbld_adjusted 36cM 4.33 18 wbst_adjusted 48cM 1.99 19 wbld_adjusted 16cM 2.07 19 wbst_adjusted 56cM 1.26

288

Appendix 2. QTL mapping results for adjusted shear force without CAPN1 (continued).

Chromosome Trait Peak F Value 20 wbld_adjusted 48cM 1.38 20 wbst_adjusted 72cM 2.3 21 wbld_adjusted 52cM 2.14 21 wbst_adjusted 68cM 2.15 22 wbld_adjusted 0cM 1.53 22 wbst_adjusted 44cM 1.61 23 wbld_adjusted 64cM 1.61 23 wbst_adjusted 68cM 1.31 24 wbld_adjusted 4cM 1.95 24 wbst_adjusted 0cM 0.87 25 wbld_adjusted 0cM 2.52 25 wbst_adjusted 0cM 3.14 26 wbld_adjusted 0cM 3.21 26 wbst_adjusted 68cM 1.63 27 wbld_adjusted 0cM 3.06 27 wbst_adjusted 0cM 2.2 28 wbld_adjusted 16cM 1.27 28 wbst_adjusted 20cM 0.67 29 wbld_adjusted 52cM 6.54 29 wbst_adjusted 52cM 6.89

289

Appendix 3. QTL mapping results for adjusted shear force with CAPN1-SNP316.

Chromosome Trait POS F Value 1 wbld_adjusted 40cM 3.96 1 wbst_adjusted 40cM 1.75 2 wbld_adjusted 60cM 2.28 2 wbst_adjusted 40cM 2.19 3 wbld_adjusted 84cM 0.7 3 wbst_adjusted 80cM 3.39 4 wbld_adjusted 32cM 1.67 4 wbst_adjusted 72cM 2.42 5 wbld_adjusted 92cM 4.22 5 wbst_adjusted 96cM 4.98 6 wbld_adjusted 116cM 1.46 6 wbst_adjusted 4cM 2.74 7 wbld_adjusted 40cM 0.81 7 wbst_adjusted 124cM 2.01 8 wbld_adjusted 72cM 1.64 8 wbst_adjusted 20cM 2.51 9 wbld_adjusted 104cM 1.9 9 wbst_adjusted 104cM 3.06 10 wbld_adjusted 16cM 2.92 10 wbst_adjusted 0cM 1.56 11 wbld_adjusted 64cM 2.54 11 wbst_adjusted 72cM 3.63 12 wbld_adjusted 64cM 1.07 12 wbst_adjusted 64cM 1.58 13 wbld_adjusted 44cM 2.54 13 wbst_adjusted 52cM 2.37 14 wbld_adjusted 64cM 2.09 14 wbst_adjusted 28cM 1.79 15 wbld_adjusted 36cM 0.96 15 wbst_adjusted 24cM 2.06 16 wbld_adjusted 80cM 2.42 16 wbst_adjusted 72cM 2.73 17 wbld_adjusted 80cM 1.47 17 wbst_adjusted 80cM 1.43 18 wbld_adjusted 40cM 4.43 18 wbst_adjusted 48cM 2.15 19 wbld_adjusted 12cM 2.09 19 wbst_adjusted 56cM 1.38

290

Appendix 3. QTL mapping results for adjusted shear force with CAPN1-SNP316 (continued).

Chromosome Trait Peak F Value 20 wbld_adjusted 20cM 1.52 20 wbst_adjusted 72cM 2.39 21 wbld_adjusted 56cM 2.05 21 wbst_adjusted 72cM 1.63 22 wbld_adjusted 0cM 1.6 22 wbst_adjusted 44cM 1.62 23 wbld_adjusted 68cM 1.9 23 wbst_adjusted 68cM 1.54 24 wbld_adjusted 4cM 2.26 24 wbst_adjusted 0cM 0.99 25 wbld_adjusted 0cM 2.88 25 wbst_adjusted 0cM 4.01 26 wbld_adjusted 0cM 2.77 26 wbst_adjusted 8cM 1.79 27 wbld_adjusted 0cM 3.12 27 wbst_adjusted 0cM 1.89 28 wbld_adjusted 16cM 1.28 28 wbst_adjusted 16cM 0.66 29 wbld_adjusted 28cM 3.91 29 wbst_adjusted 16cM 2.86

291

Appendix 4. QTL mapping results for adjusted shear force with CAPN1- SNP316 and CAPN1-SNP530.

Chromosome Trait POS F Value 1 wbld_adjusted 36cM 3.67 1 wbst_adjusted 40cM 1.84 2 wbld_adjusted 56cM 2.16 2 wbst_adjusted 36cM 2.17 3 wbld_adjusted 84cM 0.56 3 wbst_adjusted 80cM 3.38 4 wbld_adjusted 32cM 1.86 4 wbst_adjusted 72cM 2.35 5 wbld_adjusted 92cM 4.05 5 wbst_adjusted 96cM 5.01 6 wbld_adjusted 116cM 1.5 6 wbst_adjusted 4cM 2.74 7 wbld_adjusted 40cM 0.78 7 wbst_adjusted 124cM 1.96 8 wbld_adjusted 72cM 1.67 8 wbst_adjusted 20cM 2.41 9 wbld_adjusted 104cM 1.77 9 wbst_adjusted 104cM 2.95 10 wbld_adjusted 16cM 2.82 10 wbst_adjusted 0cM 1.54 11 wbld_adjusted 64cM 2.28 11 wbst_adjusted 72cM 3.78 12 wbld_adjusted 64cM 1.06 12 wbst_adjusted 64cM 1.6 13 wbld_adjusted 44cM 2.47 13 wbst_adjusted 52cM 2.33 14 wbld_adjusted 68cM 2 14 wbst_adjusted 28cM 1.75 15 wbld_adjusted 88cM 0.95 15 wbst_adjusted 24cM 2.1 16 wbld_adjusted 80cM 2.52 16 wbst_adjusted 72cM 2.62 17 wbld_adjusted 80cM 1.37 17 wbst_adjusted 0cM 1.41 18 wbld_adjusted 40cM 4.24 18 wbst_adjusted 48cM 2.05 19 wbld_adjusted 12cM 2.14 19 wbst_adjusted 56cM 1.34

292

Appendix 4. QTL mapping results for adjusted shear force with CAPN1- SNP316 and CAPN1-SNP530 (continued).

Chromosome Trait Peak F Value 20 wbld_adjusted 20cM 1.46 20 wbst_adjusted 72cM 2.37 21 wbld_adjusted 56cM 2.01 21 wbst_adjusted 72cM 1.54 22 wbld_adjusted 0cM 1.59 22 wbst_adjusted 44cM 1.63 23 wbld_adjusted 64cM 1.98 23 wbst_adjusted 68cM 1.54 24 wbld_adjusted 4cM 2.48 24 wbst_adjusted 0cM 0.99 25 wbld_adjusted 0cM 2.87 25 wbst_adjusted 0cM 3.87 26 wbld_adjusted 0cM 2.85 26 wbst_adjusted 8cM 1.72 27 wbld_adjusted 0cM 3.34 27 wbst_adjusted 0cM 1.79 28 wbld_adjusted 16cM 1.31 28 wbst_adjusted 16cM 0.65 29 wbld_adjusted 28cM 4.15 29 wbst_adjusted 16cM 3.08

293

Appendix 5. QTL mapping results for ageing rate without MSTN F94L and CAPN1.

Chromosome Traits Peak F Value 1 logwbst1_26 48cM 4.5 1 logwbld1_26 64cM 3.13 2 logwbst1_26 0cM 1.24 2 logwbld1_26 56cM 1.84 3 logwbst1_26 76cM 4.1 3 logwbld1_26 44cM 3.09 4 logwbst1_26 52cM 1.06 4 logwbld1_26 40cM 5.06 5 logwbst1_26 36cM 1.08 5 logwbld1_26 8cM 2.56 6 logwbst1_26 72cM 3.34 6 logwbld1_26 0cM 2.24 7 logwbst1_26 116cM 4.5 7 logwbld1_26 56cM 1.68 8 logwbst1_26 72cM 1.65 8 logwbld1_26 4cM 1.09 9 logwbst1_26 52cM 2.89 9 logwbld1_26 28cM 2.17 10 logwbst1_26 96cM 0.91 10 logwbld1_26 48cM 2.77 11 logwbst1_26 80cM 4.18 11 logwbld1_26 24cM 1.26 12 logwbst1_26 12cM 1.21 12 logwbld1_26 12cM 2.6 13 logwbst1_26 0cM 1.27 13 logwbld1_26 0cM 4.39 14 logwbst1_26 40cM 1.87 14 logwbld1_26 16cM 1.38 15 logwbst1_26 0cM 1.37 15 logwbld1_26 0cM 0.67 16 logwbst1_26 84cM 0.74 16 logwbld1_26 0cM 2.21 17 logwbst1_26 88cM 0.94 17 logwbld1_26 80cM 0.67 18 logwbst1_26 80cM 2.63 18 logwbld1_26 52cM 0.78 19 logwbst1_26 44cM 2.88 19 logwbld1_26 28cM 1.35

294

Appendix 5. QTL mapping results for ageing rate without MSTN F94L and CAPN1(continued).

Chromosome Trait Peak F Value 20 logwbst1_26 56cM 4.03 20 logwbld1_26 4cM 1.07 21 logwbst1_26 32cM 1.37 21 logwbld1_26 0cM 1.74 22 logwbst1_26 56cM 2.49 22 logwbld1_26 76cM 1.07 23 logwbst1_26 68cM 2.01 23 logwbld1_26 60cM 0.56 24 logwbst1_26 0cM 1.76 24 logwbld1_26 0cM 2.86 25 logwbst1_26 0cM 0.59 25 logwbld1_26 20cM 1.44 26 logwbst1_26 0cM 0.93 26 logwbld1_26 60cM 0.98 27 logwbst1_26 0cM 2.07 27 logwbld1_26 32cM 2.28 28 logwbst1_26 20cM 3.43 28 logwbld1_26 52cM 0.33 29 logwbst1_26 4cM 3.2 29 logwbld1_26 0cM 0.77

295

Appendix 6. QTL mapping results for ageing rate with MSTN F94L.

Chromosome Traits POS F Value 1 logwbst1_26 48cM 4 1 logwbld1_26 48cM 2.99 2 logwbst1_26 4cM 1.17 2 logwbld1_26 56cM 1.62 3 logwbst1_26 76cM 3.34 3 logwbld1_26 44cM 3.02 4 logwbst1_26 0cM 0.98 4 logwbld1_26 40cM 4.84 5 logwbst1_26 36cM 1.04 5 logwbld1_26 8cM 2.49 6 logwbst1_26 72cM 2.86 6 logwbld1_26 0cM 1.9 7 logwbst1_26 116cM 4.18 7 logwbld1_26 56cM 1.73 8 logwbst1_26 72cM 1.34 8 logwbld1_26 4cM 1.31 9 logwbst1_26 52cM 3.47 9 logwbld1_26 28cM 2.2 10 logwbst1_26 96cM 1.03 10 logwbld1_26 0cM 2.59 11 logwbst1_26 80cM 3.42 11 logwbld1_26 24cM 1.16 12 logwbst1_26 16cM 1.4 12 logwbld1_26 12cM 2.6 13 logwbst1_26 0cM 1.12 13 logwbld1_26 0cM 4.63 14 logwbst1_26 68cM 1.44 14 logwbld1_26 4cM 1.34 15 logwbst1_26 0cM 1.42 15 logwbld1_26 0cM 0.64 16 logwbst1_26 68cM 0.64 16 logwbld1_26 0cM 2.48 17 logwbst1_26 48cM 0.88 17 logwbld1_26 80cM 0.78 18 logwbst1_26 80cM 2.74 18 logwbld1_26 52cM 0.85 19 logwbst1_26 44cM 3.25 19 logwbld1_26 28cM 1.39

296

Appendix 6. QTL mapping results for ageing rate with MSTN F94L (continued).

Chromosome Trait Peak F Value 20 logwbst1_26 56cM 3.98 20 logwbld1_26 4cM 0.95 21 logwbst1_26 32cM 1.17 21 logwbld1_26 4cM 2.14 22 logwbst1_26 56cM 2.68 22 logwbld1_26 76cM 0.84 23 logwbst1_26 68cM 1.6 23 logwbld1_26 68cM 0.6 24 logwbst1_26 0cM 1.69 24 logwbld1_26 0cM 2.81 25 logwbst1_26 0cM 0.61 25 logwbld1_26 20cM 1.4 26 logwbst1_26 0cM 1.19 26 logwbld1_26 60cM 1.03 27 logwbst1_26 0cM 2.25 27 logwbld1_26 32cM 2.6 28 logwbst1_26 20cM 3.24 28 logwbld1_26 52cM 0.34 29 logwbst1_26 4cM 2.89 29 logwbld1_26 0cM 0.91

297

Appendix 7. QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316.

Chromosome Traits POS F Value 1 logwbst1_26 48cM 4.55 1 logwbld1_26 52cM 3.95 2 logwbst1_26 0cM 1.56 2 logwbld1_26 104cM 1.78 3 logwbst1_26 76cM 3.34 3 logwbld1_26 84cM 3.25 4 logwbst1_26 48cM 0.89 4 logwbld1_26 40cM 4.95 5 logwbst1_26 36cM 1.02 5 logwbld1_26 8cM 2.53 6 logwbst1_26 72cM 2.92 6 logwbld1_26 116cM 2.35 7 logwbst1_26 120cM 2.98 7 logwbld1_26 52cM 2 8 logwbst1_26 120cM 1.43 8 logwbld1_26 0cM 1.14 9 logwbst1_26 48cM 2.96 9 logwbld1_26 28cM 2.47 10 logwbst1_26 68cM 1.11 10 logwbld1_26 0cM 2.72 11 logwbst1_26 80cM 2.6 11 logwbld1_26 24cM 1.19 12 logwbst1_26 16cM 1.02 12 logwbld1_26 16cM 2.2 13 logwbst1_26 72cM 1.29 13 logwbld1_26 0cM 5.4 14 logwbst1_26 68cM 2.43 14 logwbld1_26 0cM 1.2 15 logwbst1_26 0cM 1.77 15 logwbld1_26 24cM 0.49 16 logwbst1_26 52cM 0.86 16 logwbld1_26 4cM 2.96 17 logwbst1_26 44cM 0.85 17 logwbld1_26 80cM 0.87 18 logwbst1_26 80cM 2.01 18 logwbld1_26 52cM 1.19 19 logwbst1_26 48cM 4.64 19 logwbld1_26 28cM 0.67

298

Appendix 7. QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316 (continued).

Chromosome Trait Peak F Value 20 logwbst1_26 56cM 3.07 20 logwbld1_26 8cM 0.99 21 logwbst1_26 32cM 1.46 21 logwbld1_26 4cM 1.9 22 logwbst1_26 56cM 2.8 22 logwbld1_26 76cM 1.07 23 logwbst1_26 8cM 1.26 23 logwbld1_26 44cM 0.72 24 logwbst1_26 0cM 1.24 24 logwbld1_26 0cM 2.71 25 logwbst1_26 8cM 0.65 25 logwbld1_26 20cM 1.43 26 logwbst1_26 4cM 0.72 26 logwbld1_26 64cM 0.61 27 logwbst1_26 0cM 1.8 27 logwbld1_26 28cM 2.43 28 logwbst1_26 24cM 3.64 28 logwbld1_26 36cM 0.15 29 logwbst1_26 60cM 2.08 29 logwbld1_26 60cM 0.85

299

Appendix 8. QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530.

Chromosome Traits POS F Value 1 logwbst1_26 48cM 3.92 1 logwbld1_26 52cM 2.96 2 logwbst1_26 0cM 1.26 2 logwbld1_26 104cM 2.29 3 logwbst1_26 76cM 3.35 3 logwbld1_26 84cM 3.22 4 logwbst1_26 48cM 0.94 4 logwbld1_26 40cM 4.82 5 logwbst1_26 36cM 0.88 5 logwbld1_26 8cM 2.45 6 logwbst1_26 76cM 2.56 6 logwbld1_26 116cM 2.28 7 logwbst1_26 120cM 3.17 7 logwbld1_26 56cM 1.74 8 logwbst1_26 120cM 1.33 8 logwbld1_26 0cM 1.19 9 logwbst1_26 48cM 3.27 9 logwbld1_26 28cM 2.61 10 logwbst1_26 68cM 1.12 10 logwbld1_26 0cM 2.98 11 logwbst1_26 28cM 2.43 11 logwbld1_26 8cM 0.99 12 logwbst1_26 16cM 1.04 12 logwbld1_26 16cM 1.88 13 logwbst1_26 72cM 0.98 13 logwbld1_26 0cM 6.03 14 logwbst1_26 68cM 2.36 14 logwbld1_26 0cM 0.85 15 logwbst1_26 0cM 1.35 15 logwbld1_26 24cM 0.34 16 logwbst1_26 52cM 0.99 16 logwbld1_26 4cM 3.02 17 logwbst1_26 44cM 1.03 17 logwbld1_26 80cM 1.31 18 logwbst1_26 80cM 2.53 18 logwbld1_26 52cM 1.05 19 logwbst1_26 44cM 4.74 19 logwbld1_26 32cM 0.8

300

Appendix 8. QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (continued).

Chromosome Trait Peak F Value 20 logwbst1_26 56cM 3.59 20 logwbld1_26 8cM 1.29 21 logwbst1_26 32cM 1.61 21 logwbld1_26 4cM 2.3 22 logwbst1_26 52cM 2.3 22 logwbld1_26 76cM 0.92 23 logwbst1_26 8cM 1.43 23 logwbld1_26 56cM 0.71 24 logwbst1_26 0cM 1.19 24 logwbld1_26 0cM 2.73 25 logwbst1_26 0cM 0.6 25 logwbld1_26 24cM 1.36 26 logwbst1_26 12cM 0.67 26 logwbld1_26 68cM 0.47 27 logwbst1_26 0cM 1.72 27 logwbld1_26 28cM 1.69 28 logwbst1_26 20cM 3.66 28 logwbld1_26 52cM 0.19 29 logwbst1_26 60cM 3.41 29 logwbld1_26 0cM 0.52

301

Appendix 9. QTL mapping results for adjusted shear force with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (for each family).

Chromosome Traits 361 t value 368 t value 398 t value 1 wbld_adjusted -0.3896 2.2269 -0.3856 2.4245 1 wbst_adjusted 2 wbld_adjusted 2 wbst_adjusted 3 wbld_adjusted 3 wbst_adjusted 0.276 2.9744 4 wbld_adjusted -0.3115 2.1947 4 wbst_adjusted 0.2039 2.3406 5 wbld_adjusted 0.3922 2.8505 5 wbst_adjusted -0.193 2.3781 0.2542 3.031 6 wbld_adjusted 6 wbst_adjusted 0.2388 2.6728 7 wbld_adjusted 7 wbst_adjusted -0.2018 2.0628 8 wbld_adjusted 8 wbst_adjusted 0.1853 2.2695 9 wbld_adjusted 9 wbst_adjusted -0.2104 2.8157 10 wbld_adjusted -0.4235 2.8518 10 wbst_adjusted 11 wbld_adjusted -0.3504 2.3571 11 wbst_adjusted -0.2207 2.8085 12 wbld_adjusted 12 wbst_adjusted -0.1892 2.0709 13 wbld_adjusted -0.3094 2.3585 13 wbst_adjusted -0.2024 2.5486 14 wbld_adjusted -0.2898 2.0571 14 wbst_adjusted 15 wbld_adjusted 15 wbst_adjusted 0.2309 2.2567 16 wbld_adjusted 0.2991 2.2806 16 wbst_adjusted 0.2086 2.4356 17 wbld_adjusted 17 wbst_adjusted 0.1899 2 18 wbld_adjusted 0.4976 3.399 0.1647 2.0191 18 wbst_adjusted 19 wbld_adjusted 0.3518 2.4555

302

Appendix 9. QTL mapping results for adjusted shear force with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (for each family) (continued).

Chromosome Traits 361 t value 368 t value 398 t value 19 wbst_adjusted 20 wbld_adjusted 20 wbst_adjusted 21 wbld_adjusted 21 wbst_adjusted 22 wbld_adjusted 22 wbst_adjusted 23 wbld_adjusted -0.3574 2.3419 23 wbst_adjusted 24 wbld_adjusted 0.3192 2.1739 24 wbst_adjusted 25 wbld_adjusted -0.3521 2.1493 25 wbst_adjusted 0.2183 2.7503 26 wbld_adjusted 0.3951 2.7227 26 wbst_adjusted 27 wbld_adjusted 0.424 2.9631 27 wbst_adjusted 0.1905 2.3041 28 wbld_adjusted 28 wbst_adjusted 29 wbld_adjusted -0.4269 2.5395 29 wbst_adjusted -0.2363 2.718

303

Appendix 10A. QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (for each family).

Chromosome Traits 361 t value 368 t value 398 t value 1 logwbst1_26 -0.0739 2.4244 0.1121 2.3729 1 logwbld1_26 -0.1008 2.7552 2 logwbst1_26 2 logwbld1_26 -0.0826 2.4305 3 logwbst1_26 -0.0737 2.8532 3 logwbld1_26 0.1008 2.4434 4 logwbst1_26 4 logwbld1_26 -0.1198 3.6088 5 logwbst1_26 5 logwbld1_26 6 logwbst1_26 -0.0732 2.5189 6 logwbld1_26 0.1781 2.4278 7 logwbst1_26 0.0689 2.4209 7 logwbld1_26 8 logwbst1_26 8 logwbld1_26 9 logwbst1_26 -0.0884 2.9081 9 logwbld1_26 0.0742 2.1907 10 logwbst1_26 10 logwbld1_26 -0.0742 2.2541 11 logwbst1_26 0.0654 2.4471 11 logwbld1_26 12 logwbst1_26 12 logwbld1_26 -0.1137 2.2755 13 logwbst1_26 13 logwbld1_26 0.0889 2.6278 -0.0661 2.0676 0.1 2.6136 14 logwbst1_26 14 logwbld1_26 15 logwbst1_26 0.0543 2.0051 15 logwbld1_26 16 logwbst1_26 16 logwbld1_26 -0.1061 2.8866 17 logwbst1_26 17 logwbld1_26 18 logwbst1_26 0.0566 2.1178 18 logwbld1_26 19 logwbst1_26 0.1139 3.5576

304

Appendix 10A. QTL mapping results for ageing rate with MSTN F94L and CAPN1-SNP316 and CAPN1-SNP530 (for each family) (continued).

Chromosome Traits 361 t value 368 t value 398 t value 19 logwbld1_26 20 logwbst1_26 20 logwbld1_26 21 logwbst1_26 0.0916 3.1541 21 logwbld1_26 0.0877 2.5949 22 logwbst1_26 0.0604 2.3122 22 logwbld1_26 23 logwbst1_26 23 logwbld1_26 24 logwbst1_26 24 logwbld1_26 0.0865 2.6949 25 logwbst1_26 25 logwbld1_26 26 logwbst1_26 26 logwbld1_26 27 logwbst1_26 -0.0637 2.1299 27 logwbld1_26 28 logwbst1_26 -0.0814 3.1242 28 logwbld1_26 29 logwbst1_26 0.1026 3.0878 29 logwbld1_26

305

Appendix 10B. Ageing rate QTL graphs

A) Without MSTN F94L + CAPN1 genotypes. B) With MSTN F94L + CAPN1 genotypes.

QTL on BTA 1 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model. (Blue line: ST muscle; Green line: LD muscle)

A) Without MSTN F94L + CAPN1 genotypes. B) With MSTN F94L + CAPN1 genotypes.

QTL on BTA 4 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model. (Blue line: ST muscle; Green line: LD muscle)

A) Without MSTN F94L + CAPN1 genotypes. B) With MSTN F94L + CAPN1 genotypes.

QTL on BTA 7 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model. (Blue line: ST muscle; Green line: LD muscle)

306

A) Without MSTN F94L + CAPN1 genotypes. B) With MSTN F94L + CAPN1 genotypes.

QTL on BTA 19 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model. (Blue line: ST muscle; Green line: LD muscle)

A) Without MSTN F94L + CAPN1 genotypes. B) With MSTN F94L + CAPN1 genotypes.

QTL on BTA 20 for ageing rate for LD and ST muscles without and with the MSTN F94L, CAPN1-SNP316 and SNP530 genotypes as fixed factors in the QTL model. (Blue line: ST muscle; Green line: LD muscle)

307

Appendix 11. 50x TAE buffer formula.

1. Weigh the ingredients (TRIZMA base buffer: 242 g, EDTA (disodium salt):

18.62 g)

2. Mix ingredients with glacial acetic acid (57.1ml) to dissolve the ingredients.

3. Add MilliQ water to 1 litre.

4. Adjust pH to 8

Appendix 12. Purification kit protocol (Ultra Clean PCR Clean-up Kit, MoBio).

1. Spin the tubes containing the PCR products shortly with a small centrifuge.

2. Prepare new tubes (labelled) and remove the PCR products to new tubes.

3. Add 250 µl of SpinBind into the new tubes with PCR products and then

vortex and spin them shortly.

4. Prepare new spin columns (labelled) and transfer the mixture of SpinBind

and PCR products into the spin columns.

5. Spin down with the maximum speed for 30 sec.

6. Discard flow-through.

7. Add 300 µl of SpinClean buffer into the spin columns and then spin them

down for 30 sec.

8. Discard flow-through.

9. Move upper tubes to new lower tubes.

10. Add 30 µl of elution buffer into the spin columns.

11. Discard the upper tubes and draw a line for the level of elution.

12. Evaporate the elution to half in a 60 ºC oven.

308

Appendix 13. Purification of sequencing PCR products.

1. Transfer the sequencing PCR products to new tubes.

2. Add 80 µl of 75% isopropanol into each tube.

3. Vortex them briefly and leave them at room temperature for 15 min.

4. Spin the tubes in 13,000 rpm for 20 min. (Keep it as cool as possible, 4 ºC)

5. Remove the supernatant from the tubes.

6. Add 250 µl of 75% isopropanol into each tube.

7. Spin the tubes in 13,000 rpm for 20 min. (Keep it as cool as possible, 4 ºC)

8. Remove the supernatant from the tubes.

9. Dry the sample in the incubator at 37 ºC.

309

Appendix 14. PCR conditions for primer sets.

Marker Forward Primer Reverse Primer name Mg2+ DMSO Program MYL7- ACTGCCCACTCTTCCTGCT CAGCCCCCTTTATGTCTCAG 5S 1.5 0 TD1# MYL7- AATGTGGCTGGAATCTGCTC CTACGGGGAAGAAGGGATGT E1 2.5 1 TD2 MYL7- CCCAGATCCAGGAGTTCAAG GTGTGAGCCAGGACAGGAGT E2 2.5 0 TD1 MYL7- ACTCCTGTCCTGGCTCACAC ACTGGCTCTGGGTGGGATT E3 2.5 0 TD1 MYL7- CCAAGCCCAGCATTTCAG CAGCACCTCCTCCAGTAAGC E4,5 2.5 1 TD2 MYL7- GCTTACTGGAGGAGGTGCTG TCTGTAACCTGGGCACAACA E6 2.5 1 TD2 MYL7- TGTTGTGCCCAGGTTACAGA TTGCCAAGTCCCAGGTAAAC 3S NA* NA NA

MYO1G AGAACCGTCCCGTCCTCA CTGTGATCTCACCCTCACCC -5S 2.5 0 TD1 MYO1G TCACAGAGGGTCGGAGTTTC CAAGAGGAAGGCTCAGGAGA -E1 2.5 1 TD2 MYO1G CCACACACTTCTCTGCCCTC CCTCGCCTCTACTTGGGTAG -E2,3 2.5 0 TD2 MYO1G CTCCTCCTGTGATTGGGAAG TACACACACTGCCATGTCCA -E4,5 2.5 0 TD2 MYO1G TCACCGTGCTGACTCATACC CCCTGATTCCTCCAACTGAA -E6 2.5 0 TD2 MYO1G AATGGCTCAGCAGGACTTGT TTTTCTTTCAAGGAGCAAGCA -E7 2.5 1 TD2 MYO1G GAGTTCGTGGAGCTGGAGAG AGCTGAAGCAGGCTTGAAGA -E8,9 2.5 0 TD1 MYO1G TTGGACACAACTGAGCGACT TCTAGGGGAGAAAGGGGAGA -E10,11 NA 0 NA MYO1G TCTCCCCTTTCTCCCCTAGA AGTCTCCAGCCGCTCTATCA -E12 NA 0 NA MYO1G AAGCCAGGATTCAGTGATGC GTGAGGGTCAGACCGAGTGT -E13 2.5 0 TD2 MYO1G ACACTCGGTCTGACCCTCAC GGCATCGTAACCTGAGCAGT -E14 2.5 0 TD1 MYO1G CCCTGAGAGATCCTCTACCC CTAGTCCCACCCCTTGTGTC -E15 2.5 0 TD1 MYO1G GAGTGGCTTGGAGGAAAGG GTGTCTGTGGTCCCCTCAGT -E16 NA 0 TD1 MYO1G TGGCAACCCACTCCAGTATT TGAGGTGTAGGCTGTCAACG -E17 2.5 0 TD1 MYO1G CTGGGGTAAAGGACGTGAAA AGTGTTGGGTTCTGGACAGG -E18 2.5 0 TD1 MYO1G GGATAGCTGTGCGGATTCAT GGATAGCTGTGCGGATTCAT -E19,20 2.5 0 TD1 MYO1G AAAGAGGCTTTCCGAAGAGG GCCGTGGCTGGAGACTATAC -E21,22 2.5 0 TD2 MYO1G GGTTTCCGACTGTCCCTACA TTCCGAATCAAGCTGCTCTT -3S NA NA NA

MBNL3 GCTGTGCAACTCCTCTTCATC CCTGATTGCTCCCTTCACAT -E1,E2 2.5 0 TD1 MBNL3 ACTCCATCAGTGCGATTTGA ATGCCCACCAGACTCTCTGT -E3 2.5 0 TD1 MBNL3 CAAACTTCCTTTGGGGACAA AAAAATGGGAGAAAAAGCACA -E4 2.5 0 TD1 MBNL3 GCCCAGGCCAGACTTTATTT TTTTTGCAATGCTGAAGCTC -E5 2.5 0 TD1

310

Appendix 14. PCR conditions for primer sets (continued).

Forward Primer Reverse Primer Marker name Mg2+ DMSO Program

TGGATATGCATGTGCTGGTT AAAACTGCCCATTCTGGCTA MBNL3-E6 2.5 0 TD1 MBNL3- TTGGCTTTCCCCTTTTCTCT TCACTGGGTCATGATAAGCA E7,E8 2.5 0 TD1

TGAGCAGTCCTTCACTTGGA TGGCCTAAGTGTGAATTCTGG MBNL3-E9 2.5 0 TD1

GGTGAAACTGGACCACCATC TACAGAGTGCAGGGGGAGTC CAPN4-5S 2.5 0 TD1

GTCTTAATGTCGGCGCTAGG AATTGAGAGGACGGATCGTG CAPN4-E1 2.5 0 TD1

GGGGTCTAGGGAATGTGCTT TATTGCTGGACAAGCCTCCT 2CAPN4-E2 2.5 1 TD2 2CAPN4- GCAAGGATAGGGACAGGTGA TGATGAAGGAGGATGGGAAC E3,4 2.5 1 TD1

GAGGCCAATGAGAGTGAGGA TGTCCATAGGTGGGTGTGTG CAPN4-E5,6 2.5 0 TD1 CAPN4- TGTCCATAGGTGGGTGTGTG CACCTCAGTGTCAGGCTCAA E7,8,9 2.5 0 TD1

GAGTGGGGACCACACAGACT GACAGTGAGGAGCAGGGAAC CAPN4-E10 2.5 0 TD1

TTCCCTGCTCCTCACTGTCT CCCACAGATGGTCCAGACTT CAPN4-E11 2.5 0 TD1 Fixed anneal CCAGAGCTACACCGACGAG TACAGAGTGCAGGGGGAGTC R2CAPN4-E1 2.5 0 56 ºC

GGGGTCTAGGGAATGTGCTT TATTGCTGGACAAGCCTCCT RCAPN4-E2 2.5 0 64 2RCAPN4- E2- F+PCAPN4- TAGGGAATGTGCTTGGAGG GGTTCCGCCGCCGCCA E-R1 2.5 0 60

LOC536988- CCCCCAAATCCTCTCCTACT CCCCCAGCTGGTATGAACTA E1 1.5 1 TD1 LOC536988- CATTCCTCTTCCTGCTTTGC CCCCTCCACTTTCTCTGTCA E2 2.5 0 TD2 2LOC536988- TGGACACTGAGTCCTTGCTG GCCCAGGCTGTCATTAAGTC E3 2.5 1 TD2 2LOC536988- AGAGGCCCGGAGATAAGAAG TAGACCAATGGGTGGAGAGG E4 2.5 0 TD1 LOC536988- TTGTGTGACCTGCTGTGTGA GCTTGGCTGATGACTCTCG E5 2.5 0 TD2 LOC536988- GTCAGTCCCTAACCGCAGAG CTCCTCCCTACCTCCAGACC E6 1.5 0 TD1 LOC536988- GCAAGAGTGGGCAGAGAAGG GTCCAAGGTCACACGACAAG E7 1.5 0 TD1 LOC536988- TTACAAAGCAGGAGACTGAGG GTGACAGGACCACGCAAAC E8 1.5 0 TD1 LOC536988- CATCCATCATCTCCCCAAGT GCAATCCTCCCCTTTTCTGT E9 2.5 0 TD2 LOC536988- GGCAGAGGCTGGTCTACTTC GATGAGAGAGGCTGGGACAG E10 1.5 0 TD1 2LOC536988- GATGAGAGAGGCTGGGACAG AGGAGAAGGGAGACCAGAGC E11 2.5 0 TD1 LOC536988- GAGGGAGGGTTCAGGGAAG GAAGGACACACGGGAGACTT E12 1.5 0 TD1 LOC536988- AAAGATGGGGAAACCAGAGC GGTCACACAGCAGGTCAGTG E13-1 1.5 0 TD1 LOC536988- CACTGACCTGCTGTGTGACC TGGGAATGAAACGGAACCTA E13-2 1.5 0 TD1

311

Appendix 14. PCR conditions for primer sets (continued).

Marker Forward Primer Reverse Primer name Mg2+ DMSO Program RCAPN5 GGATGACCTTCGAGGACTTG CACGTCTTGTGGATGCTCAG -E7-2 3.5 0 TD2 Fixed RCAPN5 anneal ACGTCTTACCTGAGCATCCA GAAGCTGTCCTTGTGGTTGA -E7-3 3.5 0 58 ºC RCAPN5 ATCAAGTGTGAGGGGGACAA CACATCATACTCGGGTGTCG -E11 3.5 0 TD2

LOXL1- CCAGAGGCCTCCTTCCTACT AGCAGCACCCGGGAATTA E1-1 2.5 1 TD2 LOXL1- AGGGCTCCGAGGCTAATTC ATGTTGCCATAGGGTGGGTA E1-2 2.5 1 TD2 LOXL1- CGACCAGGGCTACGTGTACT GTGGGGGTCTTGACTTAGCA E1-3 2.5 1 TD1 LOXL1- TACTCCCTGACGTCCCTGAC AAAAGCTGGATCTCCCCATT E2 2.5 0 TD2 LOXL1- TCATTATCAGGCCCTGCTTT CCTTGTCCCAGTGATGGATT E3 2.5 1 TD2 2LOXL1- GGAGGAGTGAGAGGTCCTGA CTCAGGACTGGCGAGAAGAC E4 2.5 0 TD2 LOXL1- GGACAAAGCACAGAGGAAGG CGTGATTGCCCCATTTAGAG E5 2.5 0 TD2 LOXL1- ATTGGTGCGTCTCTTGGTCT CAGGGCAGGTATCCTCTGTG E6 2.5 0 TD2 LOXL1- GGCTGACAGTCGAATGAGGT CAAAATAACGTGGTAGGGGAGA E7 2.5 0 TD2 Fixed R2LOXL anneal CGACCAGGGCTACGTGTACT GCTGGAAGGGGTAGACGAC 1-E1-2 2.5 0 56 ºC

TATCACTGATGTCAAACCTG ACTCAGGCACCAAATAGCTG LOX-i5 3.5 0 TD2 *NA = not available as PCR could not be optimised; # = touchdown PCR program

312

Appendix 15. Candidate genes and map locations.

Gene name Description/ Functionality Map location

Family of motor proteins; one of Myosin regulatory 79 cM on myosin light chains need for muscle light chain 7 (MYL7) chromosome 4 structure stability

Member of myosin 1 family; 79 cM on Myosin IG (MYO1G) involved in lysosome transport chromosome 4

Inhibits terminal muscle Muscleblind-like 3 116 cM on differentiation; associated with (MBNL3) chromosome 17 myotonia

Component of both large calpain Calpain, small subunits; calpain is a multiplex 41.9 cM on subunit 1 (CAPN4 or enzyme known as a major factor chromosome 18 CPNS1) contributing to the tenderization

One the large subunits of calpain LOC536988 gene 49 cM on without the calmodulin-like domain (calpain 5) chromosome 15 IV

Member of the lysyl oxidase family; similar function to LOX which plays Lysyl oxidase-like 1 an important role in forming cross- 33 cM on (LOXL1) links which stabilize the structure of chromosome 21 collagen and elastin in the extracellular matrix

313

Appendix 16. Candidate gene variants: locations and sequences.

Variants Genotype Location Sequencce MYL7 SNP1 C/G Intron 2 TCACACCAGC/GGACCCGGCTCAGAGCT MYL7 SNP2 G/T Intron 2 GCTGCCTGG/TCCCGGCCTGTACTCCT MYO1G SNP1 C/T 5’ AGCAGT/CTCCTGTC MYO1G SNP2 C/T 5’ CTGTCCTC/TACTACTAGA MYO1G SNP3 G/A 5’ CAAAG/AGCCTGGGTATCAGCAT MYO1G SNP4 C/T Intron 1 TCCCTGTC/TCCCATCCCCAA MYO1G SNP5 G/A Exon 7 AGAGCCACCACC A/GGGCAGTGACG MYO1G SNP6 C/T Intron 7 C/TTCCCTCTGCTGAAACCCA MYO1G SNP7 G/T Intron 13 GCCAAGCTTGG/TGTGGGGGC MBNL3 insertion/deletions 15T~17T Intron 3 GTGATTTTTTTTTTTTTTTTAAGGTTT LOXL1 SNP1 A/G Exon 1 CCGCAGCGCGA/GGCGGCGGCCT LOXL1 SNP2 C/T Intron1 TC/TCTTCCCACTCCTCACT LOXL1 SNP3 G/A Intron2 GTGGGGAAGGG/AAATGGGGAGAT LOX SNP1 T/C Intron 5 AGAACTACTTT/CTAAACCAAC LOX SNP2 A/T Intron 5 AGATATATTA/TAAAAAATTCTT CAPN4 insertion/deletion AC 5’ UTR ACGCGCGAACGAGGCGGTGAG CAPN4 SNP1 A/C Exon 2 GGCGGAGGCGGA/CGGC CAPN4 SNP2 A/C Exon 2 GGCGGC/AGGCGGCG CAPN4 SNP3 A/C Exon 2 GCGGCGGC/AGGCG CAPN4 SNP4 T/C Exon 2 GCGGT/CGGCGGCGGCGGAACCGCCAT CAPN4 3 base repeat insertion/deletion GCC Exon 2 GGCGGCGGCGGAACCGCCATGCGCATC CAPN4 SNP5 T/C Exon 1 GCCATT/CAGGTAAGGCGGG CAPN4 SNP6 T/C Exon 10 TGCAGTCACATCTTC/TGTGGGTCCTGCTG CAPN5 SNP1 G/A 5’ UTR GCCGGAGCCGG/AAGGCCACTAAGTGG CAPN5 SNP2 G/A 5’ UTR GCATCCCCG/AAATCCGCTGGA CAPN5 SNP3 T/C Intron 1 GGCCCCAGT/CTGTCCAGGGC CAPN5 SNP4 T/C Exon 2 GCTCCCAC/TGACCTGCA CAPN5 SNP5 T/C Exon 2 TCATCCCTC/TGCTTCCCGCGAGTCAC CAPN5 SNP6 T/C Exon4 CACGGCGGAT/CGCGTTGGT CAPN5 SNP7 T/C Exon 4 GGCCTCATCAGT/CGCCTCCATC CAPN5 SNP8 T/C Intron 4 CCTCCATCAAGGTGAGC/TGC CAPN5 SNP9 C/G Exon 5 GGCAGTGACAGCC/GGCAGACATGGA CAPN5 SNP10 C/T Exon 5 CAAGGTGCGCCTGGGCCAC/T CAPN5 SNP11 G/T Exon5 GGCCTG/TCTGGCCTTCTTCA CAPN5 SNP12 C/G Intron 5 TGC/GGGCGGTGAAAGGGGAAGCAC CAPN5 SNP13 T/A Intron 5 GCACCAGGTGGGT/AGGGCTGAGCGC CAPN5 SNP14 A/G Intron 5 GATTTGCA/GGGGAGGATGGCATT CAPN5 SNP15 A/G Intron 6 GCGGGGAGGGA/GGACCGGCCGGCT CAPN5 SNP16 T/C Eoxn 7 TGTGCC/tGCTACTTCACGGACAT CAPN5 SNP17 G/A Eoxn 7 TCCACAAGACG/ATGGGAGGAGGCCC CAPN5 SNP18 T/C Intron7 TTCGGTATGGTT/CTCACTGCATGTC CAPN5 SNP19 T/C Intron 9 TGCGTGGGGGCT/CGGCCGGGGGCCAG CAPN5 SNP20 T/C Exon 10 AT/CGAGCCTCCCCGCACCTGCT CAPN5 SNP21 G/A Exon 11 GGACAAAGTCCG/ACTCGGCCGTGC CAPN5 SNP22 G/T Intron 11 TGTTTCCCCG/TCTCAGAATGGCT CAPN5 SNP23 T/C Exon 12 TGAAAGCCAAC/TGCGGAC CAPN5 SNP24 A/T Exon 12 GTCCCA/TCGTGTCCCCAGGGC CAPN5 SNP25 G/C Exon 12 TCCCAGACTTG/CCTGTCTCTGT

314

Appendix 17A. Genotype numbers for single candidate gene variants.

Variants Genotypes(no.) MSTN-F94L AA(54) AC(189) CC(116) CAPN1-SNP316 AA(22) AG(147) GG(167) CAPN1¬-SNP530 AA(44) AG(144) GG(151) CAPN4-two bases insertion/deletion* DD(3) ND(77) NN(269) (GCC)8/(GCC)8 CAPN4-3 base (6) (GCC)8/(GCC)10(119) (GCC)10/(GCC)10(222) (GCC)10/(GCC)12(18) repeat CAPN5-SNP16 TT(3) CT(95) TT(262) CAPN5-SNP21 GG(39) GA(176) AA(142) CAPN5-SNP17 AA(88) GA(153) GG(129) MYL7-SNP1 CC(69) CG(193) GG(98) MYO1G-SNP2 TT(8) TC(83) CC(268) MYO1G-SNP3 AA(1) AG(76) GG(278) MYO1G-SNP5 GG(17) GA(100) AA(248) LOX-SNP1 CC(52) CT(173) TT(138) LOXL1-SNP1 AA(83) AG(183) GG(75) IGF1-SNP1 CC(56) CT(188) TT(120) IGF1-SNP2 TT(15) CT(114) CC(253) IGF1R-SNP1 AA(2) AC(64) CC(297) FST-SNP5 GG(25) AG(162) AA(175) FST-SNP7 AA(3) AG(68) GG(293) FSTL1-SNP1 TT(4) TC(140) CC(219) FSTL1_SNP2 GG(5) AG(134) AA(207) SNIP1-SNP3 TT(20) CT(146) CC(199) SST-SNP2 CC(313) TC(39) *D = deletion; N = no deletion

315

Appendix 17B. Genotype numbers for significant pairs of candidate gene variants.

CAPN1-SNP316/CC CAPN1-SNP316/GC CAPN1-SNP316/GG SNIP1-SNP3/CC 4 79 101 SNIP1-SNP3/CT 11 61 60 SNIP1-SNP3/TT 7 7 5

SST-SNP2/CC 20 124 143 SST-SNP2/CT 2 16 21

IGF-SNP1/CC 6 27 20 IGF-SNP1/CT 16 77 83 IGF-SNP1/TT 0 43 62

FST-SNP5/AA 5 68 92 FST-SNP5/AG 17 68 62 FST-SNP5/GG 0 10 10

CAPN1-SNP530/AA CAPN1-SNP530/AG CAPN1-SNP530/GG SNIP1-SNP3/CC 28 76 79 SNIP1-SNP3/CT 12 59 61 SNIP1-SNP3/TT 1 8 11

FST-SNP7/AA 1 1 1 FST-SNP7/AG 13 35 19 FST-SNP7/GG 27 106 131

LOX-SNP1/CC 14 24 13 LOX-SNP1/CT 20 77 63 LOX-SNP1/TT 7 40 75

LOXL1-SNP1/AA 8 42 32 LOXL1-SNP1/AG 24 64 77 LOXL1-SNP1/GG 6 25 34

IGF1-SNP1/CC 12 24 19 IGF1-SNP1/CT 25 85 63 IGF1-SNP1/TT 4 33 69

MYO1G-SNP3/AA 19 36 20 MYO1G-SNP3/AG 21 103 126 MYO1G-SNP3/GG 1 4 5

316

Appendix 17B. Genotype numbers for significant pairs of candidate gene variants (continued).

MSTN-F94L/AA MSTN-F94L/AC MSTN-F94L/CC CAPN4 3 base repeat/A (GCC)8/(GCC)8 2 3 1 CAPN4 3 base repeat/B (GCC)8/(GCC)10 19 60 35 CAPN4 3 base repeat/C(GCC)10/(GCC)10 39 119 69 CAPN4 3 base repeat/D (GCC)10/(GCC)12 0 7 11

FSTL1-SNP1/CC 32 116 66 FSTL1-SNP1/CT 21 71 47 FSTL1-SNP1/TT 0 1 3

FSTL1-SNP2/AA 29 108 65 FSTL1-SNP2/AG 21 69 43 FSTL1-SNP2/GG 0 2 3

CAPN5-SNP16/CC 44 137 78 CAPN5-SNP16/CT 10 48 36 CAPN5-SNP16/TT 0 2 0

IGF1-SNP2/CC 35 112 84 IGF1-SNP2/CT 17 65 30 IGF1-SNP2/TT 2 11 2

IGF1-SNP1/CC 14 23 17 IGF1-SNP1/CT 23 107 55 IGF1-SNP1/TT 17 58 44

CAPN5-SNP21/AA 20 66 54 CAPN5-SNP21/GA 23 100 51 CAPN5-SNP21/GG 7 23 7

IGFR-SNP1/AA 1 1 0 IGFR-SNP1/AC 11 35 17 IGFR-SNP1/CC 41 152 99

FST-SNP5/AA 46 88 38 FST-SNP5/AG 7 87 65 FST-SNP5/GG 0 12 13

317

Appendix 17B. Genotype numbers for significant pairs of candidate gene variants (continued).

MSTN- MSTN- MSTN-

F94L/AA F94L/AC F94L/CC FST-SNP7/AA 1 1 1 FST-SNP7/AG 42 153 94 FST-SNP7/GG 0 1 0

MYO1G-SNP5/AA 38 135 71 MYO1G-SNP5/AG 14 46 39 MYO1G-SNP5/GG 2 8 6

SST-SNP2/CC 42 163 102 SST-SNP2/CT 11 20 8

LOXL1-SNP1 /AA 8 42 32 LOXL1-SNP1 /AG 24 64 77 LOXL1-SNP1 /GG 6 25 34

318

Appendix 18. Effect of candidate gene variants on shear force day 1 of ageing.

A general linear regression model (model 3 from section 2.5.8) was used to investigate the relationships between the 20 variants in the candidate genes and shear force at day 1 of ageing. This model was used to identify the effect of the single variants when there were only three fixed factors (cohort, breed and sire) fitted into the model.

The analysis showed that variants in the follistatin precursor and follistatin-related protein 1 genes (FST-SNP7, FSTL1-SNP1 and FSTL1-SNP2) were associated with shear force of the ST muscle at day 1 of ageing (Table 18.1). There were no significant relationships between any of variants and shear force at day 1 of ageing for the LD muscle.

Table 18.1. Tests of significance for candidate gene variants on shear force at day 1 ageing of the ST muscle. Name of Variant Significance Name of Variant Significance MYL7 SNP1 NS IGF1R SNP1 NS MYO1G SNP2 NS IGF1 SNP1 NS MYO1G SNP3 NS IGF1 SNP2 NS MYO1G SNP5 NS SNIP1 SNP3 NS CAPN4 two base NS FST SNP5 NS insertion/deletion * CAPN4 3 base repeat NS FST SNP7 0.01 * CAPN5 SNP16 NS FSTL1 SNP1 0.01 * CAPN5 SNP21 NS FSTL1 SNP2 0.006 LOX SNP1 NS SST SNP2 NS LOXL1 SNP1 NS NS: not significant, * (P<0.05)

The FST-SNP7 was associated with shear force at day 1 in the ST muscle (wbst1)

(P=0.01). Cattle with the AA and AG genotypes for FST-SNP7 had higher shear

319 force values for the ST muscle (Figure 18.1). The meat from heterozygous cattle at day 1 was 6.6% tougher than that of homozygous cattle (GG). There were no significant additive (P=0.887) or dominance effects (P=0.093) found. This may be due to the low number of samples with the AA genotype resulting in a large standard error for the AA genotype. That is, the dominance effect may have been significant if there were a greater number of cattle with the AA genotype.

5.8 a,b b 5.6 a

5.4

5.2 5

4.8 wbst1(KgF) 4.6 4.4 4.2 AA AG GG FST-SNP7 genotypes

Figure 18.1 Effect of FST-SNP7 genotypes on ST day 1 shear force. Bars represent standard errors. Letters indicate significance differences between genotypes.

The FSTL1-SNP1 was also associated with ST shear force at day 1 (wbst1)

(P=0.013). Cattle with the TT genotype for the FSTL1-SNP1 had higher shear force values for the ST muscle (Figure 18.2). The meat from homozygous TT cattle at day

1 was 16% tougher than that of homozygous CC cattle. The additive effect was significant (P=0.02). The estimated allelic substitution effect for the C allele was -

0.425 ± 0.182 kgF.

320

7 b a,b 6 a

5 4 3

wbst1(KgF) 2 1 0 CC TC TT FSTL1-SNP1 genotypes

Figure 18.2 Effect of FSTL1-SNP1 genotypes on ST day 1 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: Y-axis scale difference due to size of effect.

Lastly, the FSTL1-SNP2 was associated with shear force at day 1 (wbst1) (P=0.006).

Cattle with the GG and AG genotypes for the FSTL1-SNP2 had higher shear force values for the ST muscle than cattle with the AA genotype (Figure 18.3). The meat of homozygous GG cattle at day 1 was 17% tougher than that of homozygous AA cattle.

The additive effect was significant (P=0.011). The estimated allelic substitution effect for the G allele was 0.418 ± 0.162 kgF.

7 a 6 b b

5 4 3 wbst1(KgF) 2 1 0 AA AG GG FSTL1-SNP2 genotypes

Figure 18.3 Effect of FSTL1-SNP2 genotypes on ST day 1 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: Y-axis scale different due to size of effect.

321

Appendix 19. Effect of candidate gene variants on shear force day 26 of ageing.

The analysis for shear force at day 26 of ageing was based on the same general linear regression model used for day 1 (model 3 from section 2.5.8). The results demonstrated that for the LD muscle, SNIP1-SNP3 had a significant effect (P=0.026) on the Warner-Bratzler shear force at day 26 of ageing for the LD muscle (wbld26).

The meat of homozygous CC cattle and heterozygous CT cattle at day 26 was 14% more tender than that of homozygous TT cattle. Furthermore, there was a significant additive effect (P=0.008) and dominance effect (P=0.037). For the additive effect, the estimated allelic substitution effect for the T allele was 0.278 ± 0.104 kgF. For the dominance effect, the estimated allelic substitution effect for the T allele was -

0.248 ± 0.118 kgF (Figure 19.1). The LOXL1-SNP1 was the only other SNP close to the significant threshold for wbld26 (P=0.084) in the LD muscle.

5 a 4.5 b b 4

3.5 3 2.5 2 wbld26(KgF) 1.5 1 0.5 0 CC CT TT SNIP1-SNP3 genotypes

Figure 19.1. Effect of SNIP1-SNP3 genotypes on LD day 26 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes.

322

For the ST muscle, there were no variants associated with day 26 shear force

(wbst26) except the CAPN4-3 base repeat genotypes (P=0.057) (Figure 19.2).

Comparing the genotypes of GCC10,10 repeats with GCC8,10 repeats, animals homozygous for the GCC10,10 repeats had 3.5% higher shear force values than heterozygous GCC8,10 animals. Thus, the number of the repeated glycine residues may be associated with shear force values for the ST muscle at day 26 of ageing.

5 a a,b 4.8 b

4.6 b

4.4

4.2 wbst26(KgF) 4

3.8

3.6 A B C D CAPN4-3 base repeat

Figure 19.2. Effect of CAPN4-3 base repeat genotypes on ST day 26 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. A: (GCC)8/(GCC)8 (number of animals = 6) B: (GCC)8/(GCC)10 (number of animals = 118) C: (GCC)10/(GCC)10 (number of animals = 220) D: (GCC)10/(GCC)12 (number of animals = 18)

323

Appendix 20. Effect of candidate gene variants on shear force day 1 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes.

A general linear regression model (model 4 from section 2.5.8) was used for investigating the relationships between 20 variants of candidate genes and 3 major known variants (MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530) related to tenderness. This model was used to identify the effect of the candidate genes when the genotypes of MSTN and CAPN1 were fitted into the model.

There were no variants affecting shear force at day 1 of ageing on the LD muscle after the MSTN and CAPN1 genotypes in the model. Thus, for the LD muscle, the results were very similar to the results from the model without MSTN and CAPN1.

Hence, the genetic effects at early stages of LD tenderization do not seem to be large for the candidate genes herein. However, there may be other genes related to tenderness at the early stage of tenderization.

However, three variants (FSTL1-SNP1, FSTL1-SNP2 and MYO1G-SNP5) showed significant effects on shear force value measured at day 1 of ageing for the ST muscle when MSTN and CAPN1 genotypes were included in the model. Another 4 variants (IGF1R-SNP1, IGF1-SNP1, FST-SNP7 and MYO1G-SNP2 showed effects which were close to significance (P<0.1) on shear force at day 1 (Table 20.1).

In general, follistatin precursor (FST) and follistatin-related protein 1 (FSTL1) were associated with shear force at day 1 ageing for the ST muscle after the effects of

MSTN and CAPN1 were taken into consideration. FSTL1-SNP1 and FSTL1-SNP2 explained 1.6% and 1.8% of the total variation in shear force at day 1 of ageing on the ST muscle, respectively. FST-SNP7 accounted for 1% of the total variation in shear force.

324

Table 20.1. Tests of significance of candidate gene variants on ST shear force at day

1 ageing. Name of Variant Significance

FSTL1-SNP1 (0.011) *

FSTL1-SNP2 (0.010) *

MYO1G SNP5 (0.045) *

MYO1G SNP3 (0.067) †

(0.071) † IGF1R SNP1 (0.058) † IGF1 SNP1 FST SNP7 (0.055) †

†(P<0.1) * (P<0.05); ** (P<0.005); *** (P<0.001) NS: not significant

The FSTL1-SNP1 was associated with the ST day 1 shear force (wbst1) (P=0.011).

The cattle with the CC and CT genotypes for FSTL1-SNP1 had higher shear force values for the ST muscle than the cattle with the TT genotype (Figure 20.1). The cattle with the CC genotype produced 12.5% tougher meat than those with the TT genotype. The cattle with the CT genotype produced 4.5% tougher meat than those with the TT genotype. The T allele was recessive, but an additive effect was observed (P=0.011). The estimated allelic substitution effect for T allele was -0.374

± 0.174 kgF on wbst1.

325

7 b 6 b a

WBST WBST (KgF) 1 2

0 CC TC TT FSTL1-SNP1 genotypes

Figure 20.1 Effect of FSTL1-SNP1 genotypes on ST day 1 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes.

The FSTL1-SNP2 was also associated with the ST day 1 shear force (wbst1) (P=0.010). Cattle with the AA genotype for FSTL1-SNP2 had lower shear force values for the ST muscle (Figure 20.2). The cattle with GG genotype produced 13.7% tougher meat than those with the AA genotype. The additive effect was significant (P=0.023). The estimated allelic substitution effect for the G allele was 0.358± 0.157 kgF.

7 b 6 a b

WBST WBST (KgF) 1 2

0 AA AG GG FSTL1-SNP2 genotypes

Figure 20.2 Effect of FSTL1-SNP2 genotypes on ST day 1 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes.

326

Lastly, the MYO1G-SNP5 significantly affected ST day 1 shear force (wbst1)

(P=0.045). Cattle with the AA genotype for MYO1G-SNP5 had lower shear force values for the ST muscle (Figure 20.3). The cattle with the GG genotype produced

5.7% tougher meat than those with the AA genotype. However, the low number (8) of animals with the GG genotype in the sample needs to be considered. The additive effect was significant (P=0.014). However, when the dominance effect was considered in the model as well, the additive effect was less significant (P=0.13).

5.8 ab 5.7 5.6 b 5.5 5.4 a 5.3

WBST WBST (KgF) 1 5.2 5.1 5 4.9 AA GA GG MYO1G-SNP5 genotypes

Figure 20.3 Effect of MYO1G-SNP5 genotypes on ST day 1 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: Y-axis scale difference due to size of effect.

327

Appendix 21. Effect of candidate gene variants on shear force day 26 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes.

When the MSTN and CAPN1 genotypes were taken into account for the shear force at day 26 of ageing (model 4 from section 2.5.8), the results demonstrate that for the

LD muscle (wbld26), only the SNIP1-SNP3 had a significant effect (P=0.005). The cattle with the TT genotype had 19% tougher meat than those with the CC genotype.

Furthermore, there was an additive effect (P=0.001) and a dominance effect

(P=0.016). For the additive effect, the estimated allelic substitution effect for T allele was 0.278 ± 0.104 kgF on wbld26. For the dominance effect, the estimated allelic substitution effect for T allele was -0.248 ± 0.118 kgF (Figure 21.1). Interestingly, the SNIP1-SNP3 accounted for 2% of the total variation of LD shear force at day 26 of ageing. In other words, the effect of SNIP1-SNP3 equalled the size of effect of the

CAPN1-SNP316 on wbld26.

The effect of LOXL1-SNP1 on LD day 26 shear force (wbld26) was close to the significance threshold (P=0.063). The cattle with the AA genotype had 8% tougher meat than those with the GG or AG genotypes (Figure 21.2). The LOXL1-SNP1 explained 1.2% of the total variation in LD shear force at day 26 of ageing. The G allele of LOXL1- SNP1 was dominant. An additive effect for LOXL1-SNP was observed. The estimated allelic substitution effect for G allele was -0.1534 ± 0.0752 kgF on wbld26.

328

5 a

b b 4

2 WBSLD26 WBSLD26 (Kg F)

0 CC CT TT SNIP1-SNP3 genotypes

Figure 21.1 Effect of SNIP1-SNP3 genotypes on LD day 26 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes.

4.3 a 4.2

4.1

4 b b 3.9 3.8

3.7 WBLD WBLD (Kg 26 F) 3.6 3.5 3.4 AA AG GG LOXL1-SNP1 genotypes

Figure 21.2 Effect of LOXL1-SNP1 genotypes on LD day 26 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: Y-axis scale difference due to size of effect.

For the ST muscle, there was one variant, FST-SNP5, significantly associated with

ST day 26 shear force (wbst26) (P=0.041). The cattle with the AA genotype for the

FST-SNP5 had approximately 6.4% higher shear force values for the ST muscle than the other genotypes (Figure 21.3). The FST-SNP5 explained 1.0% of the variance.

For the additive effect, the estimated allelic substitution effect for G allele was

329

-0.1443 ± 0.0862 kgF on wbst26. FST-SNP7 was close to significance for LD day

26 shear force (P=0.087) (Figure 21.4). There was a dominance effect discovered for

FST-SNP7 but the variation accounted by FST-SNP7 was less than 1% (0.378 ±

0.176 kgF). However, the low number of animals with the AA genotype (3) of FST-

SNP7 may have biased the result.

4.9 a 4.8

4.7 b b

4.6 4.5 4.4

4.3 WBST WBST (Kg 26 F) 4.2 4.1 4 AA AG GG FST-SNP5 genotypes

Figure 21.3 Effect of FST-SNP5 genotypes on ST day 26 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes. Note: Y- axis scale difference due to size of effect.

6 b

5 a a,b

2 WBST WBST (Kg 26 F)

0 AA AG GG FST-SNP7 genotypes

Figure 21.4 Effect of FST-SNP7 genotypes on ST day 26 shear force. Bars represent standard errors. Letters indicate significant differences between genotypes.

330

Appendix 22. Effect of candidate gene variant interactions on shear force day 1 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes.

In the analysis of interactions affecting shear force at day 1, two interactions significantly affected LD day 1 shear force and five interactions significantly affected ST day 1 shear force (Table 22.1). All these significant interactions except for the interactions between SNIP1-SNP3 and CAPN1-SNP316 /CAPN1-SNP530 did not show any effect on other tenderness traits (e.g. day 26 shear force, adjusted shear force). This implies that these interactions may only have an influence the early stages of tenderization and not the end of tenderization.

Table 22.1. Significant interactions between genes affecting wbld1 and wbst1.

Traits DNA variants P-values wbld1 SST-SNP2 and CAPN1-SNP316 0.01 wbld1 SNIP1-SNP3 and CAPN1-SNP530 0.034 wbst1 CAPN4 3 base repeat and MSTN F94L 0.048 wbst1 SNIP1-SNP3 and CAPN1-SNP316 <0.001 wbst1 FSTL1-SNP1 and CAPN1-SNP316 0.042 wbst1 FSTL1-SNP2 and CAPN1-SNP316 0.042 wbst1 SNIP1-SNP3 and CAPN1-SNP530 0.002

Only the interaction between SNIP1-SNP3 and CAPN1-SNP530 had an effect on both muscles. The number of the SNIP1-SNP3 significant interactions with other genes was greater than other candidate genes, suggesting an important role of

SNIP1-SNP3 in the early stages of tenderization.

The interaction between SST-SNP2 and CAPN1-SNP316 contributed 2.11% to the total sum of squares of LD day 1 shear force (Table 22.2). SST-SNP2 itself did not demonstrate any effect on LD day 1 shear force (wbld1). The CC genotype of SST-

331

SNP2 resulted in higher shear force values in the animals with the GG genotype of

CAPN1-SNP316 than the animals with the TC genotype of SST-SNP2 (Figure 22.1).

This phenomenon is different from those animals with the CC/GC genotype of

CAPN1-SNP316 (Figure 22. 1). Usually, animals with the GG genotype of CAPN1-

SNP316 had higher LD day 1 shear force (wbld1) values than the other genotypes of

CAPN1-SNP316. However, the wbld1 values for the animals with the GG genotypes of CAPN1-SNP316 were not greater than other 2 genotypes of CAPN1-SNP316 when the animals were also carrying the TC genotype of SST-SNP2.

4 SST-SNP2/CC 3

wbld1 (KgF) SST-SNP2/TC 2

0 CAPN1-SNP316/CC CAPN1-SNP316/GC CAPN1-SNP316/GG

Figure 22.1. Interaction of SST-SNP2 and CAPN1-SNP316 genotypes on LD day 1 shear force. Bars represent standard errors.

Table 22.2. Proportion of the total sum of squares accounted for by the interaction between SST-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 1 shear force.

Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 0.39 0.419 CAPN1-SNP316 4.06 <.001 CAPN1-SNP530 1.45 0.041 SST-SNP2 0.17 0.379 MSTN F94L and SST-SNP2 0.34 0.473 CAPN1-SNP316 and SST-SNP2 2.11 0.01 CAPN1-SNP530 and SST-SNP2 0.01 0.978 332

The SNIP1-SNP3 itself did not affect LD day 1 shear force (wbld1). However, the

SNIP1-SNP3 interaction with CAPN1-SNP530 significantly affected wbld1 and accounted for 2.27% of the total sum of squares. This is a slightly larger effect than the contribution of CAPN1-SNP530 alone (1.41%) (Table 22.3). The GG genotype of CAPN1-SNP530 was shown to be associated with lower LD day 1 shear force values than the other genotypes of CAPN1-SNP530 (Chapter 5). In the animals with the TT genotype of SNIP1-SNP3, the GG genotype of CAPN1-SNP530 did not show any difference between the AG and GG genotypes of CAPN1-SNP530. Due to the insufficient number of animals with the AA genotypes of CAPN1-SNP530 and the

TT genotype of SNIP1-SNP3, however, the size of the effect of AA genotype of

CAPN1-SNP530 in this class should be viewed cautiously given to the large standard error (Figure 22.2).

9 8 7 6

5 CAPN1-SNP530/AA 4 CAPN1-SNP530/AG 3 CAPN1-SNP530/GG wbld1(KgF) 2 1 0 -1 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT -2

Figure 22.2. Interaction of SNIP1-SNP3 and CAPN1-SNP530 genotypes on LD day 1 shear force. Bars represent standard errors.

333

Table 22.3 Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 1 shear force. Proportion of the total sum of DNA variants P-value squares (%) MSTN F94L 0.29 0.512 CAPN1-SNP316 3.90 <.001 CAPN1-SNP530 1.41 0.039 SNIP1_SNP3 0.91 0.123 MSTN F94L and SNIP1-SNP3 0.63 0.568 CAPN1-SNP316 and SNIP1-SNP3 1.51 0.137 CAPN1-SNP530 and SNIP1-SNP3 2.27 0.034

For ST day 1 shear force (wbst1), the CAPN4-3 base repeat variant interacted with

MSTN F94L to significantly affect wbst1, but only accounted for 1.76% of the total sum of squares (Table 22.4). Compared with the effects of MSTN F94L or CAPN1-

SNP316, the effect of this interaction is relatively small. As the number of each class of the repeats of CAPN4 with the various MSTN F94L genotypes is not balanced, the prediction of the size of effect on wbst1 for each combination could not be calculated.

Therefore, the interaction of the CAPN4-3 base repeat with MSTN F94L on ST day 1 shear force should only be regarded as suggestive.

Table 22.4. Proportion of the total sum of squares accounted for by the interaction between CAPN4 -3 base repeat and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 1 shear force. Proportion of the total sum of DNA variants P-value squares (%) MSTN F94L 4.86 <.001 CAPN1-SNP316 3.02 <.001 CAPN1-SNP530 0.07 0.816 CAPN4 -3 base repeat 0.32 0.621 MSTN F94L.CAPN4 -3 base repeat 1.76 0.048 CAPN1-SNP316 and CAPN4 -3 base repeat 0.50 0.735 CAPN1-SNP530 and CAPN4 -3 base repeat 0.89 0.429

The estimate for this interaction on wbst1 was almost 7 kgF (Table 22.5). The effect of the interaction between SNIP1-SNP3 and CAPN1-SNP316 explained 4.59% of the

334 total sum of squares, which is higher than that explained by CAPN1-SNP316 itself

(3.02%) and is similar to the size of effect of MSTN F94L (4.86%) (Table 22.5).

Combining the effects of CAPN1-SNP316 and the interaction of SNIP1-SNP3 and

CAPN1-SNP316, the contribution of CAPN1-SNP316 to ST day 1 shear force (wbst1) is actually greater than that of MSTN F94L alone. There was an unexpected outcome in that the interaction between SNIP1-SNP3/TT and CAPN1-SNP316/GC decreased the wbst1 values. This is in contrast to the interaction between SNIP1-SNP3/TT and

CAPN1-SNP316/GG which greatly increased the wbst1 values (Figure 22.3).

Table 22.5. Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 1 shear force. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 4.86 <.001 CAPN1-SNP316 3.02 <.001 CAPN1-SNP530 0.07 0.798 SNIP1-SNP3 0.10 0.739 SNIP1-SNP3 and MSTN F94L 0.55 0.496 CAPN1-SNP316 and SNIP1-SNP3 4.59 <.001 CAPN1-SNP530 and SNIP1-SNP3 2.80 0.002

5 CAPN1-SNP316/CC 4 CAPN1-SNP316/GC

wbst1(F/Kg) 3 CAPN1-SNP316/GG 2

0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 22.3. Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on ST day 1 shear force. Bars represent standard errors.

335

The GG genotype of the CAPN1-SNP530 had a negative effect and the TT genotype of the SNIP1-SNP3 had positive effect on ST day 1 shear force (wbst1) values.

Therefore, the GG genotype of the CAPN1-SNP530 resulted in similar wbst1 values for the cattle with the TT genotype of the SNIP1-SNP3 and cattle with the other genotypes of the SNIP1-SNP3. However, the interaction between SNIP1-SNP3/TT and CAPN1-SNP530/AG did not decrease the wbst1 values (Figure 22.4).

In contrast, the interaction between SNIP1-SNP3/TT and CAPN1-SNP530/AG increased the ST day 1 shear force (wbst1) values. However, due to the insufficient number of animals with the AA genotype of CAPN1-SNP530 and the TT genotype of SNIP1-SNP3, there was a large error for this class and the effect of the interaction could only be approximated. Given this caveat, 2.8% of the total sum of the squares was accounted for by the interaction between SNIP1-SNP3 and CAPN1-SNP530.

This is half of the effect of the interaction between SNIP1-SNP3 and CAPN1-

SNP316 (Table 22.5).

4 CAPN1-SNP530/AA 3 CAPN1-SNP530/AG

wbst1(F/Kg) CAPN1-SNP530/GG 2

0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 22.4. Interaction of SNIP1-SNP3 and CAPN1-SNP530 genotypes on ST day 1 shear force. Bars represent standard errors.

336

FSTL1-SNP1 and FSTL1-SNP2 had a very similar effect on ST day 1 shear force

(wbst1). Both of these two variants (FSTL1-SNP1 and FSTL1-SNP2) significantly interacted with CAPN1-SNP316 on wbst1 (Table 22.6 and Table 22.7). However, when the FSTL1-SNP1 and FSTL1-SNP2 were fitted into the model simultaneously, the effect disappeared for which ever variant was fitted last. This suggests that the effect of the FSTL1-SNP1 and FSTL1-SNP2, therefore, may be the one and the same.

Table 22.6. Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 1 shear force. Proportion of the total sum DNA variants P-value of squares (%) MSTN F94L 4.86 <.001 CAPN1-SNP316 3.03 <.001 CAPN1-SNP530 0.07 0.817 FSTL1-SNP1 1.62 0.01 MSTN F94L and FSTL1-SNP1 0.44 0.474 CAPN1-SNP316 FSTL1- and 1.44 0.042 SNP1 CAPN1-SNP530 FSTL1- and 1.09 0.102 SNP1

Table 22.7 Proportion of the total sum of squares accounted for by the interaction between FSTL1-SNP2 and 3 variants (MSTN F94L, CAPN1- SNP316 and SNP530) on ST day 1 shear force. Proportion of the total sum of DNA variants P-value squares (%) MSTN F94L 5.00 <.001 CAPN1-SNP316 2.80 <.001 CAPN1-SNP530 0.07 0.831 FSTL1-SNP2 1.89 0.006 MSTN F94L and FSTL1-SNP2 0.56 0.377 CAPN1-SNP316 and FSTL1-SNP2 1.49 0.042 CAPN1-SNP530 and FSTL1-SNP2 1.14 0.1

337

Appendix 23. Effect of candidate gene variant interactions on shear force day 26 of ageing with MSTN F94L, CAPN1 SNP316 and CAPN1 SNP530 genotypes.

There were also several interactions affecting shear force at day 26 for both the LD and ST muscles (wbld26 and wbst26). Two interactions significantly affected wbld26 and three interactions significantly affected wbst26 (Table 23.1). The interaction between CAPN1-SNP316 and SNIP1-SNP3 is the only interaction affecting shear force on both the LD and ST muscles. The effects of the other significant interactions were muscle specific.

Table 23.1 Significant interactions between genes affecting LD and ST day 26 shear force (wbld26 and wbst26). Traits DNA variants P-values wbld26 LOXL1-SNP1 and CAPN1-SNP316 0.041 wbld26 SNIP1-SNP3 and CAPN1-SNP316 0.007 wbst26 FST-SNP5 and MSTN F94L 0.037 wbst26 IGF1-SNP2 and CAPN1-SNP316 0.01 wbst26 SNIP1-SNP3 and CAPN1-SNP316 <0.001

Furthermore, with the exception of the interaction between FST-SNP5 and MSTN

F94L that affected tenderness (wbst26) and muscle weight (ST weight), there were no interactions affecting shear force and muscle weight for the same muscle.

The interaction between the TT genotype of SNIP1-SNP3 and the GG genotypes of

CAPN1-SNP316 showed similar results to adjusted shear force. The interaction between CAPN1-SNP316 and SNIP1-SNP3 was significant and explained 2.7% of the variation on LD day 26 shear force (wbld26). The effect of the interaction was slightly larger than the single effects of either the CAPN1-SNP316 or SNIP1-SNP3

(Table 23.2). For the ST muscle, MSTN F94L, which is known to affect tenderness of the ST muscle, accounted for 4.6% of the variation in ST day 26 shear force

338

(wbst26). The CAPN1-SNP316 only contributed 1.5% of the variation and SNIP1-

SNP3 did not show any individual effect. However, the interaction between CAPN1-

SNP316 and SNIP1-SNP3 explained 3.4% of the variation. The effect of the interaction was greater than the individual effect of SNIP1-SNP3 and similar to the size of the effect of MSTN F94L (Table 23.3).

Table 23.2 Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 26 shear force. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 0.16 0.656 CAPN1-SNP316 2.16 0.004 CAPN1-SNP530 0.30 0.447 SNIP1-SNP3 2.15 0.004 SNIP1-SNP3 and MSTN F94L 1.14 0.197 CAPN1-SNP316 and SNIP1- 2.68 0.007 SNP3 CAPN1-SNP530 and SNIP1- 0.89 0.319 SNP3

Table 23.3 Proportion of the total sum of squares accounted for by the interaction between SNIP1-SNP3 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 26 shear force.

Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 4.63 <.001 CAPN1-SNP316 1.54 0.005 CAPN1-SNP530 0.24 0.429 SNIP1-SNP3 0.06 0.802 SNIP1-SNP3 and MSTN F94L 1.13 0.097 CAPN1-SNP316 and SNIP1- 3.49 <.001 SNP3 CAPN1-SNP530 and SNIP1- 0.72 0.287 SNP3

339

The interaction between TT genotype of the SNIP1-SNP3 and GG genotype of the

CAPN1-SNP316 showed higher shear force values at day 26 for both the LD and ST muscles than the other genotype combinations of SNIP1-SNP3 and of CAPN1-

SNP316. These results were similar to the effects of the interactions on adjusted shear force (Figure 23.1 and 23.2).

5 CAPN1-SNP316/CC 4 CAPN1-SNP316/GC 3 wbld26(KgF) CAPN1-SNP316/GG 2

0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 23.1. Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on LD day 26 shear force. Bars represent standard errors.

4 CAPN1-SNP316/CC 3 CAPN1-SNP316/GC

wbst26(KgF) CAPN1-SNP316/GG 2

0 SNIP1-SNP3/CC SNIP1-SNP3/CT SNIP1-SNP3/TT

Figure 23.2. Interaction of SNIP1-SNP3 and CAPN1-SNP316 genotypes on ST day 26 shear force. Bars represent standard errors.

340

Although the LOXL1-SNP1 did not show any significant individual effects on LD day 26 shear force (wbld26), the interaction between CAPN1-SNP316 and LOXL1-

SNP1 was comparable to the effect of CAPN1-SNP316 alone on wbld26. Both explained around 2.16% of variation of the LD day 26 shear force (Table 23.4). The interaction between AA genotype of LOXL1-SNP1 and the GG genotype of CAPN1-

SNP316 resulted in significantly higher wbld26 values than all other genotype combinations except for the interaction between the AA genotype of LOXL1-SNP1 and the CC genotype of CAPN1-SNP316 (Figure 23.3).

Table 23.4 Proportion of the total sum of squares accounted for by the interaction between LOXL1-SNP1 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on LD day 26 shear force.

Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 0.20 0.633 CAPN1-SNP316 2.17 0.007 CAPN1-SNP530 0.28 0.517 LOXL1-SNP1 1.21 0.061 MSTN F94L and LOXL1-SNP1 0.58 0.603 CAPN1_SNP316 and LOXL1-SNP1 2.16 0.041 CAPN1_SNP530 and LOXL1-SNP1 0.63 0.564

5 4.5 4 3.5 3 2.5 2 CAPN1-SNP316/CC

wbld26(KgF) 1.5 CAPN1-SNP316/GC 1 0.5 CAPN1-SNP316/GG 0

Figure 23.3. Interaction of LOXL1-SNP1 and CAPN1-SNP316 genotypes on LD day 26 shear force. Note: Y-axis scale difference due to size of effect.

341

The interactions between FST-SNP5 and MSTN F94L and between IGF1-SNP2 and

CAPN1-SNP316 only affected shear force at day 26 of the ST muscle. The effects were small. The FST-SNP5 interaction with MSTN F94L explained 1.3% of the proportion of the total sum of squares (Table 23.5). This is the only evidence that there may be interaction between the follistatin and myostatin genes affecting shear force.

Table 23.5. Proportion of the total sum of squares accounted for by the interaction between FST-SNP5 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 26 shear force. Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 4.72 <.001 CAPN1-SNP316 1.55 0.007 CAPN1-SNP530 0.25 0.436 FST-SNP5 1.00 0.04 MSTN F94L and FST-SNP5 1.31 0.037 CAPN1-SNP316 and FST-SNP5 0.58 0.288 CAPN1-SNP530 and FST-SNP5 0.03 0.98

The IGF1-SNP2 interaction with CAPN1-SNP316 explained 1.73% of the proportion of the total sum of squares. IGF1-SNP2 was not individually associated with ST day 26 shear force. However, its interaction with CAPN1-SNP316 was larger than the proportion of the total sum of squares for CAPN1-SNP316 alone for this trait (Table 23.6).

Table 23.6. Proportion of the total sum of squares accounted for by the interaction between IGF1-SNP2 and 3 variants (MSTN F94L, CAPN1-SNP316 and SNP530) on ST day 26 shear force.

Proportion of the total DNA variants P-value sum of squares (%) MSTN F94L 4.62 <.001 CAPN1-SNP316 1.54 0.007 CAPN1-SNP530 0.25 0.444 IGF1-SNP2 0.09 0.74 MSTN F94L and IGF1-SNP2 0.19 0.747 CAPN1-SNP316 and IGF1-SNP2 1.73 0.01 CAPN1-SNP530 and IGF1-SNP2 0.79 0.157

342

Appendix 24. Variant interactions for LD adjusted shear force.

CAPN4-two MSTN- CAPN1- CAPN1- bases F94L SNP316 SNP530 insertion/deletion Main effect <.001 0.029 0.343 MSTN-F94L 0.399 0.668 0.906 CAPN1-SNP316 0.399 0.209 0.253 CAPN1-SNP530 0.668 0.209 0.503 CAPN4-two bases insertion/deletion 0.906 0.253 0.503 CAPN5-SNP16 0.85 0.748 0.462 0.607 CAPN5-SNP21 0.964 0.981 0.738 0.745 IGF1R1 0.715 0.947 0.393 0.612 IGF1-SNP1 0.969 0.816 0.665 0.887 IGF1-SNP2 0.715 0.795 0.304 0.788 LOXL1-SNP1 0.6 0.123 0.28 0.418 LOX-SNP1 0.704 0.253 0.448 0.651 MYL7-SNP1 0.983 0.71 0.268 0.905 MYO1G-SNP2 0.984 0.696 0.323 0.513 MYO1G-SNP3 0.358 0.411 0.301 0.512 CAPN4-ten bases insertion/deletion 0.902 0.238 0.666 0.869 SNIP1-SNP3 0.113 <.001 0.004 0.04 FSTL1_SNP2 0.75 0.842 0.523 0.644 FST-SNP5 0.964 0.511 0.901 0.616 FST-SNP7 0.252 0.309 0.097 0.472 FSTL1-SNP1 0.771 0.852 0.515 0.664 SST-SNP2 0.605 0.304 0.224 0.178 MYO1G-SNP5 0.542 0.542 0.139 0.258 CAPN5-SNP17 0.962 0.967 0.123 0.709

343

Appendix 24. Variant interactions for LD adjusted shear force (continued).

CAPN5- CAPN5- IGF1- IGF1- IGF1R1 SNP16 SNP21 SNP1 SNP2 Main effect 0.468 0.928 0.813 0.987 0.483 MSTN-F94L 0.85 0.964 0.715 0.969 0.715 CAPN1-SNP316 0.748 0.981 0.947 0.816 0.795 CAPN1-SNP530 0.462 0.738 0.393 0.665 0.304 CAPN4-two bases 0.745 insertion/deletion 0.607 0.612 0.887 0.788 CAPN5-SNP16 0.938 0.881 0.97 0.907 CAPN5-SNP21 0.938 0.968 0.971 0.942 IGF1R1 0.881 0.968 0.641 0.826 IGF1-SNP1 0.97 0.971 0.641 0.613 IGF1-SNP2 0.907 0.942 0.826 0.613 LOXL1-SNP1 0.233 0.116 0.441 0.839 0.615 LOX-SNP1 0.44 0.895 0.636 0.58 0.686 MYL7-SNP1 0.934 0.985 0.416 0.518 0.838 MYO1G-SNP2 0.912 0.898 0.828 0.86 0.94 MYO1G-SNP3 0.409 0.616 0.674 0.833 0.624 CAPN4-ten bases insertion/deletion 0.841 0.889 0.843 0.742 0.767 SNIP1-SNP3 0.044 0.115 0.086 0.169 0.006 FSTL1_SNP2 0.917 0.98 0.955 0.978 0.792 FST-SNP5 0.903 0.925 0.918 0.578 0.516 FST-SNP7 0.03 0.698 0.708 0.714 0.283 FSTL1-SNP1 0.948 0.985 0.973 0.949 0.708 SST-SNP2 0.174 0.408 0.804 0.881 0.401 MYO1G-SNP5 0.524 0.85 0.716 0.621 0.732 CAPN5-SNP17 0.946 0.981 0.921 0.76 0.941

344

Appendix 24. Variant interactions for LD adjusted shear force (continued).

LOXL1- LOX- MYL7- MYO1G- MYO1G- SNP1 SNP1 SNP1 SNP2 SNP3 Main effect 0.307 0.703 0.885 0.753 0.301 MSTN-F94L 0.6 0.704 0.983 0.984 0.358 CAPN1-SNP316 0.123 0.253 0.71 0.696 0.411 CAPN1-SNP530 0.28 0.448 0.268 0.323 0.301 CAPN4-two bases insertion/deletion 0.418 0.651 0.905 0.513 0.512 CAPN5-SNP16 0.233 0.44 0.934 0.912 0.409 CAPN5-SNP21 0.116 0.895 0.985 0.898 0.616 IGF1R1 0.441 0.636 0.416 0.828 0.674 IGF1-SNP1 0.839 0.58 0.518 0.86 0.833 IGF1-SNP2 0.615 0.686 0.838 0.94 0.624 LOXL1-SNP1 0.093 0.348 0.755 0.467 LOX-SNP1 0.093 0.629 0.896 0.569 MYL7-SNP1 0.348 0.629 0.509 0.706 MYO1G-SNP2 0.755 0.896 0.509 0.791 MYO1G-SNP3 0.467 0.569 0.706 0.791 CAPN4-ten bases insertion/deletion 0.719 0.74 0.766 0.958 0.596 SNIP1-SNP3 0.03 0.152 0.077 0.048 0.04 FSTL1_SNP2 0.926 0.794 0.489 0.307 0.732 FST-SNP5 0.378 0.202 0.669 0.671 0.671 FST-SNP7 0.357 0.643 0.548 0.382 0.497 FSTL1-SNP1 0.839 0.723 0.509 0.297 0.743 SST-SNP2 0.153 0.533 0.392 0.988 0.629 MYO1G-SNP5 0.673 0.552 0.688 0.828 0.467 CAPN5-SNP17 0.41 0.702 0.865 0.954 0.692

345

Appendix 24. Variant interactions for LD adjusted shear force (continued).

CAPN4-ten SNIP1- FST- bases FSTL1_SNP2 SNP3 SNP5 insertion/deletion Main effect 0.769 0.101 0.81 0.893 MSTN-F94L 0.902 0.113 0.75 0.961 CAPN1-SNP316 0.238 <.001 0.842 0.964 CAPN1-SNP530 0.666 0.004 0.523 0.511 CAPN4-two bases insertion/deletion 0.869 0.04 0.644 0.901 CAPN5-SNP16 0.841 0.044 0.917 0.616 CAPN5-SNP21 0.889 0.115 0.98 0.903 IGF1R1 0.843 0.086 0.955 0.925 IGF1-SNP1 0.742 0.169 0.978 0.918 IGF1-SNP2 0.767 0.006 0.792 0.578 LOXL1-SNP1 0.719 0.03 0.926 0.516 LOX-SNP1 0.74 0.152 0.794 0.378 MYL7-SNP1 0.766 0.077 0.489 0.202 MYO1G-SNP2 0.958 0.048 0.307 0.669 MYO1G-SNP3 0.596 0.04 0.732 0.671 CAPN4-ten bases insertion/deletion 0.053 0.941 0.746 SNIP1-SNP3 0.053 0.141 0.058 FSTL1_SNP2 0.941 0.141 0.901 FST-SNP5 0.746 0.058 0.901 FST-SNP7 0.309 0.089 0.592 0.546 FSTL1-SNP1 0.953 0.137 0.969 0.931 SST-SNP2 0.187 0.024 0.837 0.142 MYO1G-SNP5 0.846 0.12 0.655 0.886 CAPN5-SNP17 0.749 0.17 0.921 0.996

346

Appendix 24. Variant interactions for LD adjusted shear force (continued).

FST- FSTL1- SST- MYO1G- SNP7 SNP1 SNP2 SNP5 Main effect 0.355 0.823 0.588 0.298 MSTN-F94L 0.252 0.771 0.605 0.542 CAPN1-SNP316 0.309 0.852 0.304 0.542 CAPN1-SNP530 0.097 0.515 0.224 0.139 CAPN4-two bases insertion/deletion 0.472 0.664 0.178 0.258 CAPN5-SNP16 0.03 0.948 0.174 0.524 CAPN5-SNP21 0.698 0.985 0.408 0.85 IGF1R1 0.708 0.973 0.804 0.716 IGF1-SNP1 0.714 0.949 0.881 0.621 IGF1-SNP2 0.283 0.708 0.401 0.732 LOXL1-SNP1 0.357 0.839 0.153 0.673 LOX-SNP1 0.643 0.723 0.533 0.552 MYL7-SNP1 0.548 0.509 0.392 0.688 MYO1G-SNP2 0.382 0.297 0.988 0.828 MYO1G-SNP3 0.497 0.743 0.629 0.467 CAPN4-ten bases insertion/deletion 0.309 0.953 0.187 0.846 SNIP1-SNP3 0.089 0.137 0.024 0.12 FSTL1_SNP2 0.592 0.969 0.837 0.655 FST-SNP5 0.546 0.931 0.142 0.886 FST-SNP7 0.533 0.479 0.47 FSTL1-SNP1 0.533 0.857 0.685 SST-SNP2 0.479 0.857 0.685 MYO1G-SNP5 0.47 0.685 0.685 CAPN5-SNP17 0.496 0.923 0.651 0.461

347

Appendix 25. Variant interactions for ST adjusted shear force.

CAPN4-two MSTN- CAPN1- CAPN1- bases SNP316 SNP530 F94L insertion/deletion Main effect <.001 0.392 0.751 MSTN-F94L 0.888 0.899 0.593 CAPN1-SNP316 0.888 0.672 0.981 CAPN1-SNP530 0.899 0.672 0.997 CAPN4-two bases insertion/deletion 0.593 0.981 0.997 CAPN5-SNP16 0.996 0.911 0.984 0.808 CAPN5-SNP21 0.862 0.706 0.844 0.917 IGF1R1 0.508 0.958 0.866 0.536 IGF1-SNP1 0.326 0.728 0.77 0.898 IGF1-SNP2 0.949 0.365 0.785 0.964 LOXL1-SNP1 0.725 0.556 0.914 0.37 LOX-SNP1 0.918 0.467 0.751 0.988 MYL7-SNP1 0.885 0.243 0.914 0.731 MYO1G-SNP2 0.519 0.1 0.364 0.153 MYO1G-SNP3 0.457 0.911 0.988 0.629 CAPN4-ten bases insertion/deletion 0.785 0.928 0.957 0.774 SNIP1-SNP3 0.559 <.001 0.121 0.6 FSTL1_SNP2 0.414 0.324 0.614 0.121 FST-SNP5 0.408 0.748 0.946 0.648 FST-SNP7 0.121 0.043 0.098 0.007 FSTL1-SNP1 0.496 0.512 0.808 0.213 SST-SNP2 0.984 0.839 0.965 0.936 MYO1G-SNP5 0.964 0.494 0.826 0.343 CAPN5-SNP17 0.414 0.504 0.474 0.134

348

Appendix 25. Variant interactions for ST adjusted shear force (continued).

CAPN5- CAPN5- IGF1- IGF1- IGF1R1 SNP16 SNP21 SNP1 SNP2 Main effect 0.766 0.431 0.583 0.908 0.824 MSTN-F94L 0.996 0.862 0.508 0.326 0.949 CAPN1-SNP316 0.911 0.706 0.958 0.728 0.365 CAPN1-SNP530 0.984 0.844 0.866 0.77 0.785 CAPN4-two bases 0.917 insertion/deletion 0.808 0.536 0.898 0.964 CAPN5-SNP16 0.734 0.78 0.822 0.724 CAPN5-SNP21 0.734 0.68 0.709 0.594 IGF1R1 0.78 0.68 0.577 0.936 IGF1-SNP1 0.822 0.709 0.577 0.607 IGF1-SNP2 0.724 0.594 0.936 0.607 LOXL1-SNP1 0.558 0.672 0.529 0.685 0.668 LOX-SNP1 0.792 0.863 0.791 0.427 0.616 MYL7-SNP1 0.001 0.205 0.929 0.194 0.712 MYO1G-SNP2 0.246 0.233 0.06 0.387 0.405 MYO1G-SNP3 0.95 0.503 0.943 0.864 0.914 CAPN4-ten bases insertion/deletion 0.865 0.918 0.909 0.856 0.926 SNIP1-SNP3 0.093 0.67 0.556 0.619 0.456 FSTL1_SNP2 0.5 0.283 0.484 0.423 0.606 FST-SNP5 0.927 0.593 0.759 0.772 0.585 FST-SNP7 0.089 0.04 0.129 0.082 0.131 FSTL1-SNP1 0.626 0.362 0.575 0.468 0.741 SST-SNP2 0.807 0.588 0.897 0.248 0.671 MYO1G-SNP5 0.441 0.32 0.638 0.497 0.64 CAPN5-SNP17 0.778 0.796 0.888 0.815 0.41

349

Appendix 25. Variant interactions for ST adjusted shear force (continued).

LOXL1- LOX- MYL7- MYO1G- MYO1G- SNP1 SNP1 SNP1 SNP2 SNP3 Main effect 0.297 0.603 0.52 0.335 0.85 MSTN-F94L 0.725 0.918 0.885 0.519 0.457 CAPN1-SNP316 0.556 0.467 0.243 0.1 0.911 CAPN1-SNP530 0.914 0.751 0.914 0.364 0.988 CAPN4-two bases insertion/deletion 0.37 0.988 0.731 0.153 0.629 CAPN5-SNP16 0.558 0.792 0.001 0.246 0.95 CAPN5-SNP21 0.672 0.863 0.205 0.233 0.503 IGF1R1 0.529 0.791 0.929 0.06 0.943 IGF1-SNP1 0.685 0.427 0.194 0.387 0.864 IGF1-SNP2 0.668 0.616 0.712 0.405 0.914 LOXL1-SNP1 0.657 0.236 0.08 0.61 LOX-SNP1 0.657 0.911 0.134 0.859 MYL7-SNP1 0.236 0.911 0.149 0.884 MYO1G-SNP2 0.08 0.134 0.149 0.368 MYO1G-SNP3 0.61 0.859 0.884 0.368 CAPN4-ten bases insertion/deletion 0.525 0.46 0.065 0.168 0.757 SNIP1-SNP3 0.326 0.413 0.772 0.1 0.815 FSTL1_SNP2 0.393 0.46 0.276 0.02 0.365 FST-SNP5 0.499 0.627 0.852 0.12 0.956 FST-SNP7 0.139 0.164 0.164 0.021 0.219 FSTL1-SNP1 0.421 0.629 0.331 0.039 0.483 SST-SNP2 0.541 0.697 0.534 0.198 0.958 MYO1G-SNP5 0.145 0.489 0.442 0.221 0.672 CAPN5-SNP17 0.509 0.13 0.13 0.109 0.607

350

Appendix 25. Variant interactions for ST adjusted shear force (continued).

CAPN4-ten SNIP1- FST- bases FSTL1_SNP2 SNP3 SNP5 insertion/deletion Main effect 0.601 0.823 0.048 0.229 MSTN-F94L 0.785 0.559 0.414 0.408 CAPN1-SNP316 0.928 <.001 0.324 0.748 CAPN1-SNP530 0.957 0.121 0.614 0.946 CAPN4-two bases insertion/deletion 0.774 0.6 0.121 0.648 CAPN5-SNP16 0.865 0.093 0.5 0.927 CAPN5-SNP21 0.918 0.67 0.283 0.593 IGF1R1 0.909 0.556 0.484 0.759 IGF1-SNP1 0.856 0.619 0.423 0.772 IGF1-SNP2 0.926 0.456 0.606 0.585 LOXL1-SNP1 0.525 0.326 0.393 0.499 LOX-SNP1 0.46 0.413 0.46 0.627 MYL7-SNP1 0.065 0.772 0.276 0.852 MYO1G-SNP2 0.757 0.1 0.02 0.12 MYO1G-SNP3 0.168 0.815 0.365 0.956 CAPN4-ten bases insertion/deletion 0.291 0.583 0.892 SNIP1-SNP3 0.291 0.463 0.72 FSTL1_SNP2 0.583 0.463 0.35 FST-SNP5 0.892 0.72 0.35 FST-SNP7 0.133 0.105 0.013 0.017 FSTL1-SNP1 0.703 0.601 0.32 0.35 SST-SNP2 0.881 0.873 0.212 0.726 MYO1G-SNP5 0.878 0.065 0.059 0.823 CAPN5-SNP17 0.831 0.044 0.488 0.784

351

Appendix 25. Variant interactions for ST adjusted shear force (continued).

FST- FSTL1- SST- MYO1G- SNP7 SNP1 SNP2 SNP5 Main effect 0.076 0.114 0.46 0.422 MSTN-F94L 0.121 0.496 0.984 0.964 CAPN1-SNP316 0.043 0.512 0.839 0.494 CAPN1-SNP530 0.098 0.808 0.965 0.826 CAPN4-two bases insertion/deletion 0.007 0.213 0.936 0.343 CAPN5-SNP16 0.089 0.626 0.807 0.441 CAPN5-SNP21 0.04 0.362 0.588 0.32 IGF1R1 0.129 0.575 0.897 0.638 IGF1-SNP1 0.082 0.468 0.248 0.497 IGF1-SNP2 0.131 0.741 0.671 0.64 LOXL1-SNP1 0.139 0.421 0.541 0.145 LOX-SNP1 0.164 0.629 0.697 0.489 MYL7-SNP1 0.021 0.331 0.534 0.442 MYO1G-SNP2 0.219 0.039 0.198 0.221 MYO1G-SNP3 0.105 0.483 0.958 0.672 CAPN4-ten bases insertion/deletion 0.133 0.703 0.881 0.878 SNIP1-SNP3 0.105 0.601 0.873 0.065 FSTL1_SNP2 0.013 0.32 0.212 0.059 FST-SNP5 0.017 0.35 0.726 0.823 FST-SNP7 0.013 0.065 0.044 FSTL1-SNP1 0.013 0.366 0.095 SST-SNP2 0.065 0.366 0.641 MYO1G-SNP5 0.044 0.095 0.641 CAPN5-SNP17 0.044 0.59 0.828 0.461

352

Appendix 26. Trait variation explained by candidate genes and interactions.

wbld_ wbst_ Lnwbld Lnwbst St_ ST_ LD ST GENE VARIANTS adjusted adjusted wbld26 wbld1 wbst26 wbst1 1_26 1_26 Compression Hydroxyproline Weight Weight

MSTN-F94L NS NS NS NS 4.62~4.72 4.86 NS NS 4.23~4.58 5.58~5.75 1.12~1.50 7.84~8.22

CAPN1-SNP316 4.69~4.76 5.20~5.21 2.16~2.17 3.90~4.06 1.54~1.55 3.02~3.03 NS NS NS NS NS NS

CAPN1-SNP530 NS NS NS 1.41~1.45 NS NS NS NS 2.37~2.72 NS NS NS CAPN4-two bases insertion/deletion and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

CAPN5-SNP10 and MSTN-F94L NS NS NS NS NS NS NS NS NS 2.45 NS NS

CAPN5-SNP21 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS 1.67 2.21

IGF1R-SNP1 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS 1.4 NS

IGF1-SNP1 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

IGF1-SNP2 and MSTN-F94L NS NS NS NS NS NS NS NS NS 3.04 NS NS

LOXL1-SNP1 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

LOX-SNP1 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

MYL7-SNP1 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

MYO1G-SNP2 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

MYO1G-SNP3 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

CAPN4 -3 base repeat and MSTN-F94L NS NS NS NS NS 1.76 NS 4.81 NS 3.1 NS NS

SNIP1-SNP3 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS

FSTL1_SNP2 and MSTN-F94L NS NS NS NS NS NS NS NS 5.18 NS NS NS

FST-SNP5 and MSTN-F94L NS NS NS NS 1.31 NS NS NS NS NS 1.76 NS

FST-SNP7 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS 2.54 NS

FSTL1-SNP1 and MSTN-F94L NS NS NS NS NS NS NS NS 4.85 NS NS NS

SST-SNP2 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS 0.73

353

Appendix 26. Trait variation explained by candidate genes and interactions (continued).

wbld_ wbst_ Lnwbld Lnwbst St_ ST_ LD ST GENE VARIANTS adjusted adjusted wbld26 wbld1 wbst26 wbst1 1_26 1_26 Compression Hydroxyproline Weight Weight MYO1G-SNP5 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS 2.77 NS CAPN5-SNP17 and MSTN-F94L NS NS NS NS NS NS NS NS NS NS NS NS CAPN4-two bases insertion/deletion and NS NS NS NS NS NS NS NS NS NS NS NS CAPN1-SNP316 CAPN5-SNP16 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP21 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS IGF1R-SNP1 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS IGF1-SNP1 and CAPN1-SNP316 NS NS NS NS NS NS NS NS 3.68 NS NS NS IGF1-SNP2 and CAPN1-SNP316 NS NS NS NS 1.73 NS NS NS NS NS NS NS LOXL1-SNP1 and CAPN1-SNP316 NS NS 2.16 NS NS NS NS NS NS NS NS NS LOX-SNP1 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS MYL7-SNP1 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS MYO1G-SNP2 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS MYO1G-SNP3 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS CAPN4 -3 base repeat and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS SNIP1-SNP3 and CAPN1-SNP316 4.9 9.45 2.68 NS 3.49 4.59 NS NS NS NS NS NS FSTL1_SNP2 and CAPN1-SNP316 NS NS NS NS NS 1.44 NS NS NS NS NS NS FST-SNP5 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS 0.94 FST-SNP7 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS FSTL1-SNP1 and CAPN1-SNP316 NS NS NS NS NS 1.49 NS NS NS NS NS NS SST-SNP2 and CAPN1-SNP316 NS NS NS 2.11 NS NS 1.66 NS NS NS NS NS MYO1G-SNP5 and CAPN1-SNP316 NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP17 and CAPN1-SNP317 NS NS NS NS NS NS NS NS NS NS NS NS

354

Appendix 26. Trait variation explained by candidate genes and interactions (continued).

wbld_ wbst_ Lnwbld Lnwbst St_ ST_ LD ST GENE VARIANTS adjusted adjusted wbld26 wbld1 wbst26 wbst1 1_26 1_26 Compression Hydroxyproline Weight Weight CAPN4-two bases insertion/deletion and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP16 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP21 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS IGF1R-SNP1 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS IGF1-SNP1 and CAPN1-SNP530 NS NS NS NS NS NS NS 4.77 NS NS NS NS IGF1-SNP2 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS LOXL1-SNP1 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS 1.6 NS LOX-SNP1 and CAPN1-SNP530 NS NS NS NS NS NS 3.14 4.34 NS 3.93 NS NS MYL7-SNP1 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS MYO1G-SNP2 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS MYO1G-SNP3 and CAPN1-SNP530 NS NS NS NS NS NS NS 2.12 NS NS NS NS CAPN4 -3 base repeat and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS 1.42 SNIP1-SNP3 and CAPN1-SNP530 2.71 NS NS 2.27 NS 2.8 NS NS NS NS NS NS FSTL1_SNP2 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS FST-SNP5 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS FST-SNP7 and CAPN1-SNP530 2.45 1.7 NS NS NS NS NS NS NS NS NS NS FSTL1-SNP1 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS SST-SNP2 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS MYO1G-SNP5 and CAPN1-SNP530 NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP17 and CAPN1-SNP531 NS NS NS NS NS NS NS NS NS NS NS NS

355

Appendix 27. Additive effects of candidate genes for tenderness traits.

wbld_ wbst_ Lnwbld Lnwbst St_Compression ST_Hyd GENE VARIANT wbld26 wbld1 wbst26 wbst1 LD Weight ST Weight pH ld pH st cl ld cl st adjusted adjusted 1_26 1_26 _kgf _mg_g CAPN4-two bases NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS insertion/deletion

CAPN5-SNP16 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

-0.223 ± -0.1098 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP21 0.105 Kg 0.0405 Kg

IGF1R-SNP1 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

-0.238 ± 0.124 -0.236 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS IGF1-SNP1 mg/g 0.115 kg

IGF1-SNP2 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

-0.1534 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS LOXL1-SNP1 0.0752 kgF 0.0322 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS LOX-SNP1 0.0154 (ln kgF)

MYL7-SNP1 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

0.0696 ± 0.0475 ± NS NS NS NS NS NS NS NS NS NS NS NS NS MYO1G-SNP2 0.0187 0.0187

MYO1G-SNP3 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

CAPN4 -3 base repeat NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

0.237 ± 0.278 ± 0.104 NS NS NS NS NS NS -0.097 ± 0.425 kg NS NS 0.2249 ± 0.078 NS NS NS NS SNIP1-SNP3 0.084 kgF kgF 0.358 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS FSTL1_SNP2 0.157 kgF -0.1443 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS FST-SNP5 0.0862 kgF 0.285 ± 0.378 ± NS NS NS NS NS NS NS NS NS 1.125 ± 0.148 NS NS NS NS FST-SNP7 0.122 kgF 0.176 kgF -0.374 ± NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS FSTL1-SNP1 0.174 kgF

SST-SNP2 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

MYO1G-SNP5 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

-0.0234 ± -0.02354 0.345 ± NS NS NS NS NS NS NS NS NS NS NS NS NS CAPN5-SNP17 0.0082 ± 0.00724 0.14

356