Bayesian genome-wide association analysis of growth and yearling ultrasound measures of carcass traits in Brangus heifers S. O. Peters, K. Kizilkaya, D. J. Garrick, R. L. Fernando, J. M. Reecy, R. L. Weaber, G. A. Silver and M. G. Thomas

J ANIM SCI 2012, 90:3398-3409. doi: 10.2527/jas.2011-4507 originally published online June 4, 2012

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://www.journalofanimalscience.org/content/90/10/3398

www.asas.org

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Bayesian genome-wide association analysis of growth and yearling ultrasound measures of carcass traits in Brangus heifers1

S. O. Peters,*†2 K. Kizilkaya,‡§ D. J. Garrick,‡# R. L. Fernando,‡ J. M. Reecy,‡ R. L. Weaber,†3 G. A. Silver,* and M. G. Thomas*||4

*Department of Animal and Range Sciences, New Mexico State University, Las Cruces 88003; †Department of Animal Sciences, University of Missouri, Columbia 65211; ‡Department of Animal Science, Iowa State University, Ames 50011; §Department of Animal Science, Adnan Menderes University, Aydin 09100, Turkey; #Institute of Veterinary, Animal and Biomedical Sciences, Massey University, Palmerston North, New Zealand; and ||Department of Animal Sciences, Colorado State University, Fort Collins 80523

ABSTRACT: Data from developing Brangus heifers variation. The SNP windows with P < 0.01 were (3/8 Brahman-Bos indicus × 5/8 Angus-Bos taurus; n ≈ considered as QTL associated with a trait in which case 802 from 67 sires) registered with International Brangus their location was queried from dbSNP and the presence Breeders Association were analyzed to detect QTL of a previously reported QTL in that location was checked associated with growth traits and ultrasound measures in CattleQTLdb. For 9 traits, QTL were mapped to 139 of carcass traits. Genotypes were from BovineSNP50 regions on 25 chromosomes. Forty-one of these QTL (Infinium BeadChip, Illumina, San Diego, CA; 53,692 were already described in CattleQTLdb, so 98 are new SNP). Phenotypes included BW collected at birth additions. The SNP windows on chromosomes 1, 3, and and ~205 and 365 d of age, and yearling ultrasound 6 were associated with multiple traits (i.e., 205- and 365- assessment of LM area, percent intramuscular fat, d BW, and ADG from birth to 205 and 365 d of age). and depth of rib fat. Simultaneous association of SNP Several chromosomes harbored regions associated with windows with phenotype were undertaken with Bayes C multiple traits; however, the SNP that comprised the analyses, using GenSel software. The SNP windows were window often varied among traits (i.e., chromosomes 1, ≈ 5 SNP in length. Analyses fitted a mixture model that 3, 4, 5, 6, 7, 9, 10, 11, 13, 14, 15, 16, 20, 21, 22, 24, 28, treated SNP effects as random, with an assumed fraction and 29). Results from whole genome association of SNP pi = 0.999 having no effect on phenotype. Bootstrap with growth and ultrasound carcass traits in developing analyses were used to obtain significance values for the Brangus heifers confirmed several published QTL and SNP windows with the greatest contribution to observed detected several new QTL.

Key words: cattle, genome, growth, quantitative trait loci, single nucleotide polymorphism

© 2012 American Society of Animal Science. All rights reserved. J. Anim. Sci. 2012.90:3398–3409 doi:10.2527/jas2012-4507 INTRODUCTION

The abundance of SNP in the bovine genome 1Financial support provided by USDA-AFRI (Grant no. 2008-35205- 18751 and 2009-35205-05100) and New Mexico Agric. Exp. Stan. Project and their ease of genotyping make them effective for (Hatch #216391). Collaboration developed from activities of National genomic prediction and genome-wide association Beef Cattle Evaluation Consortium. Authors acknowledge Camp Cooley studies (Georges, 2007; Gibbs et al., 2009; Goddard Ranch (Franklin, TX) for supplying DNA and phenotypes from Brangus heifers, and Robert Schnabel, University of Missouri, for providing SNP and Hayes, 2009). These processes were enhanced with information for BovineSNP50. construction of a high density SNP genotyping assay 2Current address: Department of Animal Sciences, Berry College, (Matukumalli et al., 2009). Since this invention, QTL Mount Berry, GA 30149 associations with measures of growth and reproduction 3Current address: Department of Animal Sciences and Industry, Kansas State University, Manhattan 66506. have been reported in beef cattle (Snelling et al., 2010; 4Corresponding author: [email protected] Bolormaa et al., 2011a,b,c; Hawken et al., 2011). Received July 24, 2011. Accepted April 26, 2012. 3398

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Genome-wide association with heifer growth 3399

Genomic selection procedures were quickly yearling, and yearling ultrasound assessment of LM area, implemented in dairy cattle after the development of the percent fat within LM, and fat thickness over the 12th rib. high density SNP genotyping assay. This success was Contemporary group and date that each of these phenotypes plausible because of the data structure, business model, were recorded were also extracted from the database. and germplasm resources for Holstein cattle (Hayes et Nongenetic effects of age at measurement were taken into al., 2009; Hayes and Goddard, 2010). The beef industry account by adjusting BW (i.e., 205 and 365 d of age) and is challenged in these types of efforts because of its ultrasound measures (i.e., 365 d of age), using formulas numerous breeds in the seedstock segment, of which from the Brangus Herd Improvement Records (2010). some are multibreed composites, and its widespread use Adjusted BW values were used to calculate preweaning of crossbred calves in the production segments of the and postweaning ADG, and ADG from birth to 365 d of industry. Nonetheless, genomic prediction and genome- age as the difference in kilograms between the relevant BW wide association studies have been reported in multibreed measures divided by 205, 160, or 365 d, respectively. The populations using high density SNP genotypes (Snelling subscript B in these formulae (i.e., ADGB-205, ADG205-365, et al., 2010; Toosi et al., 2010; Bolormaa et al., 2011a,b,c). ADGB-365) represent the BW collected at birth. Initial genome-wide association studies used In this study, cows that gave birth to these heifers pedigree-based REML estimations with Bonferroni were managed by the various cooperators of Camp adjustments to infer QTL from single SNP effects. Cooley Ranch and New Mexico State University, and These methods were challenged to avoid false positives were bred by estrous synchronization and then AI and overestimation of QTL effects; therefore, Bayesian followed by brief periods of exposure to natural service approaches were developed as they offer methods to sires. These breeding seasons ranged from 60 to 90 d. mitigate these challenges (Garrick, 2009; Stromberg, After weaning, heifers were developed at Camp Cooley 2009; Cantor et al., 2010). In this study, we conducted Ranch and the New Mexico State University Campus whole genome association analyses using high density Farm, as described by Luna-Nevarez et al. (2010). SNP genotypes and growth and yearling ultrasound There were pedigree connections via 5 AI sires that measures of carcass traits from Brangus heifers (3/8 were common to the 2 groups of heifers and numerous Brahman × 5/8 Angus). The QTL were inferred from other familial relations via historic AI sires registered Bayes C analyses, which considered SNP as random with International Brangus Breeders Association. Mean effects and evaluated these effects simultaneously. inbreeding coefficient of these heifers was 8.24 ± 0.001 and 67 sires were represented in the data. It should also MATERIALS AND METHODS be noted that in these data there was only 1 calf per dam.

Animals were handled and managed accord- DNA and Genotypes ing to Institutional Animal Care and Use Committee Guidelines. A single 5-mL blood sample from each heifer was collected with vacutainer tubes coated with EDTA and Heifers and Phenotypes shipped to New Mexico State University. Tubes were centrifuged 1,875 × g at 4°C for 30 min and white Data were from heifers registered with International blood cell supernatant (i.e., buffy coat) recovered, using Brangus Breeders Association (San Antonio, TX; 3/8 procedures described by Beauchemin et al. (2006) and Brahman × 5/8 Angus). As typical of seedstock animals, Thomas et al. (2007). The Flexigene kit and procedure 890 heifers had various phenotypes, blood samples were (#51204; Qiagen, Valencia, CA) was used to extract obtained from 855 heifers, 835 heifers had adequate blood DNA from the samples. Genotyping was performed by sample for DNA extraction, and successful genotypes Advanced Genomics Technology Center (Fairfax, VA). were obtained from an average of 802 heifers. Genotyping Genotyping used 100 ng of DNA in 100 µL of nuclease- success ranged from 761 to 811 heifers for 9 traits. Heifers free water per sample and BovineSNP50 (i.e., Infinium were raised in 2 locations, Camp Cooley Ranch (Franklin, BeadChip, Illumina, San Diego, CA), as described by TX) in east central Texas, and the Chihuahuan Desert Matukumalli et al. (2009). Genotype call rates averaged Rangeland Research Center and Campus Farm of New 98.1 ± 0.001% for 53,692 SNP. Genotypes were in the Mexico State University, described by Luna-Nevarez et Illumina A/B allele format and were used to compute al. (2010). The following information was queried from a value at each locus coded as 0, 1, or 2, representing the databases of these organizations for each heifer: date the number of B alleles. If missing genotypes existed and year of birth (2005 to 2007), calving season (spring for a particular locus, the average value for that locus in or autumn), pedigree, birth method (AI, embryo transfer, the population was inserted so each heifer had complete or natural service), BW measures at birth, weaning, and genotype data. Figure 1 presents the minor allele

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 3400 Peters et al. frequencies estimated using procedures of Garrett et al. described by Kizilkaya et al. (2010) and Habier et al. (2008). (2011). Bayes-C used all SNP data simultaneously and assumed a common variance for all SNP loci. The model Descriptive Statistics and Heritability equation was:

K Descriptive statistics for growth, composition, and y = Xb + ∑x jβ jδ j + e j =1 ADG were estimated using PROC MEANS (SAS Inst. Inc., Cary, NC). Assumptions of normality of these data where y was the vector of the phenotypic values, X was were tested and visualized using PROC UNIVARIATE the incidence matrix for fixed effects, b was the vector of and GPLOT, respectively. Independent sources of fixed effects that accounted for cohort groups (defined for variation were tested for significance in a mixed model animals with the same calving season, location, and trait within PROC MIXED, so the information could be used contemporary group) defined as classes and dam age (in to construct models for estimating heritability. Fixed years) as classes, K was the number of SNP loci (53,692), effects considered were year and location of birth, calving xj was the column vector representing the SNP covariate season, age of dam, birth method, and contemporary at locus j coded as the number of B alleles, βj was the group classes and covariates for age at measurement. random substitution effect for locus j, which conditional 2 2 Random effects fitted in these models included sire on σ β was assumed normally distributed N (0, σ β ) when 2 (i.e., mean = 0, variance = σs ; Z statistic used to test δj = 1 but βj = 0 when δj = 0, with δj being a random 0/1 2 2 if Ho: σs = 0) and residual (mean = 0, variance = σe ). variable indicating the absence (with probability π) or Narrow sense heritability estimates were obtained presence (with probability 1-π) of locus j in the model, and from MTDFREML (K. G. Boldman, USDA-ARS, e was the vector of the random residual effects assumed 2 2 2 unpublished data) to use as priors in Bayesian analyses. normally distributed N (0, σ e ). The variance σ β (or σ e) Standard error of heritability estimates were based on was a priori assumed to follow a scaled inverse Chi-square the average information matrix and the “delta” method with vβ = 4 (or ve = 10) degrees of freedom and scale 2 2 (Dodenhoff et al., 1998). It should also be noted that parameter S β (or S e). The scale parameter was derived models in this effort, and the following whole genome as a function of the assumed known genetic variance of association analyses, were of direct effects as dams the population, based on the average SNP allele frequency only had 1 offspring each and none of the dams were and number of SNP assumed to have nonzero effects genotyped. It was therefore not possible to fit a genomic following Fernando et al. (2007). Parameter π was 0.999. effect for the dam or to adjust for individual dam effects. This procedure was implemented in GenSel (Fernando Dam (i.e., maternal) effects, therefore, contributed to the and Garrick, 2009). The program used Markov-Chain residual variance. Monte Carlo approaches to estimate the effect of each SNP among all SNP at each interation. This procedure involved Whole Genome Association a burn-in period of 1,000 iterations from which results were discarded, followed by 45,000 iterations from which Whole genome association of SNP with phenotypes results were accumulated to obtain the posterior mean were undertaken with Bayes-C, which was a modification effect of each SNP. of the Bayes-B method and used sampling procedures The effect of any particular QTL may be distributed across numerous SNP in single locus or multilocus linkage disequilibrium with QTL, resulting in individual SNP effects that tend to underestimate the real QTL effect (Fan et al., 2011; Onteru et al., 2011). Accordingly, the posterior means of SNP effects were collectively used to predict the genomic merit of sliding chromosomal regions, including 5 consecutive SNP based on physical map order. At each window, the genetic variance was calculated as the variance of the genomic merit for that window and expressed as a proportion of the variance of whole genome merit. Specifically, variance was calculated as a simple variance of the samples of the breeding values for the window or whole genome, accounting for the fact that the expected value of both Figure 1. Histogram of minor allele frequencies in Brangus heifers (n these variables is 0. Window genetic variance was ≈ 802) genotyped with BovineSNP50 (Illumina, San Diego, CA; i.e., 53,692 SNP). computed by multiplying the number of alleles that

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Genome-wide association with heifer growth 3401 Table 1. Descriptive statistics and heritability (h2) represent the SNP covariates for each consecutive estimates for growth, ADG, and ultrasound measures SNP in a window by their respective posterior means of carcass traits in Brangus heifers. for substitution effects. After computing these 5 SNP window breeding values for all animals in the population, Traits N Mean ± SE h2 ± SE the variance of these breeding values was calculated. Birth weight, kg 882 35.18 ± 0.38 0.23 ± 0.076 Windows that contributed the greatest genetic variance 205-d weight, kg 878 239.13 ± 2.56 0.27 ± 0.105 were considered to be the strongest signals of association. 365-d weight, kg 830 361.91 ± 3.35 0.38 ± 0.097 There were 10,738 unique SNP windows with at ADG(B-205d), kg/d 878 1.17 ± 0.01 0.29 ± 0.100 ADG , kg/d 830 0.75 ± 0.005 0.22 ± 0.081 least 5 consecutive SNP in the genome. The proportion (205-365) ADG(B-365), kg/d 880 0.98 ± 0.01 0.34 ± 0.098 of variance accounted for by a typical window was 9.3 LM area1, cm2 830 63.10 ± 0.05 0.50 ± 0.110 –5 × 10 (1/10,738.4). The SNP windows with the greatest Intramuscular fat1, % 830 4.25 ± 0.03 0.39 ± 0.106 genetic variance were considered to be the most important Rib fat1, cm 830 0.61 ± 0.01 0.40 ± 0.112 regions associated with the trait and were inferred as 1Adjusted to 365 d of age. The subscript B in formulae of ADG represent QTL at that locus. Sometimes, windows accounting for the BW collected at birth. a high proportion of genetic variance were contiguous, in which case QTL was defined as comprising 6, 7, or were constructed according to the null hypothesis of no even 10 consecutive SNP. A detected QTL was viewed QTL in the identified SNP window. Bootstrap samples as supported by a previously reported QTL if their were reanalyzed using the Bayes C model that was used regions overlapped. The position of previously reported for the raw data to construct the distribution of the test QTL in centimorgan was approximately scaled to mega- statistics for each putative QTL. The 1,000 estimates of base (Mb), using procedures described by Ogorevc et genetic variance for the SNP window corresponding to al. (2009) and Mai et al. (2010), and all loci herein were the QTL were accumulated for comparison to the test described using procedures of Matukumalli et al. (2009; statistic represented by the genetic variance of the SNP Bos taurus version 4.0). The estimated proportion of window identified in the analysis of the real data. genetic variance contributed by sliding windows of The decision criterion for significance was: if consecutive SNP was plotted against genomic marker only 1 bootstrap statistic from the 1,000 simulated locations using the R statistical package (Institute for exceeded the test statistic from the raw data, the Statistics and Mathematics, Wien, Austria). comparison-wise P-value was determined to be 0.001 < P < 0.002. Only SNP windows with P < Bootstrap Analyses for Hypothesis Testing 0.01 were considered as QTL associated with that trait. Multiple testing was taken into account using The procedure for bootstrap analysis and hypothesis the probability of false positives as described in testing to determine the significance of detected QTL Fernando et al. (2004), rather than considering the was described by Fan et al. (2011) and Onteru et al. probability of a type-I error across all the analyses. (2011). In brief, 1,000 bootstrap data sets for each When SNP were found to be of significance in these trait were produced by sampling a residual effect and boot strap analyses, their location was queried from adding it to the genomic merit computed, using the dbSNP (2011) and the occurrence of a previously posterior means of the 53,692 SNP, excluding those reported QTL was queried from CattleQTLdb SNP that defined a QTL. That is, bootstrap samples (Release 14; 2011).

Table 2. Posterior means ± SE of variance components explained by whole genome SNP markers (i.e., BovineSNP50, Illumina, San Diego, CA; 53,692 SNP) for growth, ADG, and ultrasound measures of carcass traits in Brangus heifers

Traits Number of animals Genetic variance Residual variance Total variance Proportion of phenotypic variance explained by SNP Birth weight 811 14.64 ± 0.17 96.52 ± 0.22 111.16 0.13 205-d weight 794 118.48 ± 3.82 2766.50 ± 6.39 2884.98 0.04 365-d weight 748 1076.33 ± 10.9 4691.22 ± 12.5 5767.55 0.19

ADG(B-205d) 806 0.01 ± 0.001 0.06 ± 0.001 0.07 0.11 ADG(205-365) 806 0.02 ± 0.001 0.08 ± 0.002 0.10 0.18 ADG(B-365) 816 0.01 ± 0.001 0.03 ± 0.001 0.04 0.19 LM area1 759 0.31 ± 0.003 1.10 ± 0.003 1.41 0.22 IMF1,2, % 761 0.19 ± 0.001 0.48 ± 0.001 0.67 0.28 Rib fat1 761 0.00073 ± 0.001 0.0037 ± 0.00001 0.004 0.17 1Adjusted to 365 d of age. 2Intramuscular fat = IMF. The subscript B in formulae of ADG represent the BW collected at birth.

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 3402 Peters et al. Table 3. Identification of SNP associated with birth weight in Brangus heifers (n = 811) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. SNP ID of Position ID of Position Boot strap Proportion Previously identified QTL2 Chr1 in window first SNP of first SNP last SNP of last SNP P-value of variance 1 5 rs29012753 48219435 ss61471396 48354950 0.001 0.0017 Yes 3 8 ss117962974 34410354 rs42785682 34658322 0.001 0.0017 Yes 5 6 ss86305510 50816331 rs29016809 51770121 0.003 0.0032 Yes 10 9 ss11796147 648931 ss86327353 1019175 0.010 0.0110 No 13 9 ss105240518 39371489 ss61467241 39774355 0.001 0.0070 No 14 9 ss117971275 52872811 ss117970972 53079949 0.006 0.0033 Yes 14 9 rs42413988 54100691 ss61475992 54571047 0.003 0.0031 Yes 14 8 ss86279944 56448146 rs43138491 56769637 0.003 0.0031 Yes 21 7 ss117965482 3827973 ss86283324 4068720 0.001 0.0011 Yes 28 9 ss86293148 42058631 ss86309667 42280086 0.002 0.0020 Yes 1Chr = chromosome number 2CattleQTLdb (2011)

RESULTS defining the window varied (i.e., chromosomes 1, 3, 4, 5, 6, 7, 9, 10, 11, 13, 14, 15, 16, 20, 21, 22, 24, 28, and 29). Table 1 contains descriptive statistics and heritability estimates for growth traits, ultrasound measures of DISCUSSION carcass characteristics, and ADG. Table 2 describes the proportion of phenotype variance accounted by SNP Bos taurus × Bos indicus crossed cattle have been windows. Supplemental Figures 1 to 9 (see Supplemental successfully used to detect QTL for growth and carcass Material found online at http://journalofanimalscience. traits in cattle (Bolormaa et al., 2011a,b,c; Imumorin et org) allow visualization of the proportion of genetic al., 2011; Hawken et al., 2011). Therefore, we expected variation explained by windows across the chromosomes to detect associations among high density SNP genotypes for each trait. Tables 3 to 11 reported the SNP that defined and growth and yearling ultrasound measures in heifers of the windows (i.e., QTL) associated with the 9 traits. The the 2-breed composite known as Brangus. The seedstock presence or absence of a QTL within CattleQTLdb (2011) organizations that contributed heifers to this study were is also noted in each table. Supplemental Table 1 contains previously involved in studies of growth and ultrasound the effect of each SNP within these windows for the 9 measures of carcass traits (Shirley et al., 2006; Lancaster traits, as well as the model frequency from Bayesian et al., 2009; Luna-Nevarez et al., 2010). Because of these analyses and frequency of the B allele. publications, we had knowledge that the heifers in this In brief, 139 regions on 25 chromosomes were study were typical of those registered with the International associated with the 9 traits studied in this population of Brangus Breeders Association. Also, the heritability Brangus heifers. Ninety-nine regions will be new listings estimates in the current study and those of previous studies in CattleQTLdb (2011). Specific SNP windows on that used records from International Brangus Breeders chromosomes 1, 3, and 6 were associated with multiple Association appeared to be within a similar range of SE traits (i.e., 205- and 365-d BW, and ADG from birth to (Moser et al., 1998; Stelzleni et al., 2002; Lancaster et al., 205 and 365 d of age). Several chromosomes harbored 2009). The proportion of phenotypic variance explained regions associated with multiple traits; however, the SNP

Table 4. Identification of SNP associated with 205-d weight in Brangus heifers (n = 794) genotyped with BovineSNP50

No. SNP ID of Position ID of Position Boot strap Proportion Previously Chr1 in window first SNP of first SNP last SNP of last SNP P-value of variance identified QTL2 3 7 ss86303944 90951609 ss64552970 91153077 0.001 0.0019 No 10 9 ss117969863 33126373 ss86273438 33492966 0.010 0.0203 Yes 16 6 ss86312250 446026 rs41787407 589370 0.001 0.0040 No 16 7 ss61536681 3144808 ss105263670 3324216 0.001 0.0011 Yes 29 5 rs42171465 25907844 ss86286136 26051874 0.001 0.0018 No Un.0043 10 ss61532133 90440 ss117964972 92340 0.001 0.0017 No 1Chr = chromosome number. 2CattleQTLdb (2011). 3Un.004 = unmapped SP on BovineSNP50 (Illumina, San Diego, CA).

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Genome-wide association with heifer growth 3403 Table 5. Identification of SNP associated with 365-d weight in Brangus heifers (n = 748) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. SNP ID of Position of ID of Position of Boot strap Proportion Previously Chr1 in window first SNP first SNP last SNP last SNP P-value of variance identified QTL2 1 7 ss61522637 109643162 ss117966342 109917172 0.002 0.0028 No 1 6 ss86318478 114010037 rs41897673 114188685 0.003 0.0032 No 1 9 ss117975147 138912430 rs43275053 139444528 0.001 0.0014 Yes 6 9 ss86338115 55038931 rs42005088 55293701 0.003 0.0035 Yes 9 9 rs42516892 57480772 rs42937117 57732397 0.002 0.0026 No 10 9 ss117969863 33126373 ss86273438 33492966 0.008 0.0089 No 20 6 ss117972683 8435198 ss105265377 8576253 0.001 0.0014 Yes 22 8 rs42011564 47229871 ss86335704 47486853 0.001 0.0015 Yes 29 8 rs42161771 2086575 ss86320129 2347847 0.014 0.0014 No 1Chr = chromosome number. 2CattleQTLdb (2011).

by SNP was also within the realm reported by Fan et al. improved if maternal effects could be taken into account. (2011) and Onteru et al. (2011). Nonetheless, direct effects are heritable, so the analysis Models with direct genetic effects were used in this still has the power to find QTL affecting direct effects of study for reasons presented in Materials and Methods. growth (Kizilkaya et al., 2010; Fan et al., 2011; Onteru The power to estimate direct effects would be greatly et al., 2011). Publications describing QTL are becoming

Table 6. Identification of SNP associated with ADG(B-205) in Brangus heifers (n = 806) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. of SNP ID of Position of ID of Position of Boot strap Proportion Previously Chr1 in window first SNP first SNP last SNP last SNP P-value of variance identified QTL2 1 7 ss117966268 70862231 ss61508107 71169378 0.0008 0.0008 No 1 7 ss117966381 121159769 rs43267277 121411624 0.0002 0.0003 No 3 7 ss86322668 63240606 ss863084484 63667144 0.0006 0.0007 No 3 9 ss61489204 68215262 ss86338115 68451692 0.0003 0.0028 No 3 9 rs42662796 81172893 rs43344383 81510110 0.0005 0.0007 No 3 9 ss86303944 90951609 ss61506197 91216203 0.0020 0.0004 No 6 7 rs29019575 88946762 ss61557767 89150760 0.0002 0.0006 No 7 7 ss86295884 18434503 ss86298195 18836717 0.0010 0.0004 No 7 5 ss86290798 91803672 ss86294154 92215011 0.0020 0.0004 No 9 10 rs43032893 56851107 ss117969560 57185609 0.0010 0.0003 No 9 7 rs43602118 62646204 ss61561477 62878884 0.0004 0.0005 No 9 7 ss86326094 99589545 ss86299907 99911919 0.0003 0.0004 No 10 6 ss46526487 46013388 ss86332513 46167453 0.0010 0.0003 No 10 6 ss61530388 71288958 ss117969909 71506530 0.0008 0.0009 No 10 8 ss86335790 71340536 ss86285742 71596515 0.0008 0.0009 Yes 11 7 ss86318620 48740271 ss61512380 49046757 0.0002 0.0003 No 13 10 ss86306247 73598551 ss86291065 73917944 0.0007 0.0009 No 14 5 rs41730524 28288825 ss61534671 28483106 0.0008 0.0005 Yes 14 6 ss117971634 28344925 ss61534686 28568867 0.0009 0.0009 Yes 14 5 rs41731629 28451198 ss86299412 28604029 0.0004 0.0004 No 14 5 rs41735977 30950352 rs41732036 31305252 0.0003 0.0005 Yes 15 7 ss86314475 65995470 ss86335571 66214629 0.0004 0.0004 No 16 9 ss61536681 3144808 ss61537121 3374361 0.0006 0.0007 No 16 6 ss86332308 7380169 ss61498648 7569529 0.0010 0.0003 No 21 6 rs43108270 23258748 ss61529171 23596278 0.0003 0.0003 No 24 7 ss61502923 43203019 ss61547545 43539439 0.0110 0.0005 No 26 5 ss117974136 46456797 ss86283448 46608558 0.0003 0.0003 No 29 5 rs43186847 22076503 rs29016832 22291347 0.0010 0.0003 No 29 5 ss86285321 22243929 ss86295434 22398514 0.0010 0.0005 No 1Chr = chromosome number. 2CattleQTLdb (2011).

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 3404 Peters et al.

Table 7. Identification of SNP associated with(205-365) ADG in Brangus heifers (n = 806) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. SNP ID of Position ID of Position Boot strap Proportion Previously Chr1 in window first SNP of first SNP last SNP of last SNP P-value of variance identified QTL2 3 8 ss86275112 125637877 ss86295607 125868208 0.001 0.0005 No 4 5 ss86287884 53654311 rs43390793 53849143 0.001 0.0004 No 6 9 rs29024115 32154208 ss86286530 32935120 0.007 0.0071 No 6 9 ss61556931 33328168 ss117968742 33620176 0.005 0.0058 No 6 9 rs42591335 42892480 ss61504167 43130903 0.009 0.0098 Yes 6 5 rs43465018 55080375 ss61545179 55222361 0.001 0.0004 No 7 5 ss61523901 55577265 ss86336860 55761848 0.001 0.0007 No 9 8 ss61523768 52865467 ss61525621 53145976 0.001 0.0006 No 9 7 ss61499765 52952159 ss117969547 53198488 0.001 0.0007 No 11 6 rs42894509 19267868 rs42569290 19401730 0.001 0.0003 Yes 15 9 ss86322971 76177610 ss86334585 76602651 0.004 0.0048 No 15 8 ss86337312 77419466 ss86315494 77662776 0.001 0.0008 No 16 7 rs41751080 55249873 ss86283813 55635029 0.001 0.0003 Yes 17 6 ss61508541 24585868 ss86293774 24887668 0.001 0.0006 Yes 24 9 ss86307923 56189555 ss86314380 56650602 0.001 0.0007 No 29 5 ss117965903 4372728 ss61565884 4573663 0.003 0.0010 No 29 5 ss61565884 4573663 ss61551529 4733645 0.001 0.0005 No 1Chr = chromosome number. 2CattleQTLdb (2011). voluminous and thus providing a need and utility for tools of this approach is an improvement in genome-wide such as CattleQTLdb (Hu et al., 2007). In the current association studies, as spurious associations can easily study, a sliding-SNP-window approach identified 139 occur from a single locus (Balding, 2006; Rosenberg et QTL associated with 9 traits. This approach accounted al., 2011). It is important to note that gene delineation for linkage disequilibrium among adjacent SNP, which efforts (i.e., fine mapping) are needed for QTL detected helped validate important chromosome regions. Use with SNP windows from a genotyping platform of

Table 8. Identification of SNP associated with ADG(B-365) in Brangus heifers (n = 816) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. SNP ID of Position ID of Position Boot strap Proportion Previously Chr1 in window first SNP of first SNP last SNP of last SNP P-value of variance identified QTL2 1 7 ss105242396 113879898 rs41897688 114148558 0.004 0.0041 No 1 7 rs43264628 114043777 rs43266821 114246014 0.003 0.0050 No 1 9 ss117966381 121159769 rs43259942 121510242 0.004 0.0034 No 1 9 ss117975147 138912430 rs432750053 139444528 0.001 0.0014 No 6 6 ss86338115 55038931 ss61545179 55222361 0.002 0.0021 No 6 5 ss42007442 55196036 rs42005088 55293701 0.002 0.0021 No 7 10 ss432511148 40540807 rs42368614 40932161 0.009 0.0090 No 9 5 rs43597980 57438109 ss61514383 57595191 0.001 0.0058 No 9 5 rs42516892 57480772 ss43113305 57618813 0.003 0.0037 No 9 7 rs42686238 57515851 rs42519306 57697893 0.008 0.0084 No 9 5 ss61563774 57643837 rs42827834 57796887 0.001 0.0014 No 9 7 ss86291594 98214839 ss86341224 98377340 0.002 0.0021 No 9 6 ss86324787 98298320 ss86334265 98450642 0.001 0.0018 No 10 9 ss117969863 33126373 ss86273438 33492966 0.008 0.0090 No 15 8 ss61505280 60329509 ss86278765 60592033 0.003 0.0032 No 16 9 ss117971976 2418748 ss86341387 2759082 0.009 0.0094 No 20 5 ss86295747 4972476 rs42661291 5089822 0.001 0.0094 No 20 8 ss86284812 5013126 ss117972856 5192679 0.003 0.0039 No 20 8 ss117972842 8396346 ss86317572 8605796 0.002 0.0023 No 22 9 ss86316001 47288808 ss86307923 47624602 0.002 0.0013 No 1Chr = chromosome number. 2CattleQTLdb (2011).

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Genome-wide association with heifer growth 3405 Table 9. Identification of SNP associated with LM area in Brangus heifers (n = 759) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. of SNP ID of Position ID of Position Boot strap Proportion Previously Chr1 in window first SNP of first SNP last SNP of last SNP P-value of variance identified QTL2 1 7 rs29024142 123704792 rs43263816 123888830 0.003 0.0035 Yes 2 5 ss86308738 112633455 rs42431462 112757656 0.01 0.0156 No 2 7 ss61473895 135463866 ss61519815 135680207 0.001 0.0022 No 4 9 ss86318042 110540812 ss86340434 110933741 0.002 0.0023 No 4 9 ss86287715 111916517 ss86300376 112250879 0.010 0.0147 No 5 9 ss61485769 58820908 ss61545754 59482668 0.001 0.0015 Yes 5 7 ss86338260 59659722 ss117967571 60075185 0.001 0.0011 Yes 5 10 ss61510435 60570822 ss86297204 61244601 0.004 0.0041 Yes 9 7 ss61561516 67183717 ss86280803 67470723 0.003 0.0032 Yes 18 13 ss61539708 54954740 ss61482909 55625797 0.001 0.0016 No 20 7 rs29013174 11906213 ss86338324 12073761 0.003 0.0033 No 20 5 ss86317797 11995292 ss61567871 12102095 0.007 0.0075 No 20 7 ss86338324 12073761 ss61530291 12301703 0.001 0.0011 No 20 7 ss86305979 19555756 ss117972639 19856633 0.001 0.0013 No 20 6 ss86295087 19664378 ss61520379 19982228 0.001 0.0011 No 21 9 ss105262810 6661836 ss86297800 6913978 0.001 0.0015 Yes 21 9 rs41968436 9141174 ss61544915 9443901 0.001 0.0015 Yes 24 6 ss61547319 29266446 ss86341099 29474704 0.002 0.0025 No 1Chr = chromosome number. 2CattleQTLdb (2011).

median marker interval of ~37 kb (Matukumalli et CattleQTLdb (2011) was probably a more interesting al., 2009). Many genes will likely underlie such broad result. McClure et al. (2010) described similar results in chromosome regions and causal mutations may be a study of Angus cattle, as only 118 of 673 QTL detected engulfed within blocks of linkage disequilibrium were previously cited in literature. These results suggest (Georges, 2007; Goddard and Hayes, 2009; Karim et al., that tools such as CattleQTLdb (2011) will grow 2011). Nonetheless, potential candidate genes underlying substantially from the use of high density genotyping QTL are of great interest and will be discussed herein. platforms in genome-wide association studies. However, Most of the 41 chromosome regions detected in this it is interesting to note that neither study found association study were found to be in CattleQTLdb (2011). The with traits indicative of marbling (i.e., intramuscular fat 98 QTL detected in this study that were not present in percent) and the region of chromosome 14 that harbors Table 10. Identification of SNP associated with intramuscular fat % in Brangus heifers (n = 761) genotyped with BovineSNP50 (Illumina, San Diego, CA)

No. SNP ID of Position ID of l Position Boot strap Proportion Previously Chr1 in window first SNP of first SNP ast SNP of last SNP P-value of variance identified QTL2 3 9 ss61475074 11985520 ss86334835 12634100 0.001 0.0020 No 5 6 ss86315413 49850261 ss86307440 50098067 0.001 0.0022 No 5 7 ss86307440 50098067 ss61569153 50562257 0.001 0.0025 No 5 7 ss86305510 50816331 ss61490671 51821819 0.010 0.0167 No 5 8 ss86289207 51145166 ss86314283 51969547 0.001 0.0131 No 5 6 ss61490674 51847011 ss86281166 52214923 0.001 0.0027 No 5 6 ss86339182 52006520 ss61555763 52454561 0.001 0.0028 No 5 7 rs29021741 53391515 ss86340726 53815448 0.001 0.0031 No 5 6 ss86310574 55037007 ss61490688 55176317 0.001 0.0016 No 7 7 ss61570619 18596377 ss86341484 18796165 0.001 0.0036 No 7 5 ss86295337 18833894 ss117969084 18953163 0.001 0.0024 No 7 5 ss86276220 18861656 ss86319066 18975319 0.010 0.0015 No 19 9 ss86290659 28449593 rs41905751 28736617 0.001 0.0015 No 19 8 rs41905353 28671068 ss61569153 28916863 0.001 0.0022 No 22 7 ss86283899 8328250 rs42857922 8558028 0.001 0.0013 No 27 9 ss86296987 35385533 ss61550098 35745332 0.001 0.0025 No 1Chr = chromosome number. 2CattleQTLdb (2011).

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 3406 Peters et al. the thyroglobulin gene, which contained a SNP that was non-SMC condensing I complex subunit G (NCAPG) one of the initial markers commercially validated for and osteopontin. Gene ontology of NACPG is mitotic beef cattle (Van Eenennaam et al., 2007). chromosome condensation and it has been reported The study of McClure et al. (2010) used 2 types of to be involved in fetal growth, stillbirth, and carcass analyses, linkage analyses using half-sib least squares weight in cattle (Kuhn et al., 2003; Eberlein et al., 2009; regression and Bayesian Markov chain Monte Carlo Setoguchi et al., 2009). Osteopontin was reported to have analyses. The authors stated that the 2 analyses differed association with growth and carcass traits in beef cattle, in their ability to detect QTL with low allele frequencies; as well as cattle selected for twinning (Allan et al., 2007; thus, the approach allowed identification of QTL that may White et al., 2007). The polygenic region of chromosome have been missed by 1 approach. Comparison of least 6 described in these reports was broad (i.e., ≥ 20 Mb), squares-REML and Bayes methodologies were discussed which parallels the region reported in this current study. in the reports of Wakefield (2008), Garrick (2009), and Chromosomes 14 and 26 were expected to harbor QTL Cantor et al. (2010). In our study, the decision was made as per previous reports (McClure et al., 2010; Snelling et to conduct the genome-wide association with Bayes C, al., 2010; Bolormaa et al., 2011a,b,c). The diacylglycerol which involved a mixture model within the framework of O-acyltransferase gene and thyroglobulin genes map to ridge regression (Kizilkaya et al., 2010; Fan et al., 2011; chromosome 14 and have been investigated in several Onteru et al., 2011). In brief, SNP effects were treated candidate gene-genotype to phenotype association studies as random and estimated simultaneously. These analyses (Marques et al., 2009; Arias et al., 2009; Pannier et al., did not use pedigree and assumed a common variance 2010). However, the SNP window(s) in the current study for each SNP. This approach offers advantages as results were distal to these loci. tend to be more biologically realistic than estimation of Karim et al. (2011) fine mapped a region of locus-specific variances influenced by SNP frequencies chromosome 14 associated with stature and various (Habier et al., 2011). The number of QTL detected in the measures of body fat in Holstein × Jersey crossed cattle. current study was less than the number described in the In brief, a region was studied that spanned chromosome reports of McClure et al. (2010) and Snelling et al. (2010); positions 24.7 to 25.6 Mb. Under this QTL, the intergenic however, this could partly be due to our conservative region downstream to the pleomorphic adenoma gene 1 interpretation of the results of bootstrap and hypothesis gene became the focus of this effort and results suggested testing procedure to identify the SNP sliding window. possible influences that both cis- and trans-acting Regions of chromosome 6 have been a focus in the elements influenced the region. That region contained discussions relating to several genome-wide association 9 genes and was in close proximity to the QTL detected studies in cattle. Specifically, Olsen et al. (2010) and in Brangus heifers for the traits of ADG(B-205) and rib Snelling et al. (2010) described this gene-dense region fat. To reiterate, additional research and fine mapping and its association with birth weight and other growth efforts are needed to delineate causative mutations traits. They discussed potential candidate genes such as containing DNA sequence polymorphisms underlying

Table 11. Identification of SNP associated with rib fat in Brangus heifers (n = 761) genotyped with BovineSNP50

No. SNP ID of Position ID of Position Boot strap Proportion Previously Chr1 in window first SNP of first SNP last SNP of last SNP P-value of variance identified QTL2 4 7 ss86293686 88599954 rs42187999 88808501 0.001 0.0029 Yes 5 5 ss86327441 50022924 ss86280507 50401370 0.001 0.0017 Yes 5 9 ss61490674 51847011 ss61555763 52454561 0.001 0.0027 Yes 5 9 ss117967882 55881324 ss61555768 56568072 0.001 0.0016 Yes 5 9 ss61485769 58820908 ss61545754 59482668 0.010 0.0100 Yes 5 9 ss61555892 71627713 ss86327552 72493060 0.010 0.0014 Yes 5 6 ss86283830 72449007 ss117967783 72677077 0.001 0.0045 Yes 5 8 ss86337303 81431996 ss86310331 81691901 0.001 0.0081 Yes 5 7 rs42812985 81913048 rs42855573 82164519 0.010 0.0117 Yes 14 7 rs41735007 28968386 ss61534725 29343444 0.001 0.0038 No 14 7 ss86340858 33680199 ss61534740 34055154 0.001 0.0017 Yes 14 10 ss61534740 34055154 ss105249048 34569564 0.010 0.0200 Yes 29 9 rs42161771 2086575 ss86288387 2371667 0.001 0.0046 Yes Un.0043 7 rs42443435 2078063 ss61501517 2157666 0.001 0.0043 No 1Chr = chromosome number. 2CattleQTLdb (2011) 3Un.004 = unmapped SP on BovineSNP50 (Illumina, San Diego, CA).

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Genome-wide association with heifer growth 3407

Allan, M. F., R. M. Thallman, R. A. Cushman, S. E. Echternkamp, these chromosome regions. Thus, chromosome 14 S. N. White, L. A. Kuehn, E. Casas, and T. P. L. Smith. 2007. continues to be of considerable interest in genome-wide Association of a single nucleotide polymorphism in SPP1 with association, as well as candidate gene studies. growth traits and twinning in a cattle population selected for Chromosome 5 has been reported to harbor QTL twinning rate. J. Anim. Sci. 85:341–347. associated with many traits of growth and fertility Arias, J. A., M. Keehan, P. Fisher, W. Coppieters, and R. Spelman. 2009. A high density linkage map of the bovine genome. BMC (Snelling et al., 2010, Bolormaa et al., 2011a,b,c; Genet. 10:18. Hawken et al., 2011). The Discussion in several of these Balding, D. J. 2006. A tutorial on statistical methods for population manuscripts described IGF1 as a positional candidate association studies. Nat. Rev. Genet. 7:781–791. gene under a QTL that could be at least 25 Mb wide. Beauchemin, V. R., M. G. Thomas, D. E. Franke, and G. A. Silver. There are several other genes known to be components of 2006. Evaluation of DNA polymorphisms involving growth hor- the GH-IGF axis that map to this region of chromosome mone relative to growth and carcass characteristics in Brahman steers. Genet. Mol. Res. 5:438–447. 5. Farber et al. (2006) described these genes within a Bolormaa, S., B. J. Hayes, R. J. Hawken, Y. Zhang, A. Reverter, and M. manually drawn GenMAPP image, which included the E. Goddard. 2011b. Detection of chromosome segments of zebu genes of IGFBP6, STAT2 and 6, and SOCS2. The STAT and taurin origin and their effect on beef production and growth. (i.e., signal transducer and activator of transcription) J. Anim Sci. 89:2050–2060. and SOCS (suppressor of cytokine signaling 2) proteins Bolormaa, S., B. J. Hayes, K. Savin, R. Hawken, W. Barendse, P. F. Arthur, R. M. Herd, and M. E. Goddard. 2011a. Genome-wide are involved in cell signaling cascades, subsequent to association studies for feedlot and growth traits in cattle. J. Anim. hormones, such as GH binding their receptors (Rawlings Sci. 89:1684–1697. et al., 2004; Schindler and Plumlee, 2008; Ahmed and Bolormaa, S., L. R. Porto Neto, Y. D. Zhang, R. J. Bunch, B. E. Farquharson, 2010). Polymorphisms within these STAT Harrison, M. E. Goddard, and W. Barendes. 2011c. A genome genes have been found to be associated with growth, wide association study of meat and carcass traits in Australian carcass, and fertility traits in studies of multibreed cattle. J. Anim. Sci. 89:2297–2309. Brangus Herd Improvement Records. 2010. Accessed Jan. 1, 2010. populations of cattle, which involved Angus × Brahman http://www.int-Brangus.org. crosses (Rincon et al., 2009; DeAtley et al., 2011; Luna- Cantor, R. M., K. Lange, and J. S. Sinsheimer. 2010. Prioritizing Nevarez et al., 2011). However, as stated previously, GWAS results: A review of statistical methods and recommenda- additional research and fine mapping efforts are needed tions for their applications. Am. J. Hum. Genet. 86:6–22. to delineate the causal mutation(s) underlying this broad Casas, E., S. D. Shackelford, J. W. Keele, M. Koohmaraie, T. P. L. Smith, and R. T. Stone. 2003. Detection of quantitative trait region of chromosome 5. loci for growth and carcass composition in cattle. J. Anim. Sci. The QTL detection methodology in this study was 81:2976–2983. based on single locus and multilocus LD between SNP CattleQTLdb. 2011. Release 14. Accessed June 6, 2011. http://www. and QTL. Garrick (2009) and He et al. (2010) discussed animalgenome.org/cgi-bin/QTLdb/BT/index. that crosses between distant breeds results in long-range dbSNP. 2011. dbSNP build 133 for cow txid 9913. Accessed June 6, LD in the F , which can be used to assist detection of 2011. http://www.ncbi.nlm.nih.gov/projects/SNP. 2 DeAtley, K. L., G. Rincon, C. R. Farber, J. F. Medrano, P. Luna- QTL via flanking markers. This strategy was successful Nevarez, R. M. Enns, D. M. VanLeeuwen, G. A. Silver, and in the studies of Casas et al. (2003) and Kim et al. (2003), M. G. Thomas. 2011. Genetic analyses involving microsatellite which involved Bos indicus × Bos taurus crossed cattle, ETH10 genotypes on bovine chromosome 5 and performance and more recently in the studies of Toosi et al. (2010), trait measures in Angus and Brahman-influenced cattle. J. Anim. Bolormaa et al. (2011a,b,c), Fan et al. (2011), and Onteru Sci. 89:2031–2041. Dodenhoff, J., L. D. Van Vleck, S. D. Kachman, and R. M. Koch. et al. (2011), involving crossbred livestock. Thus, it 1998. Parameter estimates for direct, maternal and grandmaternal appears this concept was applicable to our genome-wide genetic effects for birthweight and weaning weight in Hereford association study, using the advanced generations present cattle. J. Anim. Sci. 76:2521–2527. in this 2-breed composite of Brangus. In summary, results Eberlein, A., A. Takasuga, K. Setoguchi, R. Pfuhl, K. Flisikowski, R. from whole genome association of SNP with growth and Fries, N. Klopp, R. Furbass, R. Weikard, and C. Kuhn. 2009. ultrasound carcass traits in developing Brangus heifers Dissection of genetic factors modulating fetal growth in cattle in- dicates a substantial role of the Non-SMC condensing I complex confirmed 41 published QTL and detected 98 additional subunit G (NCAGP) gene. Genetics 183:951–964. QTL that can be added to CattleQTLdb (2011). These Fan, B., S. K. Onteru, Z.-Q. Du, D. J. Garrick, K. J. Stalder, and M. QTL spanned 25 bovine chromosomes. F. Rothschild. 2011. Genome-wide association study identifies loci for body composition and structural soundness traits in pigs. LITERATURE CITED PLoS One 6:e14726. Farber, C. R., P. M. Corva, and J. F. Medrano. 2006. Genome-wide Ahmed, S. F., and C. Farquharson. 2010. Review: the effect of GH and isolation of growth and obesity QTL using mouse speed congenic IGF1 on linear growth and skeletal development and their modu- strains. BMC Genomics 7:1–17. lation by SOCS proteins. J. Endocrinol. 206:249–259.

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 3408 Peters et al.

Fernando, B. L., and D. J. Garrick. 2009. User manual for a portfo- Hu, Z. L., E.R. Fritz, and J. M. Reecy. 2007. AnimalQTLdb: A live- lio of genomic selection related analyses. Animal Breeding and stock QTL database tool set for positional QTL information min- Genetics, Iowa State University, Ames. Accessed June 6, 2011. ing and beyond. Nucleic Acid Res. 35:D604–D609. http://taurus.ansci.iastate.edu/Site/Welcome.html. Imumorin, I. G., E.-H. Kim, Y.-M. Lee, D.-J. De Konig, J. A. van Fernando, R. L., D. Habier, C. Stricker, J. C. M. Deckers, and L. R. Arendonk, M. De Donato, J. F. Taylor, and J.-J. Kim. 2011. Totir. 2007. Genomic selection. Special issue: statistics in animal Genome scan for parent-of-origin QTL effects on bovine growth breeding and quantitative genetics: the next 60 years. Acta. Agric. and carcass traits. Front. Genet. 2:44. Scand., Sect. A – Animal Science. 57:192–195. Karim, L., H. Takeda, L. Lin, T. Druet, J. A. C. Arias, D. Burain, N. Fernando, R. L., D. Nettleton, B. R. Southey, J. C. M. Deckers, M. F. Cambisano, S. R. Davis, F. Farnir, B. Grisart, B. L. Harris, M. D. Rothschild, and M. Soller. 2004. Controlling the proportion of Keehan, M. D. Littlejohn, R. J. Spelman, and M. Georges, and W. false positives in multiple dependent tests. Genetics 166:611–619. Coppieters. 2011. Variants modulating the expression of a chro- Garrett, A. J., G. Rincon, J. F. Medrano, M. A. Elzo, G. A. Silver, and mosome domain encompassing PLAG1 influence bovine stature. M. G. Thomas. 2008. Promoter region of the bovine growth hor- Nat. Genet. 43:405–413. mone receptor gene: single nucleotide polymorphism discovery Kim, J. J., F. Farner, J. Savell, and J. F. Taylor. 2003. Detection of in cattle and association with performance in Brangus bulls. J. quantitative trait loci for growth and beef fatness traits in a cross Anim. Sci. 86:3315–3323. between Bos taurus (Angus) and Bos indicus (Brahman) Cattle. Garrick, D. 2009. The revolution in genetic improvement. Lohmann J. Anim. Sci. 81:1933–1942. Information 45:9–17. Kizilkaya, K., R. L. Fernando, and D. J. Garrick. 2010. Genomic pre- Georges, M. 2007. Mapping, fine mapping, and molecular dissec- diction of simulated multibreed and purebred performance using tion of quantitative trait loci in domestic animals. Annu. Rev. observed fifty thousand single nucleotide polymorphism geno- Genomics Hum. Genetic. 8:131–62. types. J. Anim. Sci. 88:544–551. Gibbs, R. A., J. F. Taylor, C. P. Van Tassell, W. Barendse, K. A. Eversol, Kuhn, C., J. Bennewitz, N. Reinsch, N. Xu, H. Thomsen, C. Looft, C. A. Gill, R. D. Green, D. L. Hamernik, S. M. Kappes, S. Lien, L. G. A. Brockmann, M. Schwerin, C. Weimann, S. Hiendleder, K. Matukumalli, J. C. McEwan, L. V. Nazareth, R. D. Schnabel, G. G. Erhardt, I. Medjugorac, M. Forster, B. Brenig, F. Reinhardt, M. Weinstock, D. A. Wheeler, P. Ajmone-Marsan, P. J. Boettcher, R. Reents, I. Russ, G. Averdunk, J. Blumel, and E. Kalm. 2003. A. R. Caetano, J. F. Garcia, O. Hanotte, P. Mariani, L. C. Skow, Quantitative trait loci mapping of functional traits in the German J. L. Williams, B. Diallo, L. Hailemariam, M. L. Martinez, C. A. Holstein cattle population. J. Dairy Sci. 86:360–368. Morris, L. O. C. Silva, R. J. Spellman, W. Mulatu, K. Y. Zhao, Lancaster, P. A., G. E. Carstens, D. H. Crews, T. H. Welsh, T. D. A. C. A. Abbey, M. Agaba, F. R., Araujo, R. J. Bunch, J. Burton, Forbes, D. W. Forrest, L. O. Tedeschi, R. D. Randel, and F. M. C. Gorni, H. Olivier, B. E. Harrison, B. Luff, M. A. Machado, J. Rouquette. 2009. Phenotypic and genetic relationships of resid- Mwakaya, G. Plastow, W. Sim, T. Smith, T. S. Stonstegard, M. B. ual feed intake with performance and ultrasound carcass traits in Thomas, A. Valentini, P. Williams, J. Womack, J. A. Wooliams, Brangus heifers. J. Anim. Sci. 87:3887–3896. Y. Liu, X. Qin, K. C. Worley, C. Gao, H. Y. Jiang. S. S. Moore, Luna-Nevarez, P., D. W. Bailey, C. C. Bailey, D. M. VanLeeuwen, R. Y. R. Ren, X. Z. Song, C. D. Bustamante, R. D. Hernandez, D. M. Enns, G. A. Silver, K. L. DeAtley, and M. G. Thomas. 2010. M. Muzny, S. Patil, A. S. Lucas, Q. Fu, M. P. Kent, R. Vega, A. Growth characteristics, reproductive performance, and evalu- Matukumalli, S. McWilliam, G. Sclep, K. Bryc, J. Choi, H. Gao, ation of their associative relationships in Brangus cattle man- J. J. Grefenstette, B. Murdoch, A. Stella, R. Villa-Angulo, M. aged in a Chihuahuan Desert production system. J. Anim. Sci. Wright, J. Aerts, O. Jann, R. Negrini, M. E. Goddard, B. J. Hayes, 88:1891–1904. D. G. Bradley, M. B. da Silva, L. P. L. Lau, G. E. Liu, D. J. Lynn, Luna-Nevarez, P., G. Rincon, J. F. Medrano, D. G. Riley, C. C. Chase, S. F. Panzitta, and K. G. Dodds. 2009. Genome-wide survey of SNP W. Coleman, D. M. VanLeeuwen, K. L. DeAtley, A. Ilsa-Trejo, G. variation uncovers the genetic structure of cattle breeds. Science A. Silver, and M. G. Thomas. 2011. Single nucleotide polymor- 324:528–532. phisms in the growth hormone–insulin-like growth factor axis in Goddard, M. E., and B. J. Hayes. 2009. Mapping genes for complex straightbred and crossbred Angus, Brahman, and Romosinuano traits in domestic animals and their use in breeding programmes. heifers: population genetic analyses and association of genotypes Nat. Rev. Genet. 10:381–391. with reproductive phenotypes. J. Anim. Sci. 89:926–934. Habier, D., R. L. Fernando, K. Kizilkaya, and D. J. Garrick. 2011. Mai, M. D., G. Sahana, F. B. Christiansen, and B. Guldbrandtsen. 2010. Extension of the Bayesian alphabet for genomic selection. BMC A genome-wide association study for milk production traits in Bioinformatics 12:186. Danish Jersey cattle using a 50K single nucleotide polymorphism Hawken, R. J., Y. D. Zhang, M. R. S. Fortes, E. Collis, W. C. Barris, chip. J. Anim. Sci. 88:3522–3528. N. J. Corbet, P. J. Williams, G. Fordyce, R. G. Holroyd, J. R. W. Marques, E., J. D. Nkrumah, E. L. Sherman, and S. S. Moore. 2009. Walkey, W. Barendse, D. J. Johnston, K. C. Prayaga, B. Tier, A. Polymorphism in positional candidate genes on BTA 14 and 26 Reverter, and S. A. Lehnert. 2012. Genome-wide association affect carcass quality in beef cattle. J. Anim. Sci. 87:2475–2484. studies of female reproduction in tropically adapted beef cattle. J. Matukumalli, L. K., C. T. Lawley, R. D. Schnabel, J. F. Taylor, M. F. Anim. Sci. 90:1398–1410. Allan, M. P. Heaton, J. O’Connell, S. S. Moore, T. P. L. Smith, Hayes, B. J., P. J. Bowman, A. J. Chamberlain, and M. E. Goddard. T. S. Sonstegard, and C. P. Van Tassell. 2009. Development and 2009. Invited review: genomic selection in dairy cattle: progress characterization of high density SNP genotyping assay for cattle. and challenges. J. Dairy Sci. 92:433–443. PLoS One 4:e5350. Hayes, B., and M. Goddard. 2010. Genome-wide association and ge- McClure, M. C., N. S. Morsci, R. D. Schnabel, J. W. Kim, P. Yao, M. nomic selection in animal breeding. Genome 53:876–883. M. Rolf, S. D. McKay, S. J. Gregg, R. H. Chapple, S. L. Northcutt, He, W., R. L. Fernando, J. C. M. Deckers, and H. Gilbert. 2010. A and J. F. Taylor. 2010. A genome scan for quantitative trait loci gene frequency model for QTL mapping using Bayesian infer- influencing carcass, post-natal growth and reproductive traits in ence. Genet. Sel. Evol. 42:21. commercial Angus cattle. Anim. Genet. 41:597–607.

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Genome-wide association with heifer growth 3409

Moser, D. W., J. K. Bertrand, I. Misztal, L. A. Kriese, and L. L. Shirley, K. L., M. G. Thomas, D. H. Keisler, D. M. Hallford, D. M. Benyshek. 1998. Genetic parameter estimates for carcass and Montrose, G. A. Silver, and M. D. Garcia. 2006. Case study: A yearling ultrasound measurements in Brangus cattle. J. Anim. Sci. Chihuahuan desert Brangus breeding program: Feed efficiency, 76:2542–2548. metabolic hormones, and puberty in heifers sired by bulls with Ogorevc, J., T. Kunej, A. Razpet, and P. Dovc. 2009. Database of cat- differing expected progeny differences for growth and scrotal cir- tle candidate genes and genetic markers for milk production and cumference. Prof. Anim. Sci. 22:48–58. mastitis. Anim. Genet. 40:832–851. Snelling, W. M., M. F. Allan, J. W. Keele, L. A. Kuehn, T. McDaneld, T. Olsen, H. G., B. J. Hayes, M. P. Kent, T. Nome, M. Svendsen, and P. L. Smith, T. S. Sonstegard, R. M. Thallman, and G. L. Bennett. S. Lien. 2010. A genome wide association study for QTL affect- 2010. Genome-wide association study of growth in crossbred ing direct and maternal effects of stillbirth and dystocia in cattle. beef cattle. J. Anim. Sci. 88:837–848. Anim. Genet. 41:273–280. Stelzleni, A. M., T. L. Perkins, A. H. Brown, F. W. Pohlman, Z. B. Onteru, S. K., B. Fan, M. T. Nikkila, D. J. Garrick, K. J. Stalder and Johnson, and B. A. Sandelin. 2002. Genetic parameter estimates M. F. Rothschild. 2011. Whole-genome association analyses for of yearling live animal ultrasonic measurements in Brangus cat- lifetime reproductive traits in pig. J. Anim. Sci. 89:988–995. tle. J. Anim. Sci. 80:3150–3153. Pannier, L., A. M. Mullen, R. M. Hamill, P. C. Stapleton, and T. Stromberg, U. L. F. 2009. Emperical Bayes and semi-Bayes ad- Sweeney. 2010. Association analysis of single nucleotide poly- justments for a vast number of estimations. Eur. J. Epidemol. morphisms in DGAT1, TG, and FABP4 genes and intramuscular 24:737–741. fat in crossbred Bos taurus cattle. Meat Sci. 85:515–518. Thomas, M. G., R. M. Enns, K. L. Shirley, M. D. Garcia, A. J.Garrett, Rawlings, J. S., K. M. Rosler, and D. A. Harrison. 2004. The JAK/ and G. A.Silver. 2007. Association of DNA polymorphisms in STAT signaling pathway. J. Cell Sci. 117:1281–1283. growth hormone and its transcriptional regulators with growth Rincon, G., E. A. Farber, C. R. Farber, J. D. Nkrumah, and J. F. and carcass traits in two populations of Brangus bulls. Genet. Medrano. 2009. Polymorphisms in the STAT6 gene and their Mol. Res. 6:222–237. association with carcass traits in feedlot cattle. Anim. Genet. Toosi, A., R. L. Fernando, and J. C. M. Deckers. 2010. Genomic selec- 40:878–882. tion in admixed and crossbred populations. J. Anim. Sci. 88:32–46. Rosenberg, N. A., L. Huang, E. M. Jewett, Z. A. Szpiech, I. Jankovic, Van Eenennaam, A. L., J. Li, R. M. Thallman, R. L. Quaas, M. E. and M. Boehnke. 2011. Genome-wide association studies in di- Dikeman, C. A. Gill, D. E. Franke, and M. G. Thomas. 2007. verse populations. Nat. Rev. Genet. 11:356–366. Validation of commercial DNA tests for quantitative beef traits. J. Schindler, C., and C. Plumlee. 2008. Interferons pen the JAK-STAT Anim. Sci. 85:891–900. pathway. Semin. Cell Dev. Biol. 19:311–318. Wakefield, J. 2008. Bayes factors for genome-wide association studies: Setoguchi, K., M. Furuta, T. Hirano, T. Nagao, T. Watanabe, Y. comparison with P-values. Genet. Epidemol. 33:79–86. Sugimoto, and A. Takasuga. 2009. Cross-breed comparison White, S. N., E. Casas, M. F. Allan, J. W. Keele, W. M. Snelling, T. L. identified a critical 591-kb region for bovine carcass weight QTL Wheeler, S. D. Shackelford, M. Koohmaraie, and T. P. L. Smith. (CW-2) on chromosome 6 and Ile-442 substitution in NCAPG as 2007. Evaluation in beef cattle of six deoxyribonucleic acid mark- a positional candidate. BMC Genet. 10:43. ers developed for dairy traits reveals an osteopontin polymorphism associated with postweaning growth. J. Anim. Sci. 85:1–10.

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013 Supplementary Material Supplementary material can be found at: http://www.journalofanimalscience.org/content/suppl/2012/10/23/90.10.33 98.DC1.html References This article cites 59 articles, 27 of which you can access for free at: http://www.journalofanimalscience.org/content/90/10/3398#BIBL Citations This article has been cited by 2 HighWire-hosted articles: http://www.journalofanimalscience.org/content/90/10/3398#otherarticles

Downloaded from www.journalofanimalscience.org at Iowa State University on April 17, 2013