Mol Biol Rep DOI 10.1007/s11033-010-0496-1

Analyses of porcine public SNPs in coding- regions by re-sequencing and phenotypic association studies

Xiaoping Li • Sang-Wook Kim • Kyoung-Tag Do • You-Kyoung Ha • Yun-Mi Lee • Suk-Hee Yoon • Hee-Bal Kim • Jong-Joo Kim • Bong-Hwan Choi • Kwan-Suk Kim

Received: 22 April 2010 / Accepted: 11 November 2010 Ó Springer Science+Business Media B.V. 2010

Abstract The Porcine SNP database has a huge number SNPs (16.2%) were found across all the five breeds, and of SNPs, but these SNPs are mostly found by computer 199 SNPs (21.7%) were breed specific polymorphisms. data-mining procedures and have not been well character- According to the SNP locations in the gene sequences, ized. We re-sequenced 1,439 porcine public SNPs from these 916 variations were categorized into 802 non-coding four commercial pig breeds and one Korean domestic SNPs (785 in intron, 17 in 30-UTR) and 114 coding SNPs breed (Korean Native pig, KNP) by using two DNA pools (86 synonymous SNPs, 28 non-synonymous SNPs). The from eight unrelated animals in each breed. These SNPs nucleotide substitution analyses for these SNPs revealed were from 419 -coding covering the 18 that 70.2% were from transitions, 20.0% from transver- autosomes, and the re-sequencing in breeds confirmed 690 sions, and the remaining 5.79% were deletions or inser- public SNPs (47.9%) and 226 novel mutations (173 SNPs tions. Subsequently, we genotyped 261 SNPs from 180 and 53 insertions/deletions). Thus, totally, 916 variations genes in an experimental KNP 9 Landrace F2 cross by the were found from our study. Of the 916 variations, 148 Sequenom MassARRAY system. A total of 33 traits including growth, carcass composition and meat quality were analyzed for the phenotypic association tests using the 132 SNPs in 108 genes with minor allele frequency Electronic supplementary material The online version of this (MAF) [0.2. The association results showed that five article (doi:10.1007/s11033-010-0496-1) contains supplementary marker-trait combinations were significant at the 5% material, which is available to authorized users. experiment-wise level (ADCK4 for rear leg, MYH3 for rear X. Li S.-W. Kim K.-T. Do Y.-K. Ha K.-S. Kim (&) leg, Hunter B, Loin weight and Shearforce) and four at the Department of Animal Science, Chungbuk National University, 10% experiment-wise level (DHX38 for average daily gain Cheongju, Chungbuk 361-763, South Korea at live weight, LGALS9 for crude lipid, NGEF for front leg e-mail: [email protected] and LIFR for pH at 24 h). In addition, 49 SNPs in 44 genes X. Li showing significant association with the traits were detec- Department of Animal Technology, Huazhong Agricultural ted at the 1% comparison-wise level. A large number of University, Wuhan 430070, China genes that function as enzymes, transcription factors or Y.-M. Lee J.-J. Kim (&) signalling molecules were considered as genetic markers School of Biotechnology, Yeungnam University, Gyeongsan, for pig growth (RNF103, TSPAN31, DHX38, ABCF1, South Korea ABCC10, SCD5, KIAA0999 and FKBP10), muscling e-mail: [email protected] (HSPA5, PTPRM, NUP88, ADCK4, PLOD1, DLX1 and S.-H. Yoon H.-B. Kim GRM8), fatness (PTGIS, IDH3B, RYR2 and NOL4) and Department of Food and Animal Biotechnology, Seoul National meat quality traits (DUSP4, LIFR, NGEF, EWSR1, ACTN2, University, Seoul, South Korea PLXND1, DLX3, LGALS9, ENO3, EPRS, TRIM29, EHMT2, RBM42, SESN2 and RAB4B). The SNPs or genes B.-H. Choi (&) National Institute of Animal Science, Suwon, South Korea reported here may be beneficial to future marker assisted e-mail: [email protected] selection breeding in pigs. 123 Mol Biol Rep

Keywords Pig public SNPs Resequencing Protein- 1,439 public SNPs coding genes Association study PCR and sequencing in five breeds

339 homozygous+ 690 SNPs+ 226 novel + 402 failed to be detected

Introduction 863 SNPs + 53 ins/del

Single nucleotide polymorphisms (SNPs) are one of the 261 SNPs most abundant genetic variations and are widespread Genotyping throughout the whole genome. In the , on 23 failed +33 homozygous + 68 MAF<0.2 + 132 MAF>0.2 average one SNP with a minor allele frequency (MAF) Association analyses greater than 1% occurs out of 300 base pairs [1, 2]. SNPs have also been found to be highly variable even in the coding 94 associated SNPs (p<0.01) regions of genes [3–7]. Compared with other polymor- phisms such as simple sequence repeats, SNPs have the 25: meat quality advantage of relatively high stability. Thus, due to their high 17: growth availability and stability, SNPs are becoming an important 23: carcass marker of choice for applications in a variety of fields such composition as population genomics, evolutionary analysis and disease Fig. 1 Steps and processes of SNP selection and validations. ins research [8–11]. In farm animals, SNPs can be used to insertion, del deletion, MAF minor allele frequency. The associated identify genome regions or genes influencing important SNPs were detected on a level of P \ 0.01 economic traits, such as growth, fatness, muscling, meat quality, and reproduction as well as disease resistance [12, 13]. So far, several causative mutations affecting these traits we annotated these sequences into coding or non-coding have been identified in pigs, included the earliest reported regions and re-sequenced them in five different breeds mutations within the meat quality genes (HAL, RN)[14, 15], using direct PCR sequencing methods and also performed a missense mutation (Asp298Asn) in MC4R affecting feed large scale phenotypic association analyses to identify intake, growth and backfat [16, 17], and the IGF2-intron3- useful DNA markers for pig growth, carcass and meat G3072A substitution affecting muscle growth and backfat quality traits. The flow chart of the whole experiment has thickness detected in a cross between Meishan and European been illustrated in the Fig. 1. White breeds [18, 19]. The pig NCBI SNP database (dbSNP) includes huge numbers of pig SNPs found by direct sequencing, in silico data mining or experimental studies [6, 7]. These publicly Materials and methods available SNPs are a valuable resource for gene linkage mapping and association studies in pigs. However, insuffi- Prediction of coding and non-coding SNPs cient annotations of the SNPs such as gene origins, genome locations and allele distributions in breeds dramatically The putative SNPs in 449 sequence tagged site (STS) limited their practical utilities. Therefore, characterizing representing 419 pig protein-coding genes were analyzed and validating these SNPs in breeds will benefit the further according to their genome locations (STS sequences are applications of these SNPs in pig genetics or breeding. In available in the NCBI STS database). These STSs are addition, the increasing availability of high throughput genomic sequences which may include intronic SNPs genotyping technologies makes it possible to conduct a (iSNP) and exonic SNPs, and the SNPs in exons can be large scale phenotypic association analyses and identify identified as coding regions (cSNPs) or non-coding regions multiple causal variants for complicated traits at one time. (UTR). A subset of cSNPs give rise to a variation in the Previously, we constructed an in silico coding gene SNP encoded amino acid residues are known as non-synony- map where we consolidated 465 SNP containing sequences mous SNPs (nsSNP), and the cSNPs that do not change from the NCBI pig SNP database (dbSNP) and assigned amino acid residues are called synonymous SNPs (sSNPs). them onto the pig QTL map based on BLAST analyses In this study, we used iSNP, sSNP, nsSNP and UTR to [20]. Totally, 449 sequences corresponding to 419 protein define the SNP locations and their characteristics. Because coding genes were submitted to the NCBI STS database the NCBI database has a huge number of porcine ESTs, the and each received an access ID. These sequences contained STS sequences were firstly manually blasted with the 1,439 putative SNPs covering 18 autosomes. In this study, porcine ESTs (identity score[90%) to find out the intronic 123 Mol Biol Rep regions of the STS sequences using the ‘‘nucleotide blast’’ Association analyses program ‘‘nr’’ option under the ‘‘basic blast’’ category in NCBI. If the porcine ESTs were unavailable, the STS Resource population and phenotype collections sequences were directly blasted against the mRNA of the human homologies, and the intronic regions of the STSs A three-generation resource population developed from a were identified due to the highly conserved intron–exon cross between five Korean native boars and ten Landrace boundaries between humans and pigs [21]. Then blastx was sows was used for the association study. A total of 404 F2 performed using the STS sequences with the detected animals with phenotypic records were genotyped for 261 porcine ESTs against the protein sequences of the human SNPs. The phenotypic traits we analyzed included four homologies to find out the non-coding and coding SNPs. growth traits (birth weight, 21-days weight, average daily The sSNPs and nsSNPs were evaluated by calculating the gain at weaning, average daily gain on test), 13 carcass frame of the consensus sequence translation. The newly composition traits (loin eye area, kalbi area, sirloin weight, detected mutations were analyzed in the same way. galmegi weight, backfat thickness, live weight, hot carcass weight, bone weight, loin weight, front leg weight, rear leg PCR amplification, sequencing and SNP discovery weight, leather weight, samgyup weight) and 16 meat quality traits (crude ash, crude protein, crude lipid, mois- In order to check the polymorphism status of the publicly ture, lipid, drip loss, water holding capacity, cooking loss, available SNPs in different pig breeds, we designed 449 shear force, loin PH at 24 h, CIE-L, CIE-A, CIE-B, Hunter sets of polymerase chain reaction (PCR) primers using L, Hunter A, Hunter B); the means and standard deviations Oligo 6 (see Table 1 and additional file 2) based on the for the 33 traits were listed in additional file 5. The detailed STS sequences. The primers were located at the two ends information about the population structure, animal man- of the STS to include as many of the putative SNPs as agement and the phenotypic measures has been described possible. The PCR product sizes averaged 137 bp in length by Choy et al. [22, 23]. and covered a total of 61,497 bp of 419 pig coding gene sequences. PCR assays were processed using two genomic SNP Genotyping DNA pools of unrelated animals from Berkshire, Landrace, Duroc, Yorkshire and Korean Native Pig (KNP) breeds, A total of 261 SNPs in 180 protein-coding genes were respectively. For each breed, equal amounts of DNA genotyped by Sequenom MassARRAY system in 450 F2 samples from four animals were mixed together to form pigs generated from the KNP and Landrace cross. The one DNA pool with a final concentration of 25 ng/ll. forward and reverse amplification primers as well as an Polymerase chain reactions were performed in a final extension primer were designed using SpectroDESIGNER volume of 20 ll using 25 ng genomic DNA, 1 9 PCR 1.3.4 (Sequenom), and the SNPs were multiplexed into 14 buffer (1.5 mM MgCl2), 4 mM of each dNTP, 10 pmol of assays using the iPLEX system (Sequenom) according to each primer and 1 U hTaq DNA polymerase (SolGent, the manufacturer’s instructions. The full list of multiplexed South Korea). The following program was used: 95°C for genotyped SNPs and primers used is available in additional 15 min, followed by 45 cycles of 95°C for 30 s, 30 s of file 2. annealing (from 50 to 68°C) and 72°C for 30 s, and a final extension step at 72°C for 5 min in a PTC200 Peltier Statistical analysis Thermal Cycler. To check the quality of amplification, 4 ll of PCR products were loaded to 2% agrose gel for analysis, Among the 261 SNPs, 138 SNPs that had at least ten and then 10 ll of PCR products identified as high quality animals for any of the two or three genotypes were chosen. were purified and directly sequenced commercially. Association tests between SNP genotypes and traits were Sequence polymorphisms were verified manually using performed with the GLM procedure of SAS (Version 9.1). Sequencher program (version 4.6). For all models, fixed effects were included for gender, F1

Table 1 Summary of SNPs in Putative Confirmed (%) Homozygous (%) Failed (%) Novel this study SNP number

Total SNPs 1,439 690 (47.9) 339 (23.5) 402 (27.9) 173 iSNPs 1,172 590 (50.3) 272 (23.2) 302 (25.8) 143 sSNPs 207 76 (36.7) 50 (24.2) 81 (39.1) 11 nsSNPs 44 16 (36.4) 14 (31.8) 14 (31.8) 11 UTR SNPs 16 8 (50.0) 3 (18.8) 5 (31.2) 8

123 Mol Biol Rep sire, and the tested SNP. Parity was also included as a fixed Korean Native pig (KNP) using two DNA pools from eight effect for pre-weaning traits: birth, weaning weight, and unrelated animals in each breed (primer sequences are average daily gain at weaning. Covariates were litter size available in additional file 2). After optimizing the PCR for the pre-weaning traits, age at slaughter for live weight conditions, 56 sets of primers still failed to amplify PCR and average daily gain on test, and age at slaughter and live amplicons due to the high GC content of the template weight for loin eye area (LEA), hot carcass weight sequences or very low annealing temperature of the prim- (HCWT), backfat thickness (Backth), and bone weight ers. Another eight PCR fragments showing good agarose (bone). For crude ash, crude protein, crude lipid, lipid, and gel quality yield were unreadable DNA chromatograms the rest of weight traits (Kalbi, front and rear leg, loin, with double peaks for the sequence analysis software. The sirloin, Galmagi, Samgyup, and leather), live weight was remaining 385 sets of working primers represented 344 fitted as a covariate. Age at slaughter was fitted as a genes and encompassed 1,215 putative public SNPs covariate for the rest of the meat quality traits. (additional file 3). To obtain significance thresholds for all combinations of By direct PCR sequencing, 690 out of the 1,215 SNPs the 33 traits and 138 SNPs, six traits were randomly cho- (56.8%) were confirmed to be polymorphic in at least one sen; birth weight, average daily gain on test, backfat breed and 339 SNPs (27.9%) were homozygous in all thickness, crude lipid, lipid weight and shear force. Indi- breeds. The remaining 178 SNPs (14.7%) close to the vidual genotypes for the tested SNP were randomly shuf- primer region or located in a fragment with insertion/ fled and assigned to F2 indivuduals, while keeping deletion mutations did not produce believable chromato- phenotypes and other fixed factors and covariates constant. grams and were removed from the results sheet. In par- After one permutation, the lowest P value among those ticular, 226 novel mutations including 173 SNPs and 53 from the tested 138 SNPs was chosen for each trait. The insertions/deletions were simultaneously identified. In permutations were repeated for 5,000 times for each trait, total, 916 DNA mutations were confirmed or newly and the P values at the 250th and 500th in small order were detected in our study, and they included 785 SNPs in chosen, and the average P values across the six traits were introns, 86 SNPs with no amino acid changes, 28 with determined to be the 5% and 10% experimental-wise amino acid changes and 17 in 30-UTR. The identified SNPs P values, respectively. The P value of 0.01 from F-distri- were also categorized according to nucleotide substitution butions was also used as the 1% comparison-wise signifi- as either transitions (A/G or C/T) or transversions (A/C, cance threshold. A/T, G/C, or G/T). Of the 916 DNA mutations, 70.2% were the result of transitions (34.5% A/G, 35.7% C/T) which was in line with human data that approximately 70% of Results SNPs were found to be the result of transitions [24, 25]. The remaining 29.8% mutations were composed of 24.0% Characteristics of the putative public SNPs

A total of 1,439 putative SNPs in 449 STS sequences from 300 419 pig protein-coding genes were analyzed according to 250 iSNP sSNP the SNP locations in the pig genome (STS sequences are 200 available in the NCBI STS database). In this study, we used iSNP, sSNP, nsSNP and UTR to define the SNP locations 150 nsSNP UTR and their characteristics. These 1,439 putative public SNPs 100 were categorized into 1,172 iSNPs (81.4%), 207 sSNPs (14.4%), 44 nsSNPs (3.06%) and 16 UTR SNPs (1.14%) by 50 blasting the pig STS sequences with pig ESTs and the number of observations 0 mRNA and protein sequences of the human homologies A/G C/T A/C A/T C/G T/G (Table 1). According to the nucleotide substitution types, iSNP 267 277 52 34 63 42 the 1,439 SNPs were also classified into 1,045 transitions 30 40 5 244 (72.6%) and 397 transversions (27.6%). The detailed sSNP information is available in additional file 1. nsSNP 1553310 UTR 455011 SNP validation in five pig breeds Substitution

PCR primers based on the 449 STS sequences were Fig. 2 Distribution of the iSNP, sSNP, nsSNP and UTR SNP assayed in Berkshire, Landrace, Duroc, Yorkshire and according to their nucleotide replacement 123 Mol Biol Rep transversions (7.10% A/C, 4.26% A/T, 7.53% C/G and illustrated in Fig. 2 and Table 1, and the detailed infor- 5.11% T/G) and 5.80% insertions/deletions. For the inser- mation is available in additional file 3. tions/deletions, 19 were multiple nucleotides mutations. Considering the distribution of polymorphic SNPs in The numbers of the various substitutions detected are different breeds, the polymorphic SNP numbers were cal- culated by occurrence in one (n = 199), two (n = 177), three (n = 151), four (n = 149) and all (n = 148) five of the breeds (Fig. 3). The 148 SNPs segregating in all breeds 450 400 consisted of 127 public and 21 novel identified SNPs and 5 350 SNPs predicted to result in an amino acid change encoded 300 at that position (additional file 4). Also, 199 SNPs were 250 200 found to exhibit polymorphisms within only one of the five 150 breeds (20, 29, 36, 42 and 72 in Berkshire, Landrace,

number of SNPs 100 Duroc, Yorkshire and KNP, respectively; additional file 5) 50 0 and the 145 SNPs were fixed in one breed, but polymorphic PCR or polymorphic polymorphic polymorphic polymorphic polymorphic sequence homozygous in the other four breeds (29, 47, 15, 24 and 30 in Berkshire, in 1 breed in 2 breed in 3 breed in 4 breed in 5 breed failed Landrace, Duroc, Yorkshire and KNP, respectively). The Series1 402 339 199 177 151 149 148 total number of identified SNPs in Berkshire, Landrace, Fig. 3 SNP validation results. The number of SNPs that failed, Duroc, Yorkshire and KNP were 440, 427, 479, 491 and homozygous, or polymorphic in breeds 527, respectively.

Table 2 SNPs displaying specific alleles in certain breeds SSC Gene name SNP site SNP type Seq. no.a Berk Duroc Land York KNP

1 GSN BV726581_264 iSNP 1 C T C C C 2YTCCY 2 MAPK9 BV726600_246 iSNP 1 C T C C C 2CYCCC 3 DCTN1 BV726656_287 iSNP 1 G G G G T 2KGGGK 3 DCTN1 BV726655_256 iSNP 1 C C C C T 2YCCCY 9 CAPN5 BV726833_333 iSNP 1 A A A G A 2ARRGA BV726833_340 iSNP 1 T T T C T 2TYYCT 12 LGALS9 BV726873_109 30-UTR 1 T C T T T 2TC[ TT T T BV726873_232 30-UTR 1 C A C C C 2CA[ CC M C BV726873_256 30-UTR 1 G A G G G 2GA[ GG G G BV726873_276 30-UTR 1 A G A A A 2AG[ AA A A 16 LIFR BV726873_496 nsSNP 1 C C C T C 2C[ TC C[ TC[ TC[ T 17 PPAPDC1B BV726992_49 iSNP 1 A G A A A 2ARAAA 18 H2AFV BV727017-256 iSNP 1 G A G G G 2GRGGG a The SNPs were sequenced in two PCR assays. 1: The SNPs were sequenced in two DNA pools containing eight animals for each breed. 2: In order to test the breed specific alleles detected in assay 1, another 16 animals in each breed were used to sequence these SNPs again

123 Mol Biol Rep

An interesting finding was that 13 SNPs in 9 genes traits, a total of 261 SNPs from 180 genes that were appeared to be fixed for specific alleles within either breeds polymorphic in either KNP or Landrace were selected (the including eight in Duroc, three in Yorkshire and two in genotyped SNPs are noted in additional file 3 and the KNP, respectively (Table 2). Among the eight specific phenotypic data were summarized in additional file 6). By alleles in Duroc, six alleles in the three genes were found using Sequenom MassARRAY system, these SNPs were when compared to the NCBI Porcine SNP data. Four genotyped in the 450 F2 animals generated from a cross alleles in the LGALS9 gene 30-UTR formed a new haplo- between five Korean native boars and ten Landrace sows type CAAG in Duroc, while another haplotype, TCGA was [22, 23]. The genotyping resulted in only 132 SNPs with formed in other breeds after sequencing DNA chromato- MAF [ 0.2, which were suitable for association analyses grams (see Table 2). The UTRs might be involved in post- (detailed information was listed in Table 3 and additional transcriptional regulation such as mRNA stability and file 7). The remaining 129 SNPs were removed due to PCR translation efficiency, and many of the identified cis-acting failure (n = 23), monomorphisms (n = 33), MAF \ 0.2 or elements for gene expressional and translational regulation limited number of genotyped animals (n = 73). These are located within the 30-UTR [25–27]. In order to confirm results showed that five marker-trait combinations were these breed specific alleles, they were re-sequenced in significant at a 5% experiment-wise level (ADCK4 for rear additional 16 genetically unrelated animals for each breed. leg, MYH3 for rear leg, Hunter B, loin weight and Shear- With the increased numbers of sequenced animals, all of force) and four at a 10% experiment-wise level (DHX38 for the breed specific alleles were found to be polymorphisms average daily gain at live weight, LGALS9 for crude lipid, (Table 2). For example, Duroc had both of the two hap- NGEF for front leg and LIFR for pH at 24 h). Furthermore, lotypes CAAG and TCGA in LGALS9 gene. However the 85 marker-trait combinations (49 SNPs in 44 genes) CAAG had much higher frequency than the other haplo- showing significant at the 1% comparison-wise level were type allele TCGA that only existed in the other four breeds, detected and put into the significant association results. which indicated CAAG may be a dominant allele in Duroc Among the totally 94 marker-trait combinations, 75 had pigs. mainly an additive effect, 40 a dominant effect, and 22 had both additive and dominant effects (P \ 0.05; Table 4, and additional file 8). Association analyses In total, 17 SNPs were associated with the four growth traits including birth weight (n = 3), 21-days weight In order to test whether any of the SNPs was associated (n = 6), average daily gain at weaning (ADGWT, n = 5) with the 33 growth, carcass composition and meat quality and average daily gain on test (ADGLT, n = 8) at the 1%

Table 3 SSC Genotyped SNPs Significantly associated SNPs (P \ 0.01) distribution of SNP number in SNPs (MAF [ 0.2) phenotypic association studies Growth Carcass composition Meat quality

1168 0 2 1 2178 1 0 0 32 1 10 0 4144 0 0 0 516102 1 2 638161 7 4 725122 0 1 85 3 20 0 98 4 10 1 10 7 4 1 0 2 12 36 20 3 4 7 13 20 11 0 2 1 14 14 9 1 1 2 15 15 8 1 2 2 16 9 4 0 1 1 17 14 6 1 2 1 18 5 4 0 1 0 Total 261 132 17 23 25

123 Mol Biol Rep

Table 4 SNPs that were associated with growth, carcass composition, and meat quality traits at least at the 1% comparison-wise level in the F2 animals SSC Gene name SNP_ID Traita P value LS mean (SD) Animal number 11 12 22

1 GBA2 BV726578_123 Cash 0.0008 1.07 (0.02) 1.02 (0.01) 1.01 (0.01) 376 GSN BV726581_290 Livewt 0.0099 85.69 (1.67) 93.67 (2.14) 88.71 (2.19) 328 HSPA5 BV726582_215 LEA 0.0028 25.82 (1.67) 31.39 (0.66) 29.74 (0.69) 220 2 AP3D1 BV726631_256 Birthwt 0.0081 1.17 (0.02) 1.25 (0.02) 1.26 (0.03) 353 3 RNF103 BV726663_256 ADGLT 0.0053 0.99 (0.10) 1.05 (0.08) 0.74 (0.08) 385 5 TSPAN31 BV726705_138 ADGLT 0.0036 1.36 (0.15) 0.91 (0.08) 0.82 (0.06) 377 WHC 0.0020 55.45 (1.02) 58.86 (0.56) 58.18 (0.38) 386 MIP BV726711_256 ADGWT 0.0012 0.17 (0.01) 0.19 (0.01) 278 Backth 0.0015 28.96 (0.65) 26.31 (0.81) 257 Cpro 0.0011 22.11 (0.11) 22.59 (0.14) 258 Lipid 0.0069 12.66 (0.35) 11.48 (0.43) 278 twentywt 0.0016 4.52 (0.12) 5.04 (0.12) 278 6 PLOD1 BV726731_312 Fleg 0.0060 7.31 (0.08) 7.75 (0.14) 7.82 (0.15) 342 Galmagi 0.0009 0.26 (0.01) 0.32 (0.01) 0.31 (0.02) 342 Samgyup 0.0072 9.42 (0.10) 9.56 (0.18) 10.05 (0.20) 342 SESN2 BV726737_510 Clipid 0.0082 2.83 (0.28) 2.15 (0.21) 1.77 (0.25) 382 Livewt 0.0034 83.18 (1.87) 89.57 (1.39) 89.36 (1.68) 368 Sirloin 0.0047 0.95 (0.02) 0.98 (0.02) 0.90 (0.02) 382 DHX38 BV726753_303 ADGLT 0.0003b 0.79 (0.08) 0.84 (0.09) 1.43 (0.15) 309 PTPRM BV726759_256 Sirloin 0.0030 0.88 (0.03) 0.98 (0.02) 0.87 (0.06) 225 NOL4 BV726762_263 Backth 0.0083 25.58 (0.84) 28.10 (0.71) 28.75 (1.21) 229 HPN BV726764_424 HCWTc 0.0024 73.50 (0.89) 70.33 (0.59) 71.64 (0.66) 393 Waterp 0.0073 74.61 (0.22) 74.06 (0.15) 73.73 (0.17) 393 RBM42 BV726766_258 Rlegc 0.0046 12.40 (0.17) 12.63 (0.12) 12.14 (0.14) 394 Samgyup 0.0028 9.72 (0.14) 9.60 (0.10) 9.23 (0.12) 394 Lipid 0.0019 13.11 (0.41) 11.67 (0.29) 11.52 (0.35) 394 ADCK4 BV726769_230 HCWTc 0.0029 69.17 (0.78) 71.53 (0.56) 72.38 (0.79) 349 Rlegc 0.0003d 11.92 (0.17) 12.59 (0.12) 12.61 (0.17) 349 Samgyup 0.0022 9.13 (0.13) 9.53 (0.10) 9.72 (0.14) 349 RAB4B BV726770_379 pH 24 h 0.0023 5.68 (0.02) 5.61 (0.02) 5.68 (0.03) 373 7 ABCF1 BV726776_256 ADGLT 0.0068 1.77 (0.25) 0.98 (0.07) 0.80 (0.07) 371 Birthwt 0.0056 1.26 (0.07) 1.27 (0.02) 1.21 (0.02) 370 EHMT2 BV726782_380 Lipid 0.0021 12.61 (0.32) 11.63 (0.36) 10.54 (0.58) 357 ABCC10 BV726787_431 ADGLT 0.0031 1.41 (0.15) 0.90 (0.08) 0.86 (0.11) 277 8 IGJ BV726812_306 Birthwt 0.0052 1.15 (0.05) 1.20 (0.02) 1.27 (0.02) 384 SCD5 BV726815_327 ADGLT 0.0094 0.79 (0.09) 1.02 (0.07) 0.79 (0.09) 381 9 KIAA0999 BV726837_119 ADGWT 0.0007 0.17 (0.01) 0.19 (0.01) 0.15 (0.01) 346 twentywt 0.0022 4.68 (0.10) 4.96 (0.12) 4.26 (0.21) 347 TRIM29 BV726842_344 Lipid 0.0055 12.78 (0.44) 11.33 (0.32) 12.25 (0.37) 346 10 EPRS BV726846_114 Clipid 0.0024 3.36 (0.40) 2.01 (0.23) 2.07 (0.25) 361 Waterp 0.0015 73.20 (0.26) 74.13 (0.15) 74.08 (0.17) 361 BV726846_50 ADGLT 0.0034 0.90 (0.13) 1.09 (0.08) 0.72 (0.09) 334 SYK BV726854_190 Cash 0.0035 0.95 (0.02) 1.03 (0.02) 1.04 (0.02) 377 12 P2RX1 BV726863_480 Livewt 0.0081 83.58 (2.05) 91.81 (1.73) 89.81 (2.27) 326 Loinwt 0.0039 3.38 (0.08) 3.72 (0.07) 3.65 (0.09) 337 CieA 0.0043 10.52 (0.35) 9.08 (0.30) 8.99 (0.39) 337

123 Mol Biol Rep

Table 4 continued SSC Gene name SNP_ID Traita P value LS mean (SD) Animal number 11 12 22

Huna 0.0056 8.72 (0.30) 7.48 (0.26) 7.47 (0.34) 337 BV726863_131 CieA 0.0014 8.79 (0.34) 9.17 (0.29) 10.54 (0.32) 380 CieB 0.0072 4.65 (0.27) 4.86 (0.23) 5.83 (0.26) 380 Huna 0.0021 7.27 (0.29) 7.57 (0.26) 8.75 (0.28) 380 ENO3 BV726865_302 Shear force 0.0053 3.84 (0.19) 3.18 (0.14) 3.41 (0.12) 293 NUP88 BV726867_167 LEA 0.0057 31.59 (1.08) 31.04 (0.66) 28.14 (0.81) 231 MYH3 BV726872_324 CieB 0.0011 5.96 (0.24) 5.04 (0.19) 5.10 (0.25) 352 Hunb 0.0013 4.52 (0.18) 3.84 (0.14) 3.85 (0.19) 352 Shear force 0.0010 3.07 (0.15) 3.62 (0.11) 3.63 (0.15) 352 Cooking loss 0.0044 73.48 (0.19) 74.12 (0.15) 74.11 (0.19) 352 LEA 0.0082 27.76 (1.00) 30.00 (0.73) 32.32 (1.09) 200 Loinwt 0.0019 3.42 (0.07) 3.57 (0.05) 3.75 (0.07) 352 Rleg 0.0001d 11.94 (0.17) 12.36 (0.13) 12.87 (0.17) 352 BV726872_412 Hunb 0.0044d 3.69 (0.15) 3.87 (0.14) 4.72 (0.22) 386 Driploss 0.0093 2.76 (0.19) 2.09 (0.19) 2.05 (0.28) 380 Shear force 0.0003d 3.74 (0.11) 3.37 (0.11) 2.97 (0.17) 386 Loinwt 0.0002d 3.74 (0.05) 3.52 (0.05) 3.41 (0.08) 386 LGALS9 BV726873_76 Clipidc 0.0004b 2.26 (0.24) 2.61 (0.24) 0.76 (0.48) 314 SUPT6H BV726876_248 twentywt 0.0098 4.58 (0.11) 4.95 (0.11) 231 FKBP10 BV726884_198 ADGWT 0.0027 0.17 (0.00) 0.19 (0.01) 0.18 (0.01) 350 twentywt 0.0055 4.59 (0.10) 5.06 (0.13) 4.81 (0.21) 351 BV726884_275 ADGWT 0.0038 0.17 (0.00) 0.19 (0.01) 0.19 (0.01) 361 twentywt 0.0066 4.57 (0.10) 5.01 (0.12) 4.88 (0.20) 362 DLX3 BV726894_111 Driploss 0.0031 2.30 (0.31) 2.24 (0.18) 3.59 (0.40) 310 13 PLCD1 BV726904_82 Fleg 0.0058 7.23 (0.15) 7.67 (0.10) 7.41 (0.08) 340 PLXND1 BV726907_140 Leather 0.0062 5.35 (0.15) 5.47 (0.12) 6.12 (0.16) 368 PLXND1 BV726906_239 Clipid 0.0012 3.89 (0.57) 1.59 (0.33) 0.95 (0.37) 365 Waterp 0.0074 73.25 (0.38) 74.19 (0.22) 74.69 (0.25) 365 14 ACTN2 BV726936_119 Cooking loss 0.0070 32.13 (0.55) 33.76 (0.33) 34.12 (0.57) 317 RYR2 BV726937_409 Backth 0.0042 28.55 (0.56) 27.19 (0.62) 24.73 (1.13) 376 twentywt 0.0032 4.52 (0.099) 4.85 (0.104) 5.13 (0.20) 344 EWSR1 BV726953_85 CieL 0.0099 49.85 (0.52) 50.97 (0.42) 52.09 (0.55) 389 15 DLX1 BV726965_236 Kalbi 0.0086 3.33 (0.06) 3.50 (0.06) 182 ATG9A BV726973_119 ADGLT 0.0018 1.51 (0.21) 0.87 (0.07) 288 Clipid 0.0071 3.92 (0.67) 2.14 (0.21) 288 Cpro 0.0008 21.26 (0.30) 22.26 (0.10) 288 NGEF BV726977_111 Fleg 0.0006b 7.77 (0.11) 7.35 (0.12) 199 CieL 0.0098 51.62 (0.61) 49.76 (0.62) 199 Cooking lossc 0.0079 32.97 (0.50) 34.53 (0.50) 217 HunL 0.0090 44.61 (0.60) 42.75 (0.61) 217 16 LIFR BV726981_354 Galmagi 0.0031 0.27 (0.01) 0.28 (0.01) 0.32 (0.01) 400 pH 24he 0.0006b 5.68 (0.02) 5.62 (0.02) 5.55 (0.03) 400 17 DUSP4 BV726991_329 Lipid 0.0026 12.00 (0.38) 11.31 (0.35) 13.14 (0.52) 339 Samgyup 0.0099 9.56 (0.13) 9.39 (0.12) 9.93 (0.17) 339 IDH3B BV726996_180 Backthe 0.0090 26.78 (0.70) 27.39 (0.64) 30.10 (0.94) 381 PTGIS BV727011_50 ADGWT 0.0037 0.17 (0.01) 0.19 (0.01) 0.16 (0.01) 357 Backth 0.0091 26.68 (0.65) 27.92 (0.59) 29.66 (0.86) 389

123 Mol Biol Rep

Table 4 continued SSC Gene name SNP_ID Traita P value LS mean (SD) Animal number 11 12 22

18 GRM8 BV727018_58 LEA 0.0087 32.76 (0.93) 29.73 (0.65) 29.07 (1.01) 215 a The full names of these traits are available in additional file 6 b Significant at the 10% experiment-wise level c The associated SNP is also located in the QTL region affecting that trait, and the QTL is detected in the same population for the association study d Significant at the 5% experiment-wise level e The associated SNP is located in the QTL region affecting that trait, and the QTL information is obtained from the pig QTL database ( http://www.animalgenome.org/QTLdb/) comparison-wise level (Additional file 8). The AP3D1, NGEF genes were significantly associated with meat color. RNF103, DHX38, ABCF1, ABCC10, IGJ, SCD5, The ENO3 and DLX3 genes on SSC12 were associated KIAA0999 and FKBP10 genes were associated only with only with Shearforce (P = 0.0053) and drip loss growth traits (P \ 0.01). Interestingly, the AP3D1 and IGJ (P = 0.0031), respectively. And the ACTN2 gene on genes were associated only with birth weight, and RNF103, SSC14 was significantly associated only with cooking loss DHX38, ABCC10 and SCD5 were associated only with (P = 0.007). ADGLT (P \ 0.01). The DHX38 polymorphism In our study, a total of 13 detected nsSNPs including 7 (BV726753_303) was significantly associated with SNPs in conserved domains (ID3, SYNPO2, SGSM2, ADGLT at 10% experiment-wise level and the two homo- SETMAR, FST, FLJ23356 and BMP2) were genotyped in zygote had extremely different values (AA = 0.793 kg/ the resource population and only six SNPs segregating with day, GG = 1.433 kg/day). a MAF [ 0.2 were performed for association analyses. The For the 13 carcass composition traits, 23 significant SNPs in ZNF532, PDE4DIP (BV726681_157), SETMAR SNPs were detected, and 13 of which were associated with and BMP2 genes were confirmed to be homozygous and loin eye area (LEA), loin weight (Loinwt) or sirloin weight SNPs in SYNPO2 and SGSM2 genes showed MAF \ 0.05. (sirloin) and backfat thickness (Backth) (p \ 0.01) (Addi- The SNPs in FLJ23356, SCG2 and PDE4DIP tional file 8). These genes (HSPA5, PTPRM, NUP88, (BV726681_196) gene displayed significant association ADCK4, PLOD1, DLX1 and GRM8) function as tran- with average daily gain on test (ADGLT), Hunter L and scription factors, enzymes, or signalling molecules in cer- multiple growth traits, respectively at the P \ 0.05 level, tain pathways that were listed as the pig candidate genes but not significant at the P \ 0.01 level. The SNPs in ID3, related to muscling. The SNP in the HSPA5 gene was NRIP1 and FST did not show any association with the traits significantly associated with LEA (P = 0.0028), and the tested in this study. two homozygote had extremely different LEA values (CC = 25.82 cm2,TT= 29.74 cm2). The SNP in the GRM8 gene also displayed a strong association with LEA Discussion (P = 0.0087), and the CC genotype has 3.69 cm2 higher than that in TT genotype (CC = 32.76 cm2, TT = The SNPs in pig SNP database were mainly found by in 29.07 cm2). Four genes (PTGIS, IDH3B, RYR2 and NOL4) silico methods, also called computer data-mining proce- were found to have functional implications in pig backfat dures, and therefore most of them have not been well- deposition. Three genes (HPN, ADCK4 and RBM42) characterized or validated. Kollers et al. [28] validated 109 clustered on pig [20] were associated with public SNPs within six commercial pig populations using a hot carcass weight, rear leg weight or Samgyup (belly fat) PCR–RFLP method. However, they directly used the NCBI weight. ‘‘ss*’’ sequences as a template which did not give any A total of 25 SNPs were significantly associated with the information about the SNP locations in the pig genome. 40 meat quality trait items (P \ 0.01) (Additional file 8). Herein, we developed a more detailed annotation of 1,439 The genes related with pH at 24 h post-mortem were public SNPs and offered useful information for the com- RAB4B and LIFR, while the MIP, RBM42, SESN2, mercial applications of these SNPs. EHMT2, TRIM29, EPRS, LGALS9, PLXND1 and DUSP4 Firstly, knowing the SNP location within the gene genes were considered to affect the lipid weight or crud structure is very useful for the functional evaluation of the lipid percentage in loin muscle. The P2RX1, EWSR1 and SNPs. In this study, we categorized the SNPs into

123 Mol Biol Rep noncoding SNPs (ncSNPs) or coding SNPs (cSNPs), and effect on phenotypic variations and a large number of annotated coding SNPs as nonsynonymous changes phenotypic associated iSNPs was found in our study. (nsSNPs) or synonymous changes (sSNPs). This kind of Therefore, all the SNPs, even the insertions/deletions, annotation offered basic information of the SNP charac- detected here may be used for DNA markers for the pig teristics for future studies. In the pig genome, nsSNPs are phenotypic variations, for which they need further possibly influencing phenotypes, since they change amino evaluation. acid then may alter the protein function [16, 17]. By direct Secondly, we have chosen four commercial breeds sequencing, we confirmed 15 public nsSNPs and simulta- (Berkshire, Landrace, Duroc and Yorkshire) and a Korean neously identified additional 12 new ones. A structure domestic breed (Korean Native pig, KNP) to validate the prediction on the detected nsSNPs revealed 13 SNPs status of the publicly available SNPs. The inclusion of the located in conserved regions across species for gene reg- four commercial breeds were due to their worldwide pop- ulators (transcription regulator, growth factor), binding ularity in the pig breeding systems and the pork industry, motifs (ATP binding, Ca2? binding, clathrin and carbo- and understanding the allele distribution of the SNPs in hydrate side chains binding) or interaction domains which these breeds may help to collect informative SNPs for the may function as biologically important roles in phenotypic association studies or linkage mapping in the reference variations or differences. The human disease-related stud- populations using these breeds. Here we randomly chose ies have been well developed and can provide suggestions four animals to form one DNA pool for the PCR assay and for the functional evaluation of these SNPs in pigs. Nine in each breed two DNA pools were used. This sampling SNPs have been identified from the corresponding human strategy made it possible to detect the SNPs with a minor genes, although they were not at the exactly same sites as allele frequency (MAF) [12.5% indicating the detected the porcine SNPs but they were located in the same func- SNPs segregated at a useful frequency in these populations. tional domains. For example, for the ID3 gene: one nsSNP The 27.9% publicly available SNPs was confirmed to be changing residue Leu to Met was at position 38 in the pig, homozygous in our study may be due to the following and one sSNP coding residue pro was found at position 37 reasons. Firstly, the MAF of the SNP was too low to be in the human, and these two SNPs are all located in the detected in the limited animals we used. Secondly, the HLH domain with transcription activity. These SNPs may polymorphisms were segregating in other breeds we did not play an important role in protein function. In order to include in this study. Finally, they were indeed not poly- further test their functions, a total of 13 nsSNPs including 7 morphisms. The 148 SNPs segregating in all the five breeds SNPs in conserved domains were genotyped in the resource would be extremely useful as markers in linkage or asso- population but only six SNPs were segregating with a ciation studies due to their wide polymorphism distribu- MAF [ 0.2. Surprisingly, no significant association for tions in breeds. these six SNPs were detected at the P \ 0.01 level, but The KNP breed was included in the present study to the several associations were significant at the P \ 0.05 level. following two reasons. Firstly, the SNPs collected from the The SNPs in ID3, NRIP1 and FST did not show any NCBI database deposited before 2005 did not contain association with the growth, carcass composition and meat the KNP breed as a template. Secondly, the Korean Native quality traits tested in this study, but may affect other traits Pig has interesting biological characteristics such as supe- we did not test such as reproduction, immune, disease rior meat qualities, fatty acid composition, and differen- resistance traits or they alternatively regulate the activity of tially expressed genes in muscle and adipose tissue other unknown genes by trans-acting. FST gene encodes a [33–35]. Understanding the genetic background based on single-chain gonadal protein that specifically inhibits fol- DNA level will benefit the utilization of this breed in pork licle-stimulating hormone release [29] and may affect the industry [36–38]. From the re-sequencing data, KNP pig reproduction ability. NRIP1 gene encodes a nuclear exhibited a higher number of total polymorphisms (527 protein (nuclear receptor interacting protein 1) which spe- SNPs) and breed specific SNPs (72) in comparison to other cifically interacts with the hormone-dependent activation breeds. Furthermore, 27 out of the 72 breed specific SNPs domain of nuclear receptors. Since this protein can mod- have not been previously reported. The polymorphisms ulate the transcriptional activity of the estrogen receptor detected here will increase the knowledge of the molecular [30] and this gene is thought to be essential for female basis of variations in the KNP breed and facilitate the fertility in mice [31], the NRIP1 gene was also predicted to marker assisted selection in this breed. Large scale or be possibly related with porcine reproduction traits. The genome-wide association (GWA) study is a favourable way ID3 gene contains a helix-loop-helix (HLH) domain typical to simultaneously identify multiple causal variants impli- of a regulator of transcription [32], and its function in pigs cated in quantitative traits [39]. In this study, we performed needs to be further investigated. In addition to the nsSNPs, phenotypic association analyses using the 261 detected the SNPs in noncoding regions may also have important SNPs covering pig 17 autosomes to identify genes affecting 123 Mol Biol Rep the growth, carcass composition and meat quality traits in a genes are members of the superfamily of ATP-binding cas- resource population developed from a cross between five sette (ABC) transporters [53, 54]. ABC transport a Korean native boars and ten Landrace sows. Totally, nine number of endogenous substrates, including inorganic marker-trait combinations were significant at the 10% anions, metal ions, peptides, amino acids, sugars and a large experiment-wise level, and 51 SNPs in 46 genes were number of hydrophobic compounds and metabolites across identified at the 1% comparison-wise level. the plasma and intracellular membranes [55, 56]. Energy The birth weight of piglets affects the pig’s postnatal metabolism is an important issue during animal growth, and survival and growth, and Garn [40] indicated that long- genes involved in this pathway usually affect growth. The term weight gains were positively and systematically SCD5 gene encoding a member of Stearoyl-CoA desaturase related to birth weight in humans. Average daily gain on (SCD) which is a key regulator of energy metabolism that test (ADGLT) is an index directly reflecting the long term catalyzes the formation of monounsaturated fatty acids from growth rate from birth to slaughter, and from our study, the saturated fatty acids [57] and may be playing a role in obesity ABCF1 gene was significantly associated with birth weight and dyslipidemia in humans [58, 59]. and ADGLT. In addition, the pigs with higher birth weight From the association study and function predictions, also had a higher ADGLT (Table 4). The SNPs in AP3D1 about ten gene polymorphisms were considered to affect and IGJ genes displayed unitary significant association pig carcass traits. The HSPA5, PLOD1, ADCK4, DLX1, with birth weight indicating these genes or chromosomal GRM8 and PTPRM genes strongly suggested their potential regions may functionally contribute to the pig prenatal effects on muscling. HSPA5 is a member of the heat-shock growth. The IGJ gene encodes immunoglobulin J chain protein-70 (HSP70) family, and involved in the folding and (IgJ) polypeptide and this peptide is necessary for the assembly of proteins in the endoplasmic reticulum (ER) polymerization of IgM and IgA, and their secretions in the [60]. PLOD1 encodes the collagen modifying enzyme lysyl plasma cells and mucosal membranes [41–43]. IGJ is hydroxylase, and has been significantly associated with important for the primary immune response [44] and bone mineral density (BMD) or bone related disorders upregulated in laboring human myometrium [45] as well as identified in humans [61, 62]. From our association anal- pregnant mice between gestational d13.5 and d16.5 [46]. It yses, this gene was significantly associated with the front likely plays a role in parturition involving immune system leg, Galmagi and Samgyup weights consistent with the activation near the end of pregnancy [43]. gene function in humans because bone traits which we did Pigs with higher ADGLT reflecting the long term higher not measure might be related with muscle mass. The DLX1 growth rate usually have higher live weight (Livewt) and gene encodes a homeobox transcription factor localized to better carcass performance. From the association studies, the nucleus where it functions as a transcriptional regulator seven polymorphisms in genes including RNF103, of signals from multiple TGF beta superfamily members. TSPAN31, DHX38, ABCF1, ABCC10, SCD and ATG9A In the embryonic and adult forebrain, DLX1 plays a crucial were significantly associated with ADGLT (P \ 0.01). role in the control of craniofacial patterning and the dif- Functional investigation in other species suggested these ferentiation and survival of inhibitory neurons [63, 64]. In genes were mostly involved in growth related to the cell humans, this gene is located in a tail-to-tail configuration cycle or cellular signalling [47–59]. The protein encoded with another member of the family, DLX2, on the long arm by the RNF103 gene contains a RING-H2 finger [47], a of chromosome 2. However, the pig DLX1 gene was motif known to be involved in protein–protein and protein– mapped to chromosome 15. Association results indicated DNA interactions. The TSPAN31 gene encodes a protein this gene was also significantly associated with Kalbi (short that is a member of the transmembrane 4 superfamily, also rib and muscle) weight (P = 0.0086). Interestingly, known as the tetraspanin family [48]. Most of these are another pair of members from this family, DLX3 and cell-surface proteins and mediate signal transduction DLX4, were mapped to pig chromosome 12 [20] were also events that play a role in the regulation of cell develop- located in a tail-to-tail configuration in the human chro- ment, activation, growth and motility [49, 50]. The protein mosome. DLX3 seemed to affect drip loss according to the encoded by the DHX38 gene is a member of the DEAD/H association analyses and DLX4 showed an association with box family of splicing factors function as an ATPase and meat quality traits at the P level of 0.05. The GRM8 gene essential for the catalytic step II in pre-mRNA splicing encodes a G protein-coupled metabotropic glutamate process [51, 52]. The DHX38 gene may be used as a receptor implicated in central nervous system for gluta- genetic marker for pig growth rate (ADGLT) selection, matergic neurotransmission [65, 66]. Eight distinct because the two homozygote of the BV726753_303 poly- metabotropic glutamate receptors have been divided into morphism had extremely different values (AA = 0.793 kg/ three groups and GRM8 is in group III, linked to the day, GG = 1.433 kg/day). ABCF1 and ABCC10 genes are inhibition of the cyclic AMP cascade [67]. Porcine GRM8 located on SSC7 [20] and the proteins encoded by these two gene is located on SSC18 [20] and the SNP in this gene 123 Mol Biol Rep displayed a strong association with LEA (P = 0.0087). gene with only two haplotypes found based on the pooling PTPRM SNP was only associated with sirloin weight sequencing DNA chromatograms in the five breeds. The C (P = 0.0030) and the protein encoded by the PTPRM gene allele was followed by the other six SNPs and formed is a member of the protein tyrosine phosphatase (PTP) hyplotype CCCGAAG and T allele formed hyplotype family [68]. The PTPs are known to be signaling molecules TTTACGA. And the underlined haplotype alleles above that regulate a variety of cellular processes including cell where CAAG was dominant in Duroc means the Duroc growth, differentiation, mitotic cycle, cell–cell aggregation breed had a dominant haplotype contributing to a higher and oncogenic transformation [69–74]. crud lipid content in loin muscle. Interestingly, it has been Backfat thickness is highly related with meat produc- shown that Duroc pigs had highest lipid content of loin tion, where pigs with moderate backfat thickness will muscle when compared with Berkshire, Landrace, York- benefit from an optimal meat quantity and quality. In this shire and some other pig breeds [80–83]. This may support study, a set of genes displaying a significant association the proposed functionality of the LGALS9 gene for pig with fatness trait from the association studies attracted our intramuscular fat (IMF) trait with the selection of haplo- attention. The NOL4 gene encodes an RNA binding type CCCGAAG to potentially increase the lipid content of nucleolar protein and the SNP in this gene only showed a loin muscle. significant association with backfat thickness Previous studies have pointed out the existence of sig- (AA = 25.58 mm, AG = 28.10 mm, GG = 28.75 mm) nificant correlations among numerous meat quality traits. (P = 0.0083). Two genes on SSC17 named PTGIS and For example, Huff-Lonergan et al. [84] indicated Loin pH IDH3B [20] and one gene RYR2 on SSC14 [20] were also at 24 h postmortem was highly correlated with pork color, significantly associated with backfat thickness (P \ 0.01). drip loss, tenderness, flavor, and cooking loss, while a In addition, PTGIS and RYR2 also displayed significant lower ultimate pH of the pork was associated with poorer associations with pig early growth (average daily gain at pork quality traits such as lighter color, higher drip loss, weaning and 21 days weight, respectively). The PTGIS less tender and less pork flavor. Loin lipid percentage was gene encodes prostacyclin (PGI2) synthase which catalyzes associated with most of the pork sensory characteristics. prostaglandin endoperoxide (PGH2) to form PGI2 [75], Pork drip loss was negatively correlated with tenderness and IDH3B encodes isocitrate dehydrogenase 3 (NAD?) (r2 = –0.30), flavor (r2 = –0.24), subjective color (r2 = beta involved in the tricarboxylic acid cycle [76]. The –0.33) and marbling (r2 = –0.12), and loin hunter L was RYR2 gene encodes a ryanodine receptor found in cardiac significantly correlated with pork firmness (r2 = –0.20), muscle sarcoplasmic reticulum. Ryanodine receptors drip loss (r2 = 0.33), cooking loss (r2 = 0.31), and ten- (RyRs) are responsible for the regulated release of Ca2? derness (r2 = –0.15) [84]. Some meat quality traits also from intracellular Ca2? stores [77] and mutation in RYR1 have strong correlations with the carcass measurements, has been proven to cause porcine stress syndrome and such as backfat thickness was positively related with produce pale, soft, exudative (PSE) meat [14]. However, marbling, loin lipid percentage, pork firmness, flavor and the functional implications of RYR2 in pigs have yet to be tenderness. However, Loin eye area was negatively related investigated. This association study indicated this gene with these traits [84]. Understanding the relationships may affect pig early growth and backfat deposition. And among overall phenotypic traits has two advantages: (1) further study need to be conducted on this gene. Indirect improvement of some difficult to measure traits via In addition to the growth and carcass traits, we identified improving those positively related traits which can be a number of genes related to meat quality. It should be easily measured, such as the improvement of meat flavor noted that two genes were significantly associated with and tenderness through increasing the backfat thickness. meat quality traits at 10% experiment-wise level. The first (2) To avoid the deterioration of meat quality resulting gene, LIFR, was significantly associated with 24 h post- from an excessively increasing growth rate and by finding a mortem pH (P = 0.0006). This gene encodes a protein that balance point between the growth and meat quality traits. belongs to the type I cytokine receptor family which Positional candidate gene study is another efficient mediates the action of the leukemia inhibitory factor, a method to identify functional important genes that impli- polyfunctional cytokine that is involved in cellular differ- cate in the pig economic traits. If one gene associated with entiation, proliferation and survival in the adult and the a certain trait is also located in the QTL affecting that trait, embryo [78, 79]. The second gene, LGALS9, was signifi- this gene may be a major effect gene or closely linked with cantly associated with crude lipid at 10% experiment-wise a causative mutation for the QTL. Kim et al. [85] has level (P = 0.0004) and the CC genotype (2.257%) had a detected the Mendelian and parent-of-origin QTL for much higher crud lipid content than the TT genotype growth and carcass composition traits in the same popu- (0.762%). Strong linkage disequilibrium (LD) was found lation we used for association study. A total of eight QTLs between this SNP and the other six SNPs detected in this were detected at 5% genome-wise level, including those 123 Mol Biol Rep affecting birth weight at 108 cM on SSC4, front leg weight a QTL for hot carcass weight and rear leg weight were at 47 cM on SSC12, two QTLs affecting bone weight at identified. Furthermore, our association studies confirmed 0 cM on SSC9 and 40 cM on SSC13, and another four their close relationships with these traits (P \ 0.01). QTLs on SSC6 where backfat thickness at 106 cM between ADCK4 displayed a very significant association with rear S0003 and S0228, hot carcass weight and Rear leg weight leg weight at a 5% experiment-wise level, but the function at 54 cM close to SW1067, and leather weight at 175 cM of this gene is unclear. (see Table 5). The human SESN2 and NOL4 genes (Gene Ten QTL affecting meat quality traits were also detected ID: 83667 and 8715, respectively) are located on chro- in this population at a 5% genome-wise level (see Table 5), mosome 1 spanning 28.48–30.06 Mb of genomic sequence. and the crude lipid percentage associated gene LGALS9 According to the previous in silico SNP map [20], these (P = 0.0004) as well as the cooking loss associated gene two genes were all mapped to pig chromosome 6, where NGEF (P = 0.0079) were found to be located in the SESN2 was close to SW1823 and the FABP3 (heart fatty SSC12 and SSC15 QTL regions, respectively. LGALS9 acid binding protein) gene, and NOL4 close to S0003. The was significantly associated with crude lipid at a 10% NOL4 gene was significantly associated with the backfat experiment-wise level and the two haplotypes had extre- thickness (P = 0.0083) and was located in the SSC6 QTL mely different crude lipid percentages which strongly region for this trait, which suggested the NOL4 gene may indicated this gene may be an important marker for this be a novel positional candidate gene marker for the QTL trait. The NGEF gene located in the cooking loss QTL affecting the backfat thickness trait in pigs. Another region was also significantly associated with front leg at interesting finding was the three closely arranged genes on 10% experiment-wise level and Hunter L as well as Cie L human chromosome 19: HPN, RBM42 and ADCK4 (Gene with P \ 0.01. Three genes (LIFR, IDH3B and MYH3) ID: 3249, 79171, 79934, respectively). They were in silico displaying significant association with pH, backfat thick- mapped to pig chromosome 6 close to SW193 [20], where ness or muscling traits (P \ 0.01) were found to be located

Table 5 Phenotypically associated genes (SNPs) in QTL regions SSC Trait Position (cM) SW interval Significantly associated genes (SNPs) (P \ 0.01)

4 Birth weight 108 S0214-SW445 5 Cooking loss 139 SW995 6 Backfat thickness 106 S0003-S0228 NOL4 (BV726762_263) 6 Hot carcass weight 54 SW1067 HPN (BV726764_424), RBM42 (BV726766_258), ADCK4 (BV726769_230) 6 Rear leg weight 54 SW1067 RBM42 (BV726766_258), ADCK4 (BV726769_230)a 6 Leather weight 175 SW2415 9 Bone weight 0 SW983 9 CieB 125 S0295-SW174 9 Hunter B 126 S0295-SW174 10 Drip loss 165 SW2067 11 pH at 24 h 38 S0182 11 Shear force 102 SW1377 12 Crude lipid 117 SWR1021 LGALS9 (BV726873_76)b 12 Front leg weight 47 S0083 12 Ham weight 108.3c SW605 MYH3 (BV726872_324)a 13 Bone weight 40 SW344 13 Crude ash 36 SW344 13 Crude protein 18 SW1378 15 Cooking loss 138 SWR1221 NGEF (BV726977_111) 16 pH at 45 min 31c SW419-S0077 LIFR (BV726981_354)b 17 Backfat thickness 35.2c S0296-SW1920 IDH3B (BV726996_180) a The gene is significantly associated with the trait at the 5% experiment-wise level b The gene is significantly associated with the trait at the 10% experiment-wise level c QTL was obtained from the pig dbQTL (http://www.animalgenome.org/QTLdb/). The QTL affecting growth and carcass composition is from Kim et al. [85], and the QTL affecting meat quality is from an unpublished report

123 Mol Biol Rep in related QTL regions detected in other populations (see 3. Cargill M, Altshuler D, Ireland J, Sklar P, Ardlie K, Patil N, Shaw Table 5). The LIFR gene significantly associated with 24 h N, Lane CR, Lim EP, Kalyanaraman N, Nemesh J, Ziaugra L, Friedland L, Rolfe A, Warrington J, Lipshutz R, Daley GQ, post-mortem pH at a 10% experiment-wise level was Lander ES (1999) Characterization of single-nucleotide poly- located in a QTL affecting pH at 45 min post-mortem on morphisms in coding regions of human genes. Nat Genet SSC16 [86, 87]. The IDH3B gene significantly associated 22(3):231–238 with backfat thickness was located in the corresponding 4. Lindblad-Toh K, Winchester E, Daly MJ, Wang DG, Hirschhorn JN, Laviolette JP, Ardlie K, Reich DE, Robinson E, Sklar P, Shah QTL on SSC17 [88]. The MYH3 gene was located on N, Thomas D, Fan JB, Gingeras T, Warrington J, Patil N, Hudson SSC12 and two SNPs of this gene both showed significant TJ, Lander ES (2000) Large-scale discovery and genotyping of association with several muscling and meat quality traits. single-nucleotide polymorphisms in the mouse. Nat Genet The association with rear leg, loin weight, Hunter B, and 24(4):381–386 5. Zimdahl H, Nyakatura G, Brandt P, Schulz H, Hummel O, shear force were significant at 5% experiment-wise level, Fartmann B, Brett D, Droege M, Monti J, Lee YA, Sun Y, Zhao and a QTL affecting ham weight was found to include this S, Winter EE, Ponting CP, Chen Y, Kasprzyk A, Birney E, gene [89]. Ganten D, Hubner N (2004) A SNP map of the rat genome generated from cDNA sequences. Science 303(5659):807 6. Fahrenkrug SC, Freking BA, Smith TP, Rohrer GA, Keele JW (2002) Single nucleotide polymorphism (SNP) discovery in Conclusions porcine expressed genes. Anim Genet 33(3):186–195 7. Grapes L, Rudd S, Fernando RL, Megy K, Rocha D, Rothschild This study annotated 1,439 porcine coding gene related MF (2006) Prospecting for pig single nucleotide polymorphisms in the human genome: have we struck gold? J Anim Breed Genet SNPs covering 18 autosomes according to their locations in 123(3):145–151 the genome as well as nucleotide substitution types. By 8. Goldstein DB, Weale ME (2001) Population genomics: linkage using direct PCR sequencing methods, 47.9% SNPs were disequilibrium holds the key. Curr Biol 11(14):R576–R579 confirmed to be polymorphisms in at least one of the five 9. Shastry BS (2002) SNP alleles in human disease and evolution. J Hum Genet 47(11):561–566 breeds tested. Furthermore, 226 novel mutations including 10. Marnellos G (2003) High-throughput SNP analysis for genetic 173 SNPs and 53 insertions/deletions were detected. The association studies. Curr Opin Drug Discov Devel 6(3):317–321 distribution of polymorphisms among the five breeds was 11. Kijas JW, Townley D, Dalrymple BP, Heaton MP, Maddox JF, calculated and compared, and the SNPs that result in amino McGrath A, Wilson P, Ingersoll RG, McCulloch R, McWilliam S, Tang D, McEwan J, Cockett N, Oddy VH, Nicholas FW, acid change were functionally evaluated by protein struc- Raadsma H (2009) International Sheep Genomics Consortium: a ture analyses. Subsequently, phenotypic association anal- genome wide survey of SNP variation reveals the genetic struc- yses with 33 economically important traits were performed ture of sheep breeds. PLoS ONE 4(3):e4668 on 261 SNPs from 180 genes covering the 18 autosomes 12. Emara MG, Kim H (2003) Genetic markers and their application in poultry breeding. Poult Sci 82(6):952–957 except chromosome 11. The functional and positional 13. Do KT, Ha Y, Mote BE, Rothschild MF, Choi BH, Lee SS, Kim study combining association analyses has revealed TH, Cho BW, Kim KS (2008) Investigation of single nucleotide numerous candidate genes affecting pig growth, carcass polymorphisms in porcine chromosome 2 quantitative trait loci and meat quality traits. Therefore, the large scale analyses for meat quality traits. Asian-Aust J Anim Sci 21(2):155–160 14. Fujii J, Otsu K, Zorzato F, de Leon S, Khanna VK, Weiler JE, of pig public SNPs combining re-sequencing and pheno- O’Brien PJ, MacLennan DH (1991) Identification of a mutation typic association studies may benefit the marker assisted in porcine ryanodine receptor associated with malignant hyper- selection breeding in the pig. thermia. Science 253(5018):448–451 15. Milan D, Jeon JT, Looft C, Amarger V, Robic A, Thelander M, Acknowledgments This work was supported by grants titled Rogel-Gaillard C, Paul S, Iannuccelli N, Rask L, Ronne H, ‘‘High-throughput sequencing and validation for genetic markers Lundstro¨m K, Reinsch N, Gellin J, Kalm E, Roy PL, Chardon P, development in Korean Native animals’’ (National Institute of Animal Andersson L (2000) A mutation in PRKAG3 associated with Science) and ‘‘Genetic improvement maximization of Korean native excess glycogen content in pig skeletal muscle. Science pigs using marker-assisted selection, and construction of commercial 288(5469):1248–1251 line production system’’ from the BioGreen 21 Program of the Korea 16. Kim KS, Larsen N, Short T, Plastow G, Rothschild MF (2000) A Rural Development Administration, ROK. missense variant of the porcine melanocortin-4 receptor (MC4R) gene is associated with fatness, growth, and feed intake traits. Mamm Genome 11(2):131–135 17. Kim KS, Reecy JM, Hsu WH, Anderson LL, Rothschild MF (2004) Functional and phylogenetic analyses of a melanocortin-4 receptor mutation in domestic pigs. Domest Anim Endocrinol References 26(1):75–86 18. Jungerius BJ, van Laere AS, Te Pas MF, van Oost BA, Andersson 1. Kruglyak L, Nickerson DA (2001) Variation is the spice of life. L, Groenen MA (2004) The IGF2-intron3–G3072A substitution Nat Genet 27(3):234–236 explains a major imprinted QTL effect on backfat thickness in a 2. Reich DE, Gabriel SB, Altshuler D (2003) Quality and com- Meishan x European white pig intercross. Genet Res pleteness of SNP databases. Nat Genet 33(4):457–458 84(2):95–101

123 Mol Biol Rep

19. de Koning DJ, Rattink AP, Harlizius B, van Arendonk JA, tool for positional cloning of candidate genes for major quanti- Brascamp EW, Groenen MA (2000) Genome-wide scan for body tative traits. Mol Cells 16(1):113–116 composition in pigs reveals important role of imprinting. Proc 38. Kim TH, Kim KS, Choi BH, Yoon DH, Jang GW, Lee KT, Natl Acad Sci USA 97(14):7947–7950 Chung HY, Lee HY, Park HS, Lee JW (2005) Genetic structure 20. Li XP, Hu ZL, Moon SJ, Do KT, Ha YK, Byun MJ, Choi BH, of pig breeds from Korea and China using microsatellite loci Rothschild MF, Reecy JM, Kim KS (2008) Development of an in analysis. J Anim Sci 83(10):2255–2263 silico coding gene SNP map in pigs. Anim Genet 39(4):446–450 39. Hirschhorn JN, Daly MJ (2005) Genome-wide association studies for 21. Batzoglou S, Pachter L, Mesirov JP, Berger B, Lander ES (2000) common diseases and complex traits. Nat Rev Genet 6(2):95–108 Human and mouse gene structure: comparative analysis and 40. Garn SM (1985) Relationship between birth weight and sub- application to exon prediction. Genome Res 10(7):950–958 sequent weight gain. Am J Clin Nutr 42(1):57–60 22. Choy YH, Jeon GJ, Kim TK, Choi BH, Cheong IC, Lee HK, Seo 41. Niles MJ, Matsuuchi L, Koshland ME (1995) Polymer IgM KS, Kim SD, Park YI, Chung HW (2002) Genetic analyses of assembly and secretion in lymphoid and nonlymphoid cell lines: carcass characteristics in crossbred pigs: cross between Landrace evidence that J chain is required for pentamer IgM synthesis. Proc sows and Korean wild boars. Asian-Aust J Anim Sci Natl Acad Sci USA 92(7):2884–2888 15:1080–1084 42. Koshland ME (1983) Presidential address: molecular aspects of B 23. Choy YH, Jeon GJ, Kim TK, Choi BH, Chung HW (2002) Ear cell differentiation, American Association of Immunologists, type and coat color on growth performances of crossbred pigs. April 1983. J Immunol 131(6):i–ix Asian-Aust J Anim Sci 15:1178–1181 43. Johansen FE, Braathen R, Brandtzaeg P (2001) The J chain is 24. Freudenberg-Hua Y, Freudenberg J, Kluck N, Cichon S, Propping essential for polymeric Ig receptor-mediated epithelial transport P, No¨then MM (2003) Single nucleotide variation analysis in 65 of IgA. J Immunol 167(9):5185–5192 candidate genes for CNS disorders in a representative sample of 44. Max E, Korsmeyer S (1985) Human J chain gene structure and the European population. Genome Res 13(10):2271–2276 expression in B lymphoid cells. J Exp Med 161(4):832–849 25. Conklin D, Jonassen I, Aasland R, Taylor WR (2002) Association 45. Havelock JC, Keller P, Muleba N, Mayhew BA, Casey BM, of nucleotide patterns with gene function classes: application to Rainey WE, Word RA (2005) Human myometrial gene expres- human 30 untranslated sequences. Bioinformatics 18(1):182–189 sion before and during parturition. Biol Reprod 72:707–719 26. Bell O, Silver J, Naveh-Many T (2005) Identification and char- 46. Zhao B, Koon D, Curtis AL, Soper J, Bethin KE (2007) Identi- acterization of cis-acting elements in the human and bovine PTH fication of 9 uterine genes that are regulated during mouse mRNA 30-untranslated region. J Bone Miner Res 20(5):858–866 pregnancy and exhibit abnormal levels in the cyclooxygenase-1 27. Li XL, Andersen JB, Ezelle HJ, Wilson GM, Hassel BA (2007) knockout mouse. Reprod Biol Endocrinol 5:28 Post-transcriptional regulation of RNase-L expression is medi- 47. Yasojima K, Tsujimura A, Mizuno T, Shigeyoshi Y, Inazawa J, ated by the 30-untranslated region of its mRNA. J Biol Chem Kikuno R, Kuma K, Ohkubo K, Hosokawa Y, Ibata Y, Abe T, 282(11):7950–7960 Miyata T, Matsubara K, Nakajima K, Hashimoto-Gotoh T (1997) 28. Kollers S, Me´gy K, Rocha D (2005) Analysis of public single Cloning of human and mouse cDNAs encoding novel zinc finger nucleotide polymorphisms in commercial pig populations. Anim proteins expressed in cerebellum and hippocampus. Biochem Genet 36(5):426–431 Biophys Res Commun 231(2):481–487 29. Meinhardt A, O’bryan MK, Mcfarlane JR, Loveland KL, Mallidis 48. Jankowski SA, Mitchell DS, Smith SH, Trent JM, Meltzer PS C, Foulds LM, Phillips DJ, de Kretser DM (1998) Localization of (1994) SAS, a gene amplified in human sarcomas, encodes a new follistatin in the rat testis. Reprod Fertil 112:233–241 member of the transmembrane 4 superfamily of proteins. Onco- 30. White KA, Yore MM, Deng D, Spinella MJ (2005) Limiting gene 9(4):1205–1211 effects of RIP140 in estrogen signaling: potential mediation of 49. Hemler ME (2003) Tetraspanin proteins mediate cellular pene- anti-estrogenic effects of retinoic acid. J Biol Chem tration, invasion, and fusion events and define a novel type of 280(9):7829–7835 membrane microdomain. Annu Rev Cell Dev Biol 19:397–422 31. White R, Leonardsson G, Rosewell I, Ann Jacobs M, Milligan S, 50. Levy S, Shoham T (2005) Protein–protein interactions in the Parker M (2000) The nuclear receptor co-repressor nrip1 (RIP140) tetraspanin web. Physiology (Bethesda) 20:218–224 is essential for female fertility. Nat Med 6(12):1368–1374 51. Schwer B, Guthrie C (1991) PRP16 is an RNA-dependent 32. Littlewood TD, Evan GI (1995) Transcription factors 2: helix- ATPase that interacts transiently with the spliceosome. Nature loop-helix. Protein Profile 2(6):621–702 349(6309):494–499 33. Kim SS, Kim JR, Moon JK, Choi BH, Kim TH, Kim KS, Kim JJ, 52. Schwer B, Guthrie C (1992) A dominant negative mutation in a Lee CK (2009) Transcriptional alteration of p53 related processes spliceosomal ATPase affects ATP hydrolysis but not binding to as a key factor for skeletal muscle characteristics in Sus scrofa. the spliceosome. Mol Cell Biol 12(8):3540–3547 Mol Cells 28:565–573 53. Richard M, Drouin R, Beaulieu AD (1998) ABC50, a novel 34. Li X, Kim SW, Choi JS, Lee YM, Lee CK, Choi BH, Kim TH, human ATP-binding cassette protein found in tumor necrosis Choi YI, Kim JJ, Kim KS (2010) Investigation of porcine FABP3 factoralpha stimulated synoviocytes. Genomics 53:137–145 and LEPR gene polymorphisms and mRNA expression for 54. Hopper E, Belinsky MG, Zeng H, Tosolini A, Testa JR, Kruh GD variation in intramuscular fat content. Mol Biol Rep (2001) Analysis of the structure and expression pattern of MRP7 37(8):3931–3939. doi:10.1007/s11033-010-0050-1 (ABCC10), a new member of the MRP subfamily. Cancer Lett 35. Choi KM, Moon JK, Choi SH, Kim KS, Choi YI, Kim JJ, Lee CK 162:181–191 (2008) Differential expression of cytochrome P450 genes regu- 55. Dean M, Rzhetsky A, Allikmets R (2001) The human ATP- late the level of adipose arachidonic acid in Sus Scrofa. Asian- binding cassette (ABC) transporter superfamily. Genome Res Aust J Anim Sci 21(7):967–971 11:1156–1166 36. Kim KS, Yeo JS, Kim JW (2002) Assessment of genetic diversity 56. Vasiliou V, Vasiliou K, Nebert DW (2009) Human ATP-binding of Korean native pig (Sus scrofa) using AFLP markers. Genes cassette (ABC) transporter family. Hum Genomics 3(3):281–290 Genet Syst 77(5):361–368 57. Miyazaki M, Ntambi JM (2003) Role of stearoyl-coenzyme A 37. Jeon JT, Park EW, Jeon HJ, Kim TH, Lee KT, Cheong IC (2003) desaturase in lipid metabolism. Prostaglandins Leukot Essent A large-insert porcine library with sevenfold genome coverage: a Fatty Acids 68(2):113–121

123 Mol Biol Rep

58. Jeffcoat R (2007) Obesity—a perspective based on the bio- 75. Wu KK, Liou JY (2005) Cellular and molecular biology chemical interrelationship of lipids and carbohydrates. Med of prostacyclin synthase. Biochem Biophys Res Commun Hypotheses 68(5):1159–1171 338(1):45–52 59. Kusunoki J, Kanatani A, Moller DE (2006) Modulation of fatty 76. Kim YO, Park SH, Kang YJ, Koh HJ, Kim SH, Park SY, Sohn U, acid metabolism as a potential approach to the treatment of Huh TL (1999) Assignment of mitochondrial NAD(?)-specific obesity and the metabolic syndrome. Endocrine 29(1):91–100 isocitrate dehydrogenase beta subunit gene (IDH3B) to human 60. Liu JS, Kuo SR, Makhov AM, Cyr DM, Griffith JD, Broker TR, chromosome band 20p13 by in situ hybridization and radiation Chow LT (1998) Human Hsp70 and Hsp40 chaperone proteins hybrid mapping. Cytogenet Cell Genet 86(3–4):240–241 facilitate human papillomavirus-11 E1 protein binding to the 77. Imagawa T, Smith JS, Coronado R, Campbell KP (1987) Purified origin and stimulate cell-free DNA replication. J Biol Chem ryanodine receptor from skeletal muscle sarcoplasmic reticulum 273(46):30704–30712 is the Ca2?-permeable pore of the calcium release channel. J Biol 61. Yeowell HN, Walker LC, Farmer BT, Heikkinen J, Myllyla R Chem 262(34):16636–16643 (2000) Mutational analysis of the lysyl hydroxylase 1 gene 78. Auernhammer CJ, Melmed S (2000) Leukemia-inhibitory factor- (PLOD) in six unrelated patients with Ehler–Danlos syndrome neuroimmune modulator of endocrine function. Endocr Rev type VI: prenatal exclusion of the disorder in one family. Hum 21(3):313–345 Mutat 16(1):90 79. Catunda AP, Go´cza E, Carstea BV, Hiripi L, Hayes H, Rogel- 62. Spotila LD, Rodriguez H, Koch M, Tenenhouse HS, Tenenhouse Gaillard C, Bertaud M, Bosze Z (2008) Characterization, chro- A, Li H, Devoto M (2003) Association analysis of bone mineral mosomal assignment, and role of LIFR in early embryogenesis density and single nucleotide polymorphisms in two candidate and stem cell establishment of rabbits. Cloning Stem Cells genes on chromosome 1p36. Calcif Tissue Int 73(2):140–146 10(4):523–534 63. Chiba S, Takeshita K, Imai Y, Kumano K, Kurokawa M, Masuda 80. Zhang S, Knight TJ, Stalder KJ, Goodwin RN, Lonergan SM, S, Shimizu K, Nakamura S, Ruddle FH, Hirai H (2003) Beitz DC (2007) Effects of breed, sex, and halothane genotype on Homeoprotein DLX-1 interacts with Smad4 and blocks a sig- fatty acid composition of pork longissimus muscle. J Anim Sci naling pathway from activin A in hematopoietic cells. Proc Natl 85(3):583–591 Acad Sci USA 100(26):15577–15582 81. Suzuki K, Shibata T, Kadowaki H, Abe H, Toyoshima T (2003) 64. Cobos I, Calcagnotto ME, Vilaythong AJ, Thwin MT, Noebels Meat quality comparison of Berkshire, Duroc and crossbred pigs JL, Baraban SC, Rubenstein JL (2005) Mice lacking Dlx1 show sired by Berkshire and Duroc. Meat Sci 64:35–42 subtype-specific loss of interneurons, reduced inhibition and 82. Lo LL, McLaren DG, McKeith FK, Fernando RL, Novakofski J epilepsy. Nat Neurosci 8(8):1059–1068 (1992) Genetic analyses of growth, real-time ultrasound, carcass, 65. Nakanishi S (1994) Metabotropic glutamate receptors: synaptic and pork quality traits in Duroc and Landrace pigs: I. Breed transmission, modulation, and plasticity. Neuron 13:1031–1037 effects. J Anim Sci 70:2373–2386 66. Duvoisin RM, Zhang C, Ramonell K (1995) A novel metabo- 83. Newcom DW, Stalder KJ, Bass TJ, Goodwin RN, Parrish FC, tropic glutamate receptor expressed in the retina and olfactory Wiegand BR (2004) Breed differences and genetic parameters of bulb. J Neurosci 15(4):3075–3083 myoglobin concentration in porcine longissimus muscle. J Anim 67. Nakanishi S (1992) Molecular diversity of glutamate receptors Sci 82:2264–2268 and implications for brain function. Science 258:597–603 84. Huff-Lonergan E, Baas TJ, Malek M, Dekkers JC, Prusa K, 68. Suijkerbuijk RF, Gebbink MF, Moolenaar WH, Geurts van Rothschild MF (2002) Correlations among selected pork quality Kessel A (1993) Fine mapping of the human receptor-like protein traits. J Anim Sci 80(3):617–627 tyrosine phosphatase gene (PTPRM) to 18p11.2 by fluorescence 85. Kim EH, Choi BH, Kim KS, Lee CK, Cho BW, Kim TH, Kim JJ in situ hybridization. Cytogenet Cell Genet 64(3–4):245–246 (2007) Detection of Mendelian and parent-of-origin quantitative 69. Ensslen-Craig SE, Brady-Kalnay SM (2005) PTP mu expression trait loci in a cross between Korean native pig and Landrace I. and catalytic activity are required for PTP mu-mediated neurite growth and body composition traits. Asian-Aust J Anim Sci outgrowth and repulsion. Mol Cell Neurosci 28(1):177–188 19:1702–1705 70. Burden-Gulley SM, Ensslen SE, Brady-Kalnay SM (2002) 86. Pierzchala M, Kuryl J, Reiner G, Bartenschlager H, Moser G, Protein tyrosine phosphatase-mu differentially regulates neurite Geldermann H (2003) Linkage and QTL mapping for Sus scrofa outgrowth of nasal and temporal neurons in the retina. J Neurosci chromosome 16. J Anim Breed Genet 120(1):126–131 22(9):3615–3627 87. Ponting CP, Schultz J, Milpetz F, Bork P (1999) SMART: 71. Ensslen SE, Rosdahl JA, Brady-Kalnay SM (2003) The receptor identification and annotation of domains from signalling and protein tyrosine phosphatase mu, PTPmu, regulates histogenesis extracellular protein sequences. Nucleic Acids Res 27(1): of the chick retina. Dev Biol 264(1):106–118 229–232 72. Ensslen SE, Brady-Kalnay SM (2004) PTPmu signaling via PKC 88. Pierzchala M, Cieslak D, Reiner G, Bartenschlager H, Moser G, delta is instructive for retinal ganglion cell guidance. Mol Cell Geldermann H (2003) Linkage and QTL mapping for Sus scrofa Neurosci 25(4):558–571 chromosome 17. J Anim Breed Genet 120(1):132–137 73. Blanchetot C, Chagnon M, Dube´ N, Halle´ M, Tremblay ML 89. Yue G, Schro¨ffel J Jr, Moser G, Bartenschlager H, Reiner G, (2005) Substrate-trapping techniques in the identification of cel- Geldermann H (2003) Linkage and QTL mapping for Sus scrofa lular PTP targets. Methods 35(1):44–53 chromosome 12. J Anim Breed Genet 120(1):95–102 74. Phillips-Mason PJ, Gates TJ, Major DL, Sacks DB, Brady-Kalnay SM (2006) The receptor protein-tyrosine phosphatase PTPmu interacts with IQGAP1. J Biol Chem 281(8):4903–4910

123