Validation of Single Nucleotide Polymorphisms Associated with Carcass Traits in a Commercial Hanwoo Population.

Four carcass traits, namely carcass weight (CW), eye muscle area (EMA), back fat thickness (BF), and marbling score (MS), are the main price decision parameters used for purchasing Hanwoo beef. The development of DNA markers for these carcass traits for use in a beef management system could result in substantial profit for beef producers in Korea. The objective of this study was to validate the association of highly significant single nucleotide polymorphisms (SNPs) identified in a previous genome-wide association study (GWAS) with the four carcass traits in a commercial Hanwoo population. We genotyped 83 SNPs distributed across all 29 autosomes in 867 steers from a Korean Hanwoo feedlot. Six SNPs, namely ARS-BFGL-NGS-22774 (Chr4, Pos:4889229), ARS-BFGL-NGS-100046 (Chr6, Pos:61917424), ARS-BFGL-NGS-39006 (Chr27, Pos:38059196), ARS-BFGL-NGS-18790 (Chr10, Pos:26489109), ARS-BFGL-NGS-43879 (Chr9, Pos:39964297), and BTB-00775794 (Chr20, Pos:20476265), were found to be associated with CW, EMA, BF, and MS. The ARS-BFGL-NGS-22774, BTB-00775794, and ARS-BFGL-NGS-39006 markers accounted for 1.80%, 1.72%, and 1.35% (p<0.01), respectively, of the phenotypic variance in the commercial Hanwoo population. Many genes located in close proximity to the significant SNPs identified in this study were previously reported to have roles in carcass traits. The results of this study could be useful for marker-assisted selection programs.

INTRODUCTION

Meat quality and yield traits such as carcass weight (CW), eye muscle area (EMA), back fat thickness (BF), and marbling score (MS) are key traits used to determine the grade and market price of Hanwoo beef. Thus, knowing the genetic potential of an animal for these traits at an early age becomes an important decision-making factor for the farmer. The ability to perform an early genetic assessment of these traits would provide significant benefits to the Korean beef industry.

Many genome-wide association studies (GWAS) have identified single nucleotide polymorphisms (SNPs) that are significantly associated with carcass traits in different cattle breeds (Nishimura et al., 2012; Lee et al., 2013b; Lu et al., 2013). In a previous GWAS study performed by Lee et al. (2013b), statistically significant SNPs were identified on Bos taurus autosome (BTA) 14 for carcass traits in a progeny-tested Hanwoo population. Before any SNP marker is delivered to the beef industry, a validation study that accurately estimates the additive genetic marker effect in a population different from the discovery population should be done. Moreover, a validation population should be considered for different management practices, environment, and genetic structure (Barendse et al., 2007; Van Eenennaam et al., 2007). Validation of the effects of genetic markers in independent populations would determine the utility of the genetic testing technology and support the process of incorporating the DNA tests into nationwide cattle evaluation programs. Validated markers could then be used to estimate the genomic breeding value of female animals sired by national bulls on small Hanwoo farms.

Further identification of genes with large effects on quantitative traits would improve the efficiency of animal selection and increase our understanding of the underlying biology of quantitative traits. Given the importance of validating markers, we genotyped 83 SNP markers, which were significantly identified in a previous GWAS performed by Lee et al. (2013b), and validated the effects of those markers in a commercial Hanwoo population in this study.

MATERIALS AND METHODS

Animals, trait measurements, and genotyping

DNA was extracted from blood samples of 867 Hanwoo steers belonging to six feedlot groups comprising animals derived from 75 KPNs (Hanwoo proven bull). Blood samples were collected in the central abattoir of Korea (Eum-Sung Abattoir, National Agricultural Cooperative Federation, Eum-Sung, Korea). Management information was obtained from the Korea Animal Improvement Association and Traceability Records for Animal Products. Approval was obtained from the Animal Care and Use Committee for this study based on the animal health and welfare guidelines of the National Institute of Animal Science, Rural Development Administration, Korea. A genome-wide SNP validation analysis was performed using 83 markers (Supplementary Tables S1 to S4), which were genotyped using the Illumina GoldenGate and Infinium genotyping assays to detect the association of SNPs with four phenotypes: CW, EMA, BF, and MS. The SNPs used in this study were selected from the previous GWAS results of Lee et al. (2013b). Phenotypic data were provided by the abattoir (Table 1) and were recorded using standard measurement methods, i.e., CW (kg) was the weight after slaughter at 32 months of age, and EMA (cm2), BF (mm), and MS were measured at the twelfth and thirteenth rib junctions after 24 hours of chilling. MS was assessed on a scale of 1 to 9 points following the beef grading system, which has been standardized by the Korea Institute for Animal Products Quality Evaluation (http://www.ekape.or.kr/view/eng/system/beef.asp).

Table 1

Observed phenotypic data from a commercial Hanwoo population

Trait	N	Mean	SD	Min.	Max.	σ2p
Carcass weight (kg)	867	418.26	42.40	302	555	1,797.39
Eye muscle area (cm2)	867	87.85	10.55	53	132	111.39
Back fat thickness (mm)	867	12.62	4.55	3	33	20.73
Marbling score (1 to 9)	867	4.45	2.12	1	9	4.50

N, number of animal; SD, standard deviation; Min., minimum value in the traits; Max., maximum value in the traits; σ2p, phenotypic variation.

Statistical analysis

In a previous GWAS (Lee et al., 2013b), highly significant SNPs were identified for CW (22 SNPs), EMA (29 SNPs), BF (9 SNPs), and MS (23 SNPs). A multiple linear regression model was implemented to identify the best SNPs and their effects. To validate significant SNPs from the GWAS analysis, we fitted all significant SNPs as fixed effects in a multiple regression model with the validation dataset (n = 867). The additive effect of each SNP on carcass traits (CW, EMA, BF, and MS) was calculated by multiple regression analysis, with values of the covariate coded as 0, 1, or 2 copies of the variant allele. The model was described according to the equation:

Where, Y is a vector of carcass traits, μ is a vector of the overall mean for carcass traits, CG is a vector of the contemporary group for birth year and season, b is a regression coefficient, Sday is a vector of the day of slaughter as a covariate effect, nSNP indicates the number of SNPs to be tested in a multiple regression model, SNP is a single-locus SNP genotype coded as 0, 1, or 2 as a covariate effect, and e is a vector of random residual .

Proportion of SNP variance and bioinformatics analysis

The phenotype data were analyzed with a basic linear model using the R programming language (R Core Team, 2015) to determine the effect of markers and their significance in the studied animals. We generated four model equations indicating four traits and their candidate SNPs from previous GWAS results (Lee et al., 2013b). Genotype was the effect of interest in the model. The additive effect was fitted as a regression of phenotype on allele count. The results of the regression analyses of carcass phenotypes on genotypes are tabulated in Supplementary Tables S1 to S4. The percentage of the phenotypic variance explained by each significant SNP was calculated using the equation:

Where, p and q are the allele frequencies for the ith SNP estimated for Hanwoo, a is the estimated additive effect of the ith SNP on the phenotype in the equation, and is a phenotypic variance for meat quality traits in Hanwoo. SNPs were considered significant at p≤5%. Identified genes that were colocated with or located near the significant SNPs were annotated based on information from GeneCards (Rebhan et al., 1998) and AmiGO (Carbon et al., 2009).

RESULTS AND DISCUSSION

The results of this validation study, including allele frequencies, are summarized in Supplementary Tables S1 to S4. In a panel of 83 SNPs distributed across 29 autosomes, six markers were significantly associated (p≤0.05) with carcass traits in the commercial Hanwoo population (Table 2). ARS-BFGL-NGS-22774 (Chr4, Pos:4889229) was associated with CW; ARS-BFGL-NGS-100046 (Chr6, Pos:61917424) and ARS-BFGL-NGS-39006 (Chr27, Pos:38059196) were associated with EMA; ARS-BFGL-NGS-18790 (Chr10, Pos:26489109) was associated with BF; and ARS-BFGL-NGS-43879 (Chr9, Pos:39964297) and BTB-00775794 (Chr20; Pos: 20476265) were associated with MS. Two SNPs, BTB-00775794 and ARS-BFGL-NGS-39006, were the most significant markers (p<0.01), for MS and EMA, respectively. These two markers, along with ARS-BFGL-NGS-18790, exhibited negative effects on MS, EMA, and BF, respectively. Furthermore, the BTB-00775794 and ARS-BFGL-NGS-39006 markers in combination with ARS-BFGL-NGS-22774 moderately explained phenotypic variation (1.72%, 1.35%, and 1.80%, respectively) with respect to the total phenotypic variation in the carcass traits. Genes located in the close proximity to the significant SNPs are compiled in Table 3.

Table 2

Results for six significant SNPs associated with meat quality traits in a commercial Hanwoo population

Marker	NCBI reference	SNP1	BTA	Position2	GWAS results3								Validation results
					Allele frequency4					p-value (F-value)	%Vg5	Trait	Allele frequency					Effect6	p-value (F-value)	%Vp7
					N	n0	n1	n2	MAF				N	n0	n1	n2	MAF
ARS-BFGL-NGS-22774	rs109952763	G	4	4889229	1,090	790	278	22	0.15	0.0004893 (12.2)	2.83	CW	857	673	171	13	0.11	12.61	0.0356 (4.45)	1.80
ARS-BFGL-NGS-100046	rs110527834	G	6	61917424	1,089	847	228	14	0.12	0.0002482 (13.53)	3.01	EMA	865	678	181	6	0.11	1.98	0.0314 (4.65)	0.70
ARS-BFGL-NGS-43879	rs109783827	A	9	39964297	1,089	557	442	90	0.29	0.0002597 (13.44)	1.51	MS	864	496	320	48	0.24	0.30	0.0149 (5.96)	0.71
ARS-BFGL-NGS-18790	rs110384732	A	10	26489109	1,085	362	559	164	0.41	0.0005577 (8.87)	2.32	BF	839	279	413	147	0.42	−0.45	0.0520 (3.79)	0.48
BTB-00775794	rs41937398	G	20	20476265	1,091	364	546	181	0.42	0.0001272 (14.8)	1.71	MS	853	288	420	145	0.42	−0.40	0.0005 (12.20)	1.72
ARS-BFGL-NGS-39006	rs109741381	C	27	38059196	1,090	442	510	138	0.36	0.0004045 (12.6)	2.60	EMA	864	307	414	143	0.41	−1.77	0.0086 (6.96)	1.35

SNPs, single nucleotide polymorphisms; NCBI, National Center for Biotechnology Information; BTA, Bos taurus autosome; GWAS, genome-wide association study; MAF, minor allele frequency; CW, carcass weight; EMA, eye muscle area; BF, back fat thickness; MS, marbling score.

SNP constitute to minor allele.

Marker positions are based on the Btau 4.1 assembly of the bovine genome sequence.

Genome-wide association study of meat quality traits in Hanwoo cattle performed by Lee et al. (2013b).

n0, major homozygous alleles; n1, heterozygous alleles; and n2, homozygous least common alleles.

Proportion of SNP variation.

Estimated effect of the minor allele derived from regression analyses denominated based on the traits: CW (kg), EMA (cm2), BF (mm), MS (1–9).

Percentage of phenotypic variation calculated from the equation presented in the Materials and Methods section.

Table 3

Genes located in close proximity to the significant SNPs associated with meat quality traits in a commercial Hanwoo population

Trait	Marker name	BTA	Position1	Effect2	Gene information
					Name	Description
Carcass weight (CW)	ARS-BFGL-NGS-22774	4	4889229	12.61	COBL	Member of the G/F-actin binding protein family
Eye muscle area (EMA)	ARS-BFGL-NGS-100046	6	61917424	1.98	RBM47	An RNA-binding motif protein
	ARS-BFGL-NGS-39006	27	38059196	−1.77	ZMAT4	Zinc ion binding
Back fat thickness (BF)	ARS-BFGL-NGS-18790	10	26489109	−0.45	OR11G2OR11H6-likeOR11H12	Olfactory receptor (OR) genes
Marbling score (MS)	ARS-BFGL-NGS-43879	9	39964297	0.30	LAMA4	Intramuscular extracellular matrix (ECM) protein family
	BTB-00775794	20	20476265	−0.40	PDE4D	Phosphodiesterase activator and regulates β2-adrenergic receptor
					RAB3C	Guanine triphosphate (GTP) binding protein family

SNP, single nucleotide polymorphism; BTA, Bos taurus autosome.

Marker positions are based on the Btau 4.1 assembly of the bovine genome sequence.

Estimated effect of the minor allele derived from regression analyses denominated based on the traits: CW (kg), EMA (cm2), BF (mm), MS (1 to 9).

The BTB-00775794 marker on BTA20 exhibited an allele substitution effect of −0.40 on MS. BTA20 was previously reported to harbor quantitative trait loci (QTLs) for carcass traits including beef marbling (Han et al., 2009; Garcia et al., 2010). Genes located in close proximity to the significant marker included cyclic adenine monophosphate (cAMP)-specific phosphodiesterase 4D (PDE4D) and Ras-related protein Rab-3C (RAB3C). PDE4D plays a role in phosphodiesterase activity in skeletal muscle (Fleming-Waddell et al., 2009; Yu, 2013). PDE4D regulates β2-adrenergic receptor, which is found mostly in adipose tissue. If β2-adrenergic regulated, the size and proportion of myofibers will increase and the body fat content, which is important to develop marblingness, will decrease. This mechanism was connected to the activity of RAB3C gene, a member of the Ras-related guanine triphosphate (GTP) binding proteins family. GTP was capable to bind the G protein, which is used to compose the β2-adrenergic signal and activate cAMP (Talton, 2006).

The ARS-BFGL-NGS-39006 (BTA27) marker had a negative effect (−1.77 cm2) on EMA. Nalaila et al. (2011) previously reported a QTL at approximately 40 cM on BTA27 to be associated with carcass traits including rib eye area in beef cattle. The zinc finger matrin type-4 (ZMAT4) gene was found to be located close to the identified SNP (~38 cM). Downregulation of ZMAT4 and β2-adrenergic receptor genes affects insulin production (Jost et al., 2002), which promotes muscle development. Cattle that are more muscular and have larger EMA values accumulate high levels of muscle glycogen and subcutaneous fat (McGilchrist et al., 2012).

The ARS-BFGL-NGS-18790 (BTA10) marker exhibited a negative effect (−0.45 mm) on BF. Many olfactory receptor (OR) genes were identified near this SNP including the OR11G2 (LOC784260), OR11H6-like (LOC784302), and OR11H12 (LOC784332) genes. In Hanwoo cattle, a large number of OR genes (1,423) were identified across chromosomes. The highest number was on BTA15 (303), and 80 OR genes were observed on BTA10 (Lee et al., 2013a). OR proteins detect G protein and promote its binding to GTP. G protein is a component of the β2-adrenergic receptor (Talton, 2006), which plays a role in accumulation of muscle glycogen and subcutaneous fat.

The ARS-BFGL-NGS-22774 (BTA4) marker exhibited a significant association with CW, with an estimated effect of 12.61 kg. Another marker in the same vicinity (4.5 cM) was also identified previously to be associated with carcass traits in Hanwoo cattle (Li and Kim, 2015). A cordon-bleu WH2 repeat protein (COBL) gene was found in close proximity to the significant SNPs. COBL is a member of the G/F-actin binding protein family, which act as actin nucleators (Renault et al., 2008; Campellone and Welch, 2010; Wayt and Bretscher, 2014). G- and F-actin are constituents of myofibrils, which are small membraneless parallel fibers in muscle (Murray, 1995). An increase in myofibrils promotes hyperplasia, which involves muscle growth and is a biological basis for meat tenderness (Koohmaraie et al., 2002) and mass (Owens et al., 1993).

The ARS-BFGL-NGS-43879 (BTA9) marker exhibited a significant association with MS. The effect of this marker on MS was 0.3. The laminin alpha 4 (LAMA4) gene was located near this marker. LAMA4 is a member of the intramuscular extracellular matrix (ECM) protein family and plays a role in myogenesis (Okazaki et al., 2002; Miner and Yurchenco, 2004). ECM signaling pathway has been found to be up regulated and play critical functions in metabolic network for intramuscular adipose tissue deposition of Hanwoo cattle (Lee et al., 2014). It was already known that the degree of intramuscular adipose deposition in the 13th rib interface of the longissimus dorsi muscle tissue will determines the level of MS.

The ARS-BFGL-NGS-100046 (BTA6) marker exhibited a significant association with EMA, with an allele substitution effect of 1.98 cm2. An RNA-binding motif protein 47 (RBM47) gene was located close to this marker (61.9 cM). A similar region and candidate gene were shown to be associated with milk production traits in Chinese Holstein dairy cattle (Hu et al., 2010). RBM47 in vertebrates tend to interact with Wing and Int (Wnt)/β-catenin signaling pathway and act as tissue-specific regulators (Guan et al., 2013; Vanharanta et al., 2014). In ruminant, Wnt will promote myogenesis if it regulated (Du et al., 2010). As studied by Jeong et al. (2013) in the Hanwoo steer, this mechanism contributes to increase intramuscular fat deposition in the longissimus dorsi.

Major loci on BTA14 that have been found significantly in the previous GWAS study are not included in the present study because these effects have already validated from many studies (Sharma et al., 2014; Li and Kim, 2015). The amount of effect for SNPs that were found significant in GWAS study might vary in the other population. The SNP effect could be influenced by the environmental conditions and different management practices of the animals used in the validation study. Furthermore, the other possible reason is the fact that many selection evidences have remained unrevealed in the association study due to the possible bias (Qanbari and Simianer, 2014), so that different significant SNPs could be found.

In conclusion, this validation study confirmed that six SNPs are associated with the four carcass traits in a commercial Hanwoo population. Many genes that lie in close proximity to the significant markers identified in this study have been previously been reported to have roles in carcass traits. The outcome of this study could be exploited for marker-assisted selection programs.