Functional Partitioning of Genomic Variance and Genome-Wide Association Study for Carcass Traits in Korean Hanwoo Cattle Using Imputed Sequence Level SNP Data

Mohammad S A Bhuiyan; Dajeong Lim; Mina Park; Soohyun Lee; Yeongkuk Kim; Cedric Gondro; Byoungho Park; Seunghwan Lee

doi:10.3389/fgene.2018.00217

Functional Partitioning of Genomic Variance and Genome-Wide Association Study for Carcass Traits in Korean Hanwoo Cattle Using Imputed Sequence Level SNP Data

Front Genet. 2018 Jun 22:9:217. doi: 10.3389/fgene.2018.00217. eCollection 2018.

Authors

Mohammad S A Bhuiyan^{1

2}, Dajeong Lim³, Mina Park⁴, Soohyun Lee¹, Yeongkuk Kim¹, Cedric Gondro⁵, Byoungho Park⁴, Seunghwan Lee¹

Affiliations

¹ Department of Animal Science and Biotechnology, Chungnam National University, Daejeon, South Korea.
² Department of Animal Breeding and Genetics, Bangladesh Agricultural University, Mymensingh, Bangladesh.
³ Division of Animal Genomics and Bioinformatics, National Institute of Animal Science, Rural Development Administration, Wanju, South Korea.
⁴ Animal Genetic Improvement Division, National Institute of Animal Science, Rural Development Administration, Seonghwan, South Korea.
⁵ College of Agriculture and Natural Resources, Michigan State University, East Lansing, MI, United States.

Abstract

Quantitative traits are usually controlled by numerous genomic variants with small individual effects, and variances associated with those traits are explained in a continuous manner. However, the relative contributions of genomic regions to observed genetic variations have not been well explored using sequence level single nucleotide polymorphism (SNP) information. Here, imputed sequence level SNP data (11,278,153 SNPs) of 2109 Hanwoo steers (Korean native cattle) were partitioned according to functional annotation, chromosome, and minor allele frequency (MAF). Genomic relationship matrices (GRMs) were constructed for each classified region and fitted in the model both separately and together for carcass weight (CWT), eye muscle area (EMA), backfat thickness (BFT), and marbling score (MS) traits. A genome-wide association study (GWAS) was performed to identify significantly associated variants in genic and exon regions using a linear mixed model, and the genetic contribution of each exonic SNP was determined using a Bayesian mixture model. Considering all SNPs together, the heritability estimates for CWT, EMA, BFT, and MS were 0.57 ± 0.05, 0.46 ± 0.05, 0.45 ± 0.05, and 0.49 ± 0.05, respectively, which reflected substantial genomic contributions. Joint analysis revealed that the variance explained by each chromosome was proportional to its physical length with weak linear relationships for all traits. Moreover, genomic variances explained by functional category and MAF class differed greatly among the traits studied in joint analysis. For example, exon regions had larger contributions for BFT (0.13 ± 0.08) and MS (0.22 ± 0.08), whereas intron and intergenic regions explained most of the total genomic variances for CWT and EMA (0.22 ± 0.09-0.32 ± 0.11). Considering different functional classes of exon regions and the per SNP contribution revealed the largest proportion of genetic variance was attributable to synonymous variants. GWAS detected 206 and 27 SNPs in genic and exon regions, respectively, on BTA4, BTA6, and BTA14 that were significantly associated with CWT and EMA. These SNPs were harbored by 31 candidate genes, among which TOX, FAM184B, PPARGC1A, PRKDC, LCORL, and COL1A2 were noteworthy. BayesR analysis found that most SNPs (>93%) had very small effects and the 4.02-6.92% that had larger effects (10^-4 × $σ_{A}^{2}$ , 10^-3 × $σ_{A}^{2}$ , and 10^-2 × $σ_{A}^{2}$ ) explained most of the total genetic variance, confirming polygenic components of the traits studied.

Keywords: GWAS; Hanwoo cattle; carcass traits; genome level SNP; variance partitioning.