Genome-Wide Association (GWAS) Analysis of Bovine Milk Fatty Acids of Canadian Holstein Cows Using Genotyping-By-Sequencing (GBS) Markers

Bovine milk is a natural source of omega-3 and -6 fatty acids (FA) in human diet. Considerable breed and individual variations in milk FA composition has led to associations with candidate genes and genome-wide markers on the BovineSNP50 BeadChip. Genotyping-by-sequencing (GBS) of DNA from 1,200 Canadian Holstein milk samples on the Illumina HiSeq 2000 generated 1,4635,202 unique tags of which 79.6% aligned to unique positions on the bovine genome giving rise to a total of 515,820 SNPS identified on all chromosomes. Imputation by Beagle was done to correct for low frequency variants and 334,342 SNPS with call rate >95%, accuracy of imputation score >50% and minor allele frequency >0.01 were retained and used for genome-wide association (GWAS) with milk FAs.  GWAS was accomplished with the single-locus linear mixed model procedure implemented in Golden Helix SVS v7.7.8 software. Fifty nine and 76 SNPs showed genome wide significance (P-value BH FDR <0.05, range from 0.049 to 7.41e-05) with eicosapentaenoic acid (EPA) and arachidonic acid (AA), respectively. Although majority of associated SNPs were found within intergenic regions, others were located on 26 genes for AA and 23 genes for EPA.  Among these genes, only five are implicated in lipogenesis. Novel SNPs and genes identified here could be used for improvement of milk EPA and AA contents. Our results demonstrate the potential of the GBS technique for studies on economically important traits in cattle. Furthermore, our study shows that GBS is a robust, cost effective method to identify population-based unique SNPs for GWAS.