Integrative analysis of GWAS, Bayesian fine-mapping, Mendelian randomization and colocalization reveals genetic determinants underlying milk-related traits in dairy cattle

Genome-wide association studies (GWAS) has identified many genetic variants associated with milk-related traits in dairy cattle. However, the causal variants or genes remain largely unknown. In this study, using a large population (> 10,000 individuals) of Chinese Holstein cattle, we performed GWAS for six milk-related traits (milk yield, fat percentage, protein percentage, fat yield, protein yield, and somatic cell score) and subsequently prioritized putative causal variants by multi-trait Bayesian fine-mapping and examined the causal genes by Mendelian randomization (MR) analysis incorporating GWAS and cis-eQTL summary data from CattleGTEx. We also conducted a colocalization analysis to identify the share putative causal variants behind the milk-related traits and gene expressions. A total of 9,688 genome-wide significant SNPs (P < 1.2 × 10−7) were identified across the GWAS results for six milk-related traits, and these SNPs were distributed in 25 unique QTL regions. Subsequently, the multi-trait Bayesian fine-mapping identified 211 independent credible sets (CS) containing putative causal variants within these QTL regions. Among these CSs, 189 CSs were significantly associated with at least one trait (average lfsr < 0.01). Notably, the lead SNPs within these significant CSs included 3 missense variants and 62 non-coding transcript variants. The MR analysis detected 268 causal associations between gene expression and milk-related traits. The colocalization analysis identified two regions containing common putative causal variants for one or multiple milk-related traits and the expressions of some genes. Our integrative analysis of GWAS, Bayesian fine-mapping, MR, and colocalization further confirmed the well-known causal associations of DGAT1 and GHR and the milk-related traits. In addition, we revealed some novel potential causal genes, including AHNAK, ARHGEF2, SOX13, FDPS, SCGB2A2, and MROH2B. These results enhance our understanding of genetic mechanisms underlying the milk-related traits in dairy cattle.