GWAS and gene networks for milk-related traits from test-day multiple lactations in Portuguese Holstein cattle

This study focused on the identification of QTL regions, candidate genes, and network related genes based on the first 3 lactations (LAC3) of milk, fat, and protein yields, and somatic cell score (SCS) in Portuguese Holstein cattle. Additionally, the results were compared with those from only first lactation (LAC1) data. The analyses were performed using the weighted single-step GWAS under an autoregressive test-day (TD) multiple lactations model. A total of 11,434,294 and 4,725,673 TD records from LAC3 and LAC1, respectively, including 38,323 autosomal SNPs and 1338 genotyped animals were used in GWAS analyses. A total of 51 (milk), 5 (fat), 24 (protein), and 4 (SCS) genes were associated to previously annotated relevant QTL regions for LAC3. TheCACNA2D1at BTA4 explained the highest proportion of genetic variance respectively for milk, fat, and protein yields. For SCS, theTRNAG-CCCat BTA14,MAPK10,andPTPN3genes, both at BTA6 were considered important candidate genes. The accessed network refined the importance of the reported genes.CACNA2D1regulates calcium density and activation/inactivation kinetics of calcium transport in the mammary gland; whereasTRNAG-CCC,MAPK10, andPTPN3are directly involved with inflammatory processes widely derived from mastitis. In conclusion, potential candidate genes (TRNAG-CCC,MAPK10, andPTPN3) associated with somatic cell were highlighted, which further validation studies are needed to clarify its mechanism action in response to mastitis. Moreover, most of the candidate genes identified were present in both (LAC3 and LAC1) for milk, fat and protein yields, except for SCS, in which no candidate genes were shared between LAC3 and LAC1. The larger phenotypic information provided by LAC3 dataset was more effective to identify relevant genes, providing a better understanding of the genetic architecture of these traits over all lactations simultaneously.