Genomic Prediction and Genome-Wide Association Analysis of Heat Tolerance for Milk Yield in Buffaloes Using a Reaction Norm Model.

The aim of this study was to evaluate the impact of incorporating genomic information on the estimation of genetic (co)variance components and the accuracy of breeding values for milk yield under varying thermal environments, and to identify SNPs associated with genes that play significant roles in heat tolerance. We analysed 58,070 test-day milk yield records from 3459 first lactations, collected between 1987 and 2018 from six herds. Genotypic data consisted of 870 animals genotyped for 45,405 SNP markers. Climatic data were obtained from INMET and combined into a temperature-humidity index (THI). Breeding values for test-day milk yield across THI values and days in milk were estimated using both genomic and pedigree-based random regression animal models. The model incorporating genomic information yielded higher estimates of heritability and additive genetic variance, along with improved accuracy under heat stress conditions and better modelling of genotype-by-environment interaction, making it a promising approach for predicting breeding values. GWAS results were reported based on the proportion of genetic variance explained by sliding windows of five consecutive SNPs, with regions explaining more than 1% of the variance in heat tolerance selected for further consideration. The ESRRG, IGSF5 and PCP4 genes emerged as strong candidates associated with heat tolerance in milk yield.