Comparison of genetic evaluations for milk yield and fat yield using a polygenic model and three genomic–polygenic models with different sets of SNP genotypes in Thai multibreed dairy cattle

Abstract

The objectives of this research were to compare variance components, genetic parameters, prediction accuracies, and ranking of animals for 305-d milk yield (MY) and 305-d fat yield (FY) using a polygenic and three genomic–polygenic models in a Thai multibreed dairy population. The genomic–polygenic models utilized 7656 SNP (GM7K), 74,144 actual and imputed 80K SNP from FImpute (GM80K-FI), and 73,600 actual and imputed 80K SNP from Findhap (GM80K-FH). Traits were gathered from 8361 first-lactation cows in 810 farms that had their first calving between 1989 and 2014. Variance components and genetic parameters were estimated using REML procedures. Fixed effects included contemporary group (herd-year-season), calving age and heterosis. Random effects were animal additive genetic and residual. Estimates of variance components, heritabilities and prediction accuracies for MY and FY from GM80K-FI and GM80K-FH were similar. They were also the highest, followed by those from GM7K, and the lowest were those from the polygenic model. Correlations estimates between MY and FY were similar across models. Different MY and FY EBV rankings existed across models. The highest rank correlations were those between rankings from GM80K-FI and GM80K-FH. The second highest rank correlations between rankings from GM7K and GM80K-FI, and GM7K and GM80K-FH. The lowest rank correlations were between rankings from the polygenic model and the three genomic–polygenic models. Rank correlations indicated that selection response would differ when choosing replacement animals based on rankings from polygenic and genomic–polygenic EBV. Accuracy of EBV indicated that the highest expected selection responses for MY and FY would be achieved by utilizing EBV from GM80K-FI and GM80K-FH. Lastly, the similarity between results from GM80K-FI and GM80K-FH suggested that genotype imputation with either FImpute or Findhap would be appropriate for this Thai multibreed dairy population.