Milk for all models. The fraction of additive genetic

Milk yield (MY) and fat yield (FY) are economically important traits for Thai dairy business. Genetic prediction for MY and FY in Thailand uses only pedigree and phenotypic information. Combining SNP genotypes of individual animals with pedigree and phenotype would be expected to increase accuracy of genetic predictions and speed up selection progress. The dataset consisted of first-lactation MY and FY records from 600 cows from 56 farms in Central Thailand collected from 2000 to 2013. The mixed model contained herd-year-season, Holstein fraction, heterozygosity of the cow and age at first calving as fixed effects (all model). Random effects were SNP genomic (GP and G), animal polygenic (GP and P) and residual. Variance components were estimated using GS3 software (GP and P). Additive genetic predictions were computed with GS3 for all models. The fraction of additive genetic variances explained by the 8,257 SNP from GGP-LD and computed with the GP model were 50% for MY and 48% for FY. Heritability estimates with the GP model were higher (0.38 for MY and 0.41 for FY) than those with the P model (0.28 for MY and 0.30 for FY). Rank correlations between GP and G model were the highest (0.99 for both MY and FY; P<0.01), followed by correlations between GP and P models (0.91 for MY and 0.75 for FY; P<0.01), and the lowest correlations were between G and P models (0.89 for MY and 0.73 for FY; P<0.01). Predictive abilities of GP model were higher (0.5201 to 0.6068; P<0.01) than G and P model (0.5046 to 0.6060; P<0.01) for MY and FY. Accuracies of GP model were higher (0.8550 to 0.9594) than G and P model (0.8295 to 0.9582) for MY and FY. SNP from GeneSeek GGP-LD 9k BeadChip not only accounted for a sizeable fraction of the additive genomic variance for MY and FY, but they also yielded animal genomic EBV whose ranking was highly correlated with rankings of both genomic-polygenic and polygenic EBV. These results indicated that utilization of GGP-LD and perhaps higher density genotyping chips, would be advantageous for improve accuracy of prediction and selection in Central Thailand.