493
Efficiency of including cows in genomic prediction versus multivariate models for fertility traits in dairy cattle in the United Kingdom
The efficiency of genomic evaluations with including cows in the reference population was compared to multivariate genomic evaluations for calving interval (CI) and non-return rate at 56 days (NR56). Cow data included constituted 10 and 15% of the reference set for CI and NR56 respectively. Analyses were based on SNP BLUP, BayesA and BayesC models. Inclusion of cow data had very little impact on accuracies and errors of predictions of genomic evaluations for CI and NR56. The bivariate genomic models resulted in an increase of 15% in accuracy for NR56. Similar increases in accuracies of 10 to 15% were observed for NR56 from the Bayesian models. All of the analyses had very little effect on accuracies for CI. The study concluded that multivariate SNP BLUP models resulted in more efficient genomic evaluations rather than including cows.
Keywords:
Multivariate genomic models
Bayesian methods
Genomic evaluations
Fertility traits
