A Continuous Genomic Evaluation System for German Holsteins

Abstract

In March 2014, a continuous genomic evaluation system was introduced in German Holsteins, in addition to routine monthly genomic evaluation. Genomic selection on embryos can lead to a shortened generation interval and substantial savings for breeders. The objectives of this study were to compare genomic breeding values (GEBV) of the continuous with the routine genomic evaluation systems and to investigate the effect of genotype imputation on genomic predictions of embryos. A total of 2440 animals genotyped with Illumina 50K, LD and EuroG10K chips were available between March and April 2014 for comparing genomic evaluations between the two systems. Average differences for all 44 evaluated traits ranged from -4% to 1% genetic standard deviations for direct genomic values (DGV) and from -10% to 5% genetic standard deviations for GEBV. Depending on the traits, correlations of DGV between the two systems were above 0.995, 0.992, and 0.988 for animals genotyped with 50K, EuroG10K and LD chips, respectively. GEBV correlations exceeded 0.98 for all the traits across all the chip types. Due to the genotype imputing step, DGV variances of embryos increased, on average, by 2% to 7% of total additive genetic variance and about half of the variance increase was attributed to Mendelian sampling. For embryos with lower call rates in original genotypes, the increase in DGV variances reached 12% of total genetic variance. The validation study showed that the continuous genomic evaluation system gave highly consistent results as the routine genomic evaluation. Further developments are needed to minimise differences in statistical methods between the two genomic evaluation systems.