Introduction of single-step genomic evaluations in German Holsteins

Abstract

In April 2025, we introduced single-step genomic evaluations for all traits that are subject to routine genetic evaluations in German Holstein. With all models, we estimate the same main effects as with the former conventional genetic evaluations of Holsteins. In addition, a fixed regression on the inbreeding coefficient was added to all the models. With the introduction of single-step models, reliabilities of GEBVs increased for all traits. This increase is especially pronounced in young animals with no own or offspring performance. As expected, the increase in reliability was greatest for many functional traits: longevity and direct calving index: +0.11, maternal calving index: +0.13, young stock survival and health index: +0.14, while it was lower for production traits: production index: +0.04. Additionally, validations with 2 and 4 years of right-truncated data confirm a substantial increase in the predictive ability of genomic GEBVs compared to the previous multi-step model: correlations of purely genomic GEBVs of young bulls with their later daughter-proven GEBVs are higher for all traits with the single-step model. Again, this increase in predictive ability is highest for the functional traits and lower for the production traits. With publication dates in April, August, and December, we conduct main runs with updated phenotypic information three times a year. In these main runs, we include MACE information from the respective current Interbull MACE run. In addition to these full runs, we conduct weekly genomic evaluations, for which we use the estimates of the SNP-effects and the residual polygenic effects from the main runs and apply them to the newly genotyped animals.