Walloon single-step genomic evaluation system integrating local and MACE EBV

Abstract

Walloon dairy cattle could be considered as a small scale population where the majority of AI bulls are imported from several foreign countries. Single-step Genomic Best Linear Unbiased Prediction (ssGBLUP) methods allow the simultaneous use of genomic, pedigree and phenotypic information and should reduce potential biases in the estimation of Genomically Enhanced Breeding Values (GEBV). Therefore, in the context of developing a Walloon genomic evaluation system, it was considered as the best option. However, in opposition to multi-step genomic predictions, ssGBLUP only uses local phenotypic information and is unable to use directly important other sources of information coming from abroad, e.g. Multiple Across Country Evaluation (MACE) results provided by Interbull. Therefore, single-step Genomic Bayesian Prediction (ssGBayes) was used as an alternative method for the Walloon genomic evaluation system. The ssGBayes approach allows combining simultaneously all available genotype, pedigree, local and foreign information in a local evaluation by considering a correct propagation of external information avoiding double counting of contributions due to relationships and due to records. In the Walloon genomic evaluation system, local information refers to Walloon EBV and associated reliabilities (REL) and foreign information refers to MACE EBV and associated REL. Furthermore, the Bayesian approach has the advantage to directly combine EBV and REL without any deregression step. The ssGBayes method computed more accurate predictions for all types of animals. For example, for genotyped animals with low Walloon REL (< 0.25) without MACE results and sired by genotyped bulls with MACE results, the average increase of REL for the studied traits was 0.39 points of which 0.14 points could be traced to the inclusion of MACE information. For other categories of genotyped animals, the contribution by MACE information was high too. The new Walloon genomic evaluation system passed the Interbull GEBV tests for several traits in July 2013. This approach has the potential to improve current genomic prediction strategies as it can be used in other settings where the combination of different sources of information is required.