Know more

Genomic evaluation predicts the genetic value of breeding candidates using three types of information: performances (or phentoypes), geneologies and genetic types. Since 2009, genomic evaluation has been done in two steps: an analysis of all animals, limited to performances and geneologies; then a genomic analysis limited to all typed animals. For 12 years, this two-step procedure has shown its efficiency but it suffers from a more or less marked bias called "selection bias", leading to the underestimation of the genetic progress realized and the underestimation of the value of reproducing animals when the phenotypic information of their offspring arrived.

With this in mind, since 2017, the UMT eBIS has prepared the conditions for the deployment of a one-step evaulation, or Single Step. Although this method has been proposed for over ten years, it is only now being deployed on an international level in very large populations. First, the HSSGBLUP software (for Hybrid Single Step Genomic Best Linear Unbiased Prediction) was developed at INRAE. It is intended to provide the means to cover the needs for both the dairy and beef sectors, and even beyond just cattle. As part of the ASAP project financed by Apis-Gene, the UMT validated this software over a large array of conditions that exist on the field, estimated the gain in precision and the reduction of the associated bias, and analyzed the impact of the change of the method on the indexes. Finally, as part of the UNIGENO project financed by CASDAR, GenEval and Idele deployed this new method, with the help and experise of the UMT: the extension to all breeds and traits, integration of specific changes in the associated models, and concertation with all the users (selection organizations and companies).

This long project resulted in the diffusion of new indexes on March 8, 2022 for national dairy breeds, in April for the international Holstein and Brown breeds, whereas the indexes for beef cattle are programmed for the beginning of 2023.