Integration of single step DGV in conventional genetic evaluations using DGV-PBLUP

Abstract

Routine single-step genomic evaluations can be costly in time and computer resources. Hence, newly genotyped animals initially receive a genomic prediction of their direct genomic values (DGV). If genomic predictions of DGVs of animals become available it may be convenient to estimate GEBV of such animals using some form of integration into conventional pedigree BLUP evaluations. DGV-PBLUP is a novel method of integration of DGV from genomic predictions, into a conventional pedigree BLUP (PBLUP) evaluation. This is done by setting the prior mean of the animal genetic effect (which usually is zero in linear mixed models) to the DGV to be incorporated. In this paper we report on the application of this methodology to the Dutch-Flemish genetic evaluation. Results showed a high correlation (0.99 or higher) between GEBV of animals associated with a genotype in single step SNP-BLUP and a GEBV in DGV-PBLUP. Run time of DGV-PBLUP evaluations were comparable to conventional pedigree evaluations and much shorter than routine single-step SNP-BLUP evaluations. DGV-PBLUP promises to be a convenient method of integration of genomic information into pedigree BLUP evaluations, without the need for sharing or accessing SNP genotypes.