Development of a genomic evaluation for milk production for a local bovine breed

The dual purpose Belgian Blue breed (DP-BBB) is a vulnerable breed rooted in the tradition of the Walloon Region of Belgium. The aim of this study was to investigate the potential in the development of a single step genomic evaluation (ssGBLUP) for DP-BBB milk production using a Bayesian procedure to integrate the Walloon estimated breeding values (EBV) as a priori known external information. This procedure combined genomic, pedigree and EBV by considering a correct propagation of all the information and no multiple considerations of contributions due to relationships. The Bayesian approach allowed to combine EBV and associated reliabilities (REL) without computation of deregressed proofs (DRP). 99 different combinations of genomic (a) and additive (1-a) relationships into a merged covariance structure for ssGBLUP were tested The optimal combination was chosen using EBV200s (available for 2,457 animals) from genetic evaluation based on milk production recorded until 31/12/2008. The pedigree file included 4,777 DP-BBB animals After usual editing, the SNP file contained 333 genotypes (from 155 cows and 178 bulls) of 34,531 useful SNP. In order to test the method, DRP2012 were computed from EBV from genetic evaluation based on milk production recorded until 31/12/2012 and associated REL for 2,462 animals. DRP2012 of 32 bulls (with EBV2008 equal to their pedigree index in 2008, with 10 EDC in 2012 and born after 2001) were regressed on genomically enhanced breeding values (GEBV2008) obtained from each of the 99 ssGBLUP or on EBV20 with REL2012 as a weight. The a of 0.78 gave the best determination coefficient (0.494). while the determination coefficient for the regression DRP2012 on EBV 200s was 0.47. Thereby, these results showed the feasibility of a modified ssGLUP for a small breed.