Accuracy of genomic estimated breeding values for crossbred performance in broilers using a purebred or crossbred reference population

Duenk, P.; Calus, M.P.L.; Wientjes, Y.C.J.; Breen, Vivian; Henshall, John; Hawken, Rachel; Bijma, P.


In breeding programs of pigs and poultry, selection is typically applied at the level of the purebred (PB) animals using PB data. This may result in a suboptimal response at the crossbred (CB) level when the genetic correlation between PB and CB performance () is lower than one. The objectives of this study, therefore, were to (1) estimate the , and (2) compare the accuracy of genomic estimated breeding values (GEBVs) for CB body weight in broilers using a PB reference population (RP) or a CB RP. Phenotype and genotype data were available for 5,274 PB and 10,395 CB offspring, and genotype data for their 163 PB sires. For comparison of cross-validation (CV) GEBV accuracies, the size of the CB RP was reduced to approximately match the size of the PB RP. Because the number of sires was
limited, validation was also performed on GEBVs of CB offspring of the sires. The estimated was 0.94 (0.04). The results show that the accuracy of sire GEBVs was slightly higher with a CB RP (0.43) compared to a PB RP (0.40). Similarly, the accuracy of CB offspring GEBVs was higher with a CB RP (0.47) compared to a PB RP (0.25). These results are likely due to (1) the that is lower than one, and (2) differences in genomic relationships between reference and validation animals. In finalizing this work, we will develop an equation that uses the accuracy of CB GEBVs for CB performance to predict the accuracy of PB GEBVs for CB