342. Breed-origin-of-alleles approach using summary statistics for multi-breed genomic prediction in dairy cattle

Systematic crossbreeding strategies between dairy cattle breeds in dairy herds are becoming more and more attractive to farmers, and this leads to a request for genomically enhanced breeding values for crossbred females in the dairy herds. Accurate genomic prediction of crossbred animals can be achieved if the genotypic and phenotypic data of the breeds involved in the crossbreeding are available to form reference populations, to estimate marker effects. However, sharing genotype and phenotype data between breed populations may be an issue due to privacy and competition. This study investigated genomic prediction of two-breed and three-breed rotational crossbred dairy cattle using summary statistics and a breed-origin of alleles model. The results indicate that the approach can yield almost as high prediction accuracies as having full information from the pure breeds.