Abstract:

We compared the effects of using three genomic relationship matrices and a pedigree based relationship matrix for optimum contribution selection (OCS) with a constraint on number of sires selected, on genetic gain, kinship and additive genetic variance. This was done by simulating a genomically selected small dairy cattle population. The cattle population underwent genomic selection for 11 generations. In total, 6,000 females and 2,500 males were genotyped per generation. We compared a pedigree-based matrix as well as VanRaden’s method 1 to construct genomic relationship matrices using either allele frequencies of the base population, of the current generation, or allele frequencies among all genotyped animals. The pedigree-based OCS resulted in higher kinship and more loss of genetic variance than using genomic OCS based on base population frequencies. Thus, genomic OCS with genomic relationship matrices using base population frequencies are preferable compared with pedigree-based OCS.