Peter Sørensen

Genetic variance for complex traits in animal breeding are often estimated using linear mixed-models that incorporate information from SNP-markers using a realized genomic-relationship matrices. In these models, individual genetic markers are weighted equally and the variation in the genome is treated as a “black box”. While this approach has proved useful in selecting animals with high genetic potential, it does not generate insight into the biological mechanisms underlying trait variation. We propose to build a linear mixed model approach to evaluate the collective effects of sets of SNPs in genomic features and open the “black box”. Using data on ADG and BF from 6,112 entire Duroc boars and a high-density SNP chip, we show here, that the QTL categories with highest relative importance of the SNP set were indeed biological meaningful