Marker assisted selection in dairy cattle using a mixed model approach

A mixed model method combining marker and phenotypic information was developed for genetic evaluation. Marker effects were treated as uncorrelated random effects, providing a convenient framework for marker assisted selection using best linear unbiased prediction (BLUP). It is computationally efficient because only a small number of additional equations are needed to include marker effects in the model. The total genetic merit of QTL and polygenes are obtained as the sum of the predicted marker and polygenic effects. A maximum likelihood method was developed to estimate each marker component required for the BLUP procedure. Computer simulations showed that MAS using the proposed approach may significantly improve selection accuracy, particularly in situations when traditional phenotypic selection is least effective, i.e., when heritability is low and the percentage of selected individuals is high.