Non-additive genetic effects in animal selection

Genetic evaluation of purebred farm animals has been carried out for about half a century, employing additive approximation to describe the genetic background. An evaluated animal has been attributed a single breeding value for each trait of the breeding goal. The predicted additive genetic value of an animal equals the average breeding value of its parents. Although the selection based on the additive approach has proved successful, there still is a possibility of increasing the reliability of the breeding value estimation by accounting for non-additive genetic effects of dominance and epistasis, disregarded in the additive model. In the non-additive model, the expected quality of the progeny equals the average of the parents plus an effect resulting from the interaction between the parents. In this case, the evaluated animal may have as many breeding values as there are possible candidates to mate to, for each trait. The dominance and epistatic effects have already been accounted for in selecting animals or populations for some crossbreeding plans (combining ability, heterosis, and recombination loss). Also, using crossbreds for the sake of the breeding value estimation of purebred animals requires removing the non-additive effects from the crossbred performance and distributing the additive component between the purebreds. Combining ability is more and more discussed as a factor for matings within breed to produce terminal progeny.