Optimizing management of non-additive genetic variation in selective breeding programmes

Intense selection, particularly in nucleus breeding programmes, has been hugely successful in achieving genetic gain and improving commercially important production and health traits. Alongside genetic drift, selection leads to changes in allele frequencies over time, and loss of genetic variance and heterozygosity, which give rise to inbreeding depression. Breeding programmes in livestock are typically based solely on additive genetic architecture, while non-additive genetic effects (e.g., dominance) are overlooked. The aim of this study is to explore methods to estimate quantitative genetic parameters and assess the impact of dominance variance on the breeding programme over time, in the presence of both additive and dominance genetic effects. We have employed stochastic simulations within the AlphaSimR package, to examine scenarios for a complex trait influenced by only additive or by both additive and dominance genetic effects, in populations of varying numbers of breeding animals, and assuming random mating and equal contributions. An initial burn-in of 10 generations was followed by 20 generations of truncation selection based on phenotypes or on genomic estimated breeding values (GEBV).<br/>We estimated GEBV using either an additive or an additive and dominance model via ridge-regression best linear unbiased prediction (RR-BLUP / SNP-BLUP). The temporal trends and long-term performance of the simulated breeding programmes were assessed in terms of genetic gain, changes in genetic variance, inbreeding depression, and prediction accuracy. We expect that accurately estimating and directly accounting for and exploiting dominance effects can enhance the long-term performance of the breeding programme and better inform the management of genetic variance and inbreeding depression. Our findings will support further work to optimize the trade-off between genetic gain and loss of genetic diversity, and to devise breeding strategies for how to best estimate and exploit non-additive variation in practice in commercial breeding programmes.