Analysis of non-additive genetic effects in Norway spruce

Genetic variance includes both additive and non-additive components, and the latter can be partitioned into dominance and epistatic variance. Additive genetic variance has been the main source of genetic improvement in most tree breeding programs and the genetic gain achieved has mostly been deployed through open-pollinated seed orchards. However, non-additive effects could be exploited in alternative deployment programs and provide additional genetic gain. Thus, knowledge of non-additive effects is essential for constructing robust genetic models and evaluating vegetative deployment systems, particularly for clonal forestry. In this study, we used data drawn from seven Norway spruce full-sib and three half-sib clonal trials to test and compare different models partitioning genetic effects into additive, dominance, and epistatic components for height and diameter. Simpler models provided more robust estimates of additive and non-additive effects, whereas more complex models provided more detail and insight in performance at single sites and G × E patterns, but with less reliable estimates. In this study, we cannot clearly distinguish the size and importance of dominance and epistatic effects. However, the total non-additive effects seem to be substantial, being around 80–90% of the additive, and were of the magnitude required to consider clonal forestry as an alternative to current deployment systems. Additive correlations largely followed the expected patterns based on seed zones and late spring frost incidence. Clonal correlations were generally lower than the additive implying that G × E interactions were stronger for non-additive effects.