Are tree breeders properly predicting genetic gain? A case study involving Corymbia species

Abstract

The estimation of quantitative genetic parameters in breeding programs is important to ensure efficient selection. In this context, knowledge of the mating system is critical, as it underpins assumptions about inter-relatedness on which variance component estimation depends. However, proper account of the breeding system is not always taken, either because it is unknown and/or because it is ignored. That eucalypts have a mixed-mating system is well-established, however many breeders use models that assume outcrossed mating with an infinite number of male parents (i.e. allogamous mating), from which genetic parameter estimates are then used to predict genetic gains. First-generation, open-pollinated progeny tests of Corymbia citriodora subsp. citriodora and C. citriodora subsp. variegata, being managed for seed production, were used to investigate the likely bias, resulting in overestimation of genetic parameters under the assumption of allogamous mating. When we assumed allogamous mating, we observed inflated predictions of additive variance and narrow-sense heritability ($\hat{h}_{a}^{2}$) of diameter at breast height and height. The overestimate of $\hat{h}_{a}^{2}$ was approximately 32% for C. citriodora subsp. citriodora and 21% for C. citriodora subsp. variegata. Inappropriate modelling of relatedness in eucalypts that assumes panmictic outcrossing when in fact these species have a mixed-mating system results in overestimates of the population genetic gain with selection.