Methods to optimize livestock breeding programs with genotype by environment interaction and genetic heterogeneity of environmental variance

Genotype by environment interaction (G × E) and genetic heterogeneity of environmental variance are both related to genetic variation in environmental sensitivity. Both phenomena can have consequences for livestock breeding programs. This thesis focuses on developing methods to optimize livestock breeding programs with G × E and genetic heterogeneity of environmental variance.The first part of this thesis deals with G × E, which means that different genotypes respond differently to environmental changes. G × E causes reranking of genotypes when the genetic correlation between performances in different environments is smaller than one. G × E can exist, for example, between a selection environment (SLE) and a production environment (PDE). Recording of half-sibs or progeny in the PDE limits the loss in genetic gain. Due to the lower loss in genetic gain, progeny testing schemes rather than sib testing schemes are recommended when the genetic correlation between SLE and PDE is lower than 0.7 - 0.8. In dairy cattle, G × E limits the possibilities for cooperation of breeding programs operating in different environments. Long-term cooperation in the presence of G × E is possible when the genetic correlation is higher than 0.8 - 0.9. When breeding dairy cattle for two environments in the presence of G × E, it is optimal to run a single breeding program with progeny testing bulls in both environments when the genetic correlation is higher than 0.7 - 0.8. When the genetic correlation is lower than 0.7 - 0.8, two environment-specific breeding programs are optimal.The second part of this thesis deals with genetic heterogeneity of environmental variance. Genetic heterogeneity of environmental variance means that genotypes differ in the magnitude of the environmental variance, for example, due to differences in environmental sensitivity. A multiple regression framework was developed to predict selection responses in mean and environmental variance. Although environmental variance can be considered as a trait with a low heritability, responses in environmental variance can be greater than 10% of the current mean environmental variance in some cases. Genetic heterogeneity of environmental variance can be exploited in livestock breeding programs to increase uniformity of animals by reducing environmental variance. This is especially of interest for traits where the current population mean is near an optimum value. Another application is to breed animals that are more robust against unpredictable environmental fluctuations.