Epsilon Open Archive

Abstract

A reaction norm describes the phenotype of an individual as a function of the environment. A reaction norm model is useful for describing traits that change gradually and continuously over an environmental gradient, for example temperature or feed quality. In this thesis the usefulness of reaction norm models for studying genotype by environment interaction (GxE) was evaluated. Field data of Nordic red dairy cattle were analysed using reaction norm models, with the purpose of describing the amount and pattern of GxE. Effects of various environmental variables on milk protein yield, days open and length of productive life were studied. The results showed that over the range of environments commonly encountered within Sweden and the Nordic countries, re-ranking of sires due to GxE was negligible. The slope of a linear reaction norm measures the sensitivity of an individual towards a change in the environment. A simulation study of environmental sensitivity showed that, in the presence of GxE, the average environmental sensitivity is expected to increase in a population selected for high phenotypic value in a continuously improving environment. As genetic variation was found for environmental sensitivity of milk protein yield, days open and length of productive life of dairy cattle it is possible to change environmental sensitivity of these traits with selection. Genetic evaluation of environmental sensitivity is possible using a reaction norm model. Response to mass selection of a trait affected by GxE was described using a linear reaction norm model combined with selection index theory. Prediction equations were derived for the genetic change in reaction norm coefficients and the average phenotypic value in any environment, depending on the environment in which the animals were evaluated and selected. A breeding program can be optimised by choosing the selection environment that gives the maximum genetic progress for a given breeding objective and environment where the animals are required to live. Equations were derived to predict this optimum selection environment.