Dynamics of polygenic characters under selection (Q1813967)

The evolution of quantitative characters influenced by genetic variation at many loci is usually analyzed by statistical methods. In this paper, an extensive analysis of selection, recombination and genetic drift is presented which is based on multi-locus population theory. The basic model used by the authors is an \(n\)-loci system with purely additive genetic effects, i.e. the character \(Z\) which is determined by the loci under consideration is given by \(Z=\sum(X_ i+X_ i^*)+E\), where \(X_ i\) and \(X_ i^*\) denote the contributions of parental alleles at locus \(i\) and \(E\) denotes a random environmental effect. Further, the model assumes random mating in a large population, discrete generations, viability selection and a distribution of environmental effects that is independent of genotypes. No (Gaussian) assumptions, however, are made about the distribution of \(Z\) or the number of loci or alleles. Recursions for the mean and higher-order moments are derived generalizing the standard results which are based on statistical arguments. It is shown by analytical as well as numerical methods, that even for weak selection the population mean can depart substantially from the predictions of the standard recursions, at least in the long run.