Optimisation of recurrent selection under mutagenesis in micro-organisms

A general model is presented for the response of a haploid population reproducing asexually to recurrent cycles of mutagenesis and selection. It is concluded that the determinants of genetic progress in such populations differ from those in sexually reproducing populations in that (1) measurement errors are relatively unimportant in impeding progress, and (2) the proportion selected has a powerful effect on genetic variance in future generations. Dividing available testing resources between two selection lines, with two or three cycles of hybridisation and selection every five to ten cycles of mutagenesis can increase gains if numerous loci can potentially mutate to plus alleles, and if the variance induced by mutation is not very much less than the error variance. Programmes involving hybridisation become more attractive in the presence of a deleterious effect of mutagenesis on mean population value.Three conditions for the success of recurrent selection programmes based on mutagenesis are (1) the production by mutagenesis of a modicum of positive genetic changes, (2) the absence of strong effects of mutagenesis on the mean value of the population, and (3) the absence of strong genotype x test-environment interactions. The exact shape of the genetic distribution following mutagenesis, the total number of mutable loci and the presence of epistatic interactions between successive mutations are less critical.