Forging links between population and quantitative genetics

Abstract

Summary An initially rare allele with a large effect on a quantitative character is expected to exhibit the following behaviour in artificial selection lines: 1. It should change in frequency, or be lost by chance, at rates predictable from the effects of the allele on the quantitative character and the selection regime imposed. 2. At the phenotypic level the behaviour of the allele should cause (a) asymmetrical responses to bidirectional selection, (b) variation among replicate lines in response to selection corresponding to the behaviour of the allele in individual lines, (c) changes in heritability in lines in which the allele increases in frequency and (d) selection response compatible with the effects of the allele. This paper reports an experimental evaluation of these predictions utilizing a rare allele of large effect (sm ^lab) detected in a sample of the Canberra outbred population of Drosophila melanogaster at a frequency of 1/120. Homosygosity for this allele reduced abdominal bristle number by more than 50%, altered the abdominal bristle pattern and reversed the sexual dimorphism for abdominal bristle number. Experiments were done to characterise sm ^lab and all evidence indicates that it is a single allele with a very large effect. Bidirectional selection for abdominal bristle number was carried out in three high and three low lines from this sample of the Canberra population. The sm ^lab allele rose in frequency and went to fixation in two of the low lines (in 10 generations) but was lost from the third. These times to fixation were slower than the expectations derived from computer simulations of the behaviour of such an allele but this can be attributed to the lower fitness of sm ^lab homozygotes. The proportions of lines with the allele fixed or lost were compatible with expectations. At the phenotypic level, the behavior of sm ^lab had the expected consequences, namely, (i) asymmetrical responses to bidirectional selection, (ii) variation in response among replicate low lines corresponding to the behaviour of sm ^lab, (iii) changes in heritabilities in the lines in which sm ^lab went to fixation, and (iv) selection responses compatible with the effects of the allele. A test for rare alleles of large effect was proposed, based on the expected pattern of change in heritability under artificial selection. This test was applied to the high selection lines but no evidence was found for important effects due to rare alleles of large effect increasing abdominal bristle number, a conclusion consistent with other independent evidence. This work provides experimental corroboration of the links between population genetics and quantitative genetics.