The role of finite population size and linkage in response to continued truncation selection : II. Dominance and overdominance.

In a direct Monte Carlo study, the effects and interactions of parental population size, close linkages between loci, degree of truncation selection and environmental variance are explored in the cases of complete dominance and overdominance with equilibrium gene frequency of .5. The base population simulated has a gene frequency of .5 at all 40 loci and is a state of gametic equilibrium. Response and fixation of genes are studied up to 30 generations.In the case of complete dominance of desirable genes, long term response is described in terms of an initial phase of change and a later phase when there is a very slow approach to the limit and plateaus are established in the case of close linkages between loci. The changes in the initial phase consitutes most of the total response. Increased selection in parental populations of 16 or less is ineffective when recombination value between adjacent loci is as low as .005. Linkage effects in reducing response appear to increase in geometric proportions with restricted population size.Results of the overdominance model provide an interesting study of the effects of population size and linkage on heterozygosity while there is selection for the heterozygotes. These effects are partially explainable in terms of differences in heterozygosity between chromosomes and heterozygosity within chromosomes. The between chromosome component of heterozygosity appears to be rapidly reduced with finite population size and common heterozygosity within chromosomes is conserved due to linkages between loci. In general, however, the effects of selection pressure, linkage and population size interact strongly. With small populations and tight linkage, increased selection pressure causes a decrease in amount of fixation. With the largest population increased selection pressure tends to increase the amount of fixation. It is clear, however, that these conclusions must be regarded tentatively, and an exact picture of what is happening will be obtained only with very extensive simulation. It seems quite clear, also, that single locus finite population theory is essentially ineffective.