Theoretical study of near neutrality. I. Heterozygosity and rate of mutant substitution

In order to clarify the nature of 'near neutrality' in molecular evolution and polymorphism, extensive simulation studies were performed. Selection coefficients of new mutations are assumed to be small so that both random genetic drift and selection contribute to determining the behavior of mutants. The model also incorporates normally distributed spatial fluctuation of selection coefficients. If the system starts from 'average neutrality', it will move to a better adapted state, and most new mutations will become 'slightly deleterious'. Monte Carlo simulations have indicated that such adaptation is attained, but that the rate of such 'progress' is very low for weak selection. In general, the larger the population size, the more effective the selection becomes. Also, as selection becomes weaker, the behavior of the mutants approaches that of completely neutral genes. Thus, the weaker the selection, the smaller is the effect of population size on mutant dynamics. Increase of heterozygosity with population size is very pronounced for subdivided populations. The significance of these results is discussed in relation to various observed facts on molecular evolution and polymorphism, such as generation-time dependency and overdispersion of the molecular clock, or contrasting patterns of DNA and protein polymorphism among some closely related species.