DISTRIBUTION OF GENE FREQUENCY AS A TEST OF THE THEORY OF THE SELECTIVE NEUTRALITY OF POLYMORPHISMS

TLDR

Gene frequency variation across populations or generations arises from breeding structure and selection, and breeding structure should affect all loci equally. Heterogeneity in loci‑specific inbreeding coefficients F = sp²/p(1‑p) can be used as evidence of selection. Statistical analysis of F shows significant heterogeneity among human polymorphic genes worldwide, indicating natural selection, while temporal data from Dacus oleae reveal no evidence of selection.

Abstract

ABSTRACT The variation in gene frequency among populations or between generations within a population is a result of breeding structure and selection. But breeding structure should affect all loci and alleles in the same way. If there is significant heterogeneity between loci in their apparent inbreeding coefficients F=sp2/p(1-p), this heterogeneity may be taken as evidence for selection. We have given the statistical properties of F and shown how tests of heterogeneity can be made. Using data from human populations we have shown highly significant heterogeneity in F values for human polymorphic genes over the world, thus demonstrating that a significant fraction of human polymorphisms owe their current gene frequencies to the action of natural selection. We have also applied the method to temporal variation within a population for data on Dacus oleae and have found no significant evidence of selection.