Statistical analyses of Drosophila and human protein polymorphisms

By using the distribution function of allelic frequencies which was recently derived by Kimura and Ota for the model of stepwise production of neutral alleles, the observed protein polymorphisms of Drosophila and man are tested for fit to the theory of neutral protein variation. The observed and theoretical distributions of alleles agree quite well except for the excess of rare alleles in the actual distributions. In human polymorphisms, the alleles with frequencies less than 1% are more numerous than expected, whereas in Drosophila, those with frequencies less than 10% are more numerous. It is pointed out that these results support my thesis that mutational pressure rather than balancing selection is the main cause for the maintenance of protein polymorphisms.