[Selective structure of the gene pool. II. Standard method using F(ST) statistics].

A new approach to the description of the selective structure of the gene pool that is determined by the type and intensity of selection for individual genes is suggested. Selective pressure is defined based on the heterogeneity of the interpopulation diversity index FST. It is assumed that a significant difference between the estimated heterogeneity for the ith gene FST(i) and the selectively neutral interpopulation differentiation of the gene pool (Fe) indicate an effect of selection on the ith gene. In order to reveal such significant differences, empirical FST(i) distributions for the human gene pools of all parts of the world and five subregions of northeastern Eurasia were approximated by different theoretical distributions. It was demonstrated that only beta-distributions yielded a good approximation of the empirical FST(i) estimations in all studied gene pools. Based on the beta-approximations, confidence intervals for FST were calculated. These intervals allowed the genes to be divided into three classes of selective structure. Class NEUTRAL comprised genes that were assumed to be selectively neutral (FST(i) approximately equal to Fe); classes LOWER DIFF and SUPER DIFF comprised selectively important genes with significantly decreased (FST(i) < Fe) and increased (FST(i) > Fe) differentiation, respectively. Positions of 80 immunologically important biochemical markers were determined in the selective structure of the gene pools of six regions of the world: Europe, Asia, Africa, Australia, America, and northwestern Eurasia. The suggested method of estimating selective structure can be used if FST = Fe, where Fe is a selectively neutral variation of genes. Fe = 4NeMe + 1)-1; the main conditions for the use of this method are the following assumptions: (1) the genetic process is stationary and (2) the effective population size (Ne), the migration rate (Me), and the selection rate (S) are constant in time and space. If these conditions are not met, a correction (numeric resampling) is required.