[Interpopulation diversity of the gene pool: beta distribution of Wright's F(ST) statistics].

The distribution of FST(i) estimates were studied in representative samples of ith genes for the gene pools of the major human populations, including the populations of Europe, Asia, Africa, Australia, America, North-eastern Eurasia, and five subregions of the latter. An average of 80 FST(i) estimates were analyzed for each sample of marker genes with a level of polymorphism (q) from 0.05 to 0.95. For each gene pool, the empirical distributions of FST(i) estimates were approximated by the main types of theoretical distributions--the normal, chi 2, Weibull, gamma, and beta distributions. In all gene pools, only beta distributions were good approximations of the empirical FST(i) distributions. The parameters of the beta distributions are reported in this study. It was demonstrated that the characteristics of beta distributions for all major gene pools studied could be interpolated to the smaller constituent gene pools. since the use of the traditional parametric tests starts from an assumption of normal distribution, they are inapplicable to analysis of FST statistics. Therefore, the obtained parameters of beta distributions should be used. These parameters allow the confidence intervals of the average FST values to be determined and permit correct comparison between the characteristics of both individual genes and gene pools to be performed.