Simple Statistical Tools to Detect Signals of Recent Polygenic Selection

Abstract

A growing body of evidence shows that most psychological traits are polygenic, that is they involve the action of many genes with small effects. However, the study of selection has disproportionately been on one or a few genes and their associated sweep signals (rapid and large changes in frequency). If our goal is to study the evolution of psychological variables, such as intelligence, we need a model that explains the evolution of phenotypes governed by many common genetic variants. This study illustrates simple statistical tools to detect signals of recent polygenic selection: a) ANOVA can be used to reveal significant deviation from random distribution of allele frequencies across racial groups. b) Principal component analysis can be used as a tool for finding a factor that represents the strength of recent selection on a phenotype and the underlying genetic variation. c) Method of correlated vectors: the correlation between genetic frequencies and the average phenotypes of different populations is computed; then, the resulting correlation coefficients are correlated with the corresponding alleles’ genome-wide significance. A significant difference between the allele frequencies for the three races was found. Post-hoc test revealed that East Asians had significantly higher frequencies of IQ increasing alleles than Africans. In contrast, the distribution of height increasing alleles did not differ among races. The second prediction, that alleles with large frequency differences between populations had a higher genome-wide significance value than alleles with small frequency differences, was confirmed by the analysis of the Pygmy vs non Pygmy data set.