Population genetics of trinucleotide repeat polymorphisms.

Trinucleotide repeats at five disease loci (DM, DRPLA, HD, SBMA and SCA1) were surveyed in phenotypically normal individuals from three continental populations. This is the first analysis to examine the population dynamics of these five disease-related trinucleotide repeats in the same individuals from worldwide populations. Roughly half of all alleles observed at each locus are shared between all continental groups. For three loci, disease prevalence in each population corresponds with the number of alleles in the upper tail of the allele-size distribution. The allele-size distributions of African, Asian and Caucasian groups show a high degree of variation, and gene diversity estimates for trinucleotide repeat loci exceed estimates derived from dinucleotide or tetranucleotide repeats. Analyses that compared infinite alleles and stepwise mutation models suggest that normal variation at trinucleotide loci is not generated by stepwise mutation alone. Trees constructed for subpopulations using trinucleotide repeat loci show accurate continental clustering. Interpopulation genetic distance estimates show remarkable similarity to distance estimates produced from tetranucleotide repeats or nuclear restriction site polymorphisms. This finding is especially noteworthy in light of the fact that trinucleotide repeat polymorphisms at these loci can cause disease, while restriction site and tetranucleotide polymorphisms appear to be selectively neutral. In contrast, genetic distance estimates from trinucleotide loci are poorly correlated with genetic distance estimates from mitochondrial sequence data.