The apportionment of dinucleotide repeat diversity in Native Americans and Europeans: a new approach to measuring gene identity reveals asymmetric patterns of divergence.

The purpose of this paper is to assess the extent of gene identity and differentiation at 33 dinucleotide repeat loci (377 total alleles) within and among three European and three Native American populations. In order to do this, we show that a maximum-likelihood method proposed for phylogenetic trees (Cavalli-Sforza and Piazza 1975) can be used to estimate gene identity (Nei 1987) with respect to any hierarchical structure. This method allows gene differentiation to be evaluated with respect to any internal node of a hierarchy. It also allows a generalization of F- and G-statistics to situations with unequal expected levels of differentiation. Our principal finding is that levels of genetic differentiation are unique to specific populations and levels of nesting. The populations of European origin show very little internal differentiation; moreover, their continental average is close to the total population defined by the aggregate of Europeans and Native Americans. By contrast, the Native American populations show moderate levels of internal differentiation, and a great distance between their continental average and the total. The results of analyses of subsets of loci that were selected to have high gene diversities in either Europeans or Native Americans closely parallel those from the total set of loci. This suggests that the principal results are unlikely to be caused by a European ascertainment bias in locus selection. In summary, our findings demonstrate that partitions of gene diversity into within- and between-populations components are heavily biased by the populations analyzed and the models fitted. Optimistically, however, more information is available to analyze population history and evolution by quantifying, as we have done, the uniqueness of patterns of differentiation.