High resolution of human evolutionary trees with polymorphic microsatellites

GENETIC variation at hypervariable loci is being used extensively for linkage analysis¹ and individual identification², and may be useful for inter-population studies^2-5. Here we show that polymorphic microsatellites (primarily CA repeats) allow trees of human individuals to be constructed that reflect their geographic origin with remarkable accuracy. This is achieved by the analysis of a large number of loci for each individual, in spite of the small variations in allele frequencies existing between populations^6,7. Reliable evolutionary relationships could also be established in comparisons among human populations but not among great ape species, probably because of constraints on allele length variation. Among human populations, diversity of microsatellites is highest in Africa, which is in contrast to other nuclear markers and supports the hypothesis of an African origin for humans.