Human evolution and migration: inferences from autosomal, Y chromosome, and mitochondrial data
We report an analysis of genetic variation in 317 individuals from 21 populations in sub-Saharan Africa, Asia, Europe, and the Indian subcontinent. We have assayed mitochondrial DNA polymorphisms (HVS1 and non-control region SNPs), 51 Y chromosome SNPs, 30 restriction site polymorphisms, 60 short tandem repeat polymorphisms, 100 Alu insertion polymorphisms, and 100 L1 insertion polymorphisms in these individuals. This large number of genotypes permits robust inferences about migration patterns and population affinities. There is remarkable consistency in the results for autosomal systems: Fst values range from 10 to 15%, and Africans have the highest level of genetic diversity. We used genetic distance trees and principal components analysis to study the relationships between populations. All data sets show that populations from the same continent cluster together, with the largest genetic distances occurring between the African and non-African populations. A series of south Indian populations are intermediate between the European and East Asian populations, and the upper-caste Brahmin population has the smallest distance to the European cluster. The roots of the L1, Alu, and RSP trees are both located within the African population cluster, supporting an African origin of modern humans (however, we discuss reasons why the root of such a tree can be misleading). We show that these polymorphisms can infer the continental ancestry of individuals (African, Asian, or European) with 90-100% accuracy, and we discuss the implications of these findings for current concepts of “race.” Finall, we show that Y and mtDNA data are very useful for inferring migration patterns, but, because each is only a single locus, more robust and statistically supportable inferences can be made when large batteries of autosomal polymorphisms are also used.