Evolutionary relationships of human populations on a global scale

Using gene frequency data for 29 polymorphic loci (121 alleles), weconducted a phylogenetic analysis of 26 representative populations fromaround the world by using the neighbor-joining (NJ) method. We alsoconducted a separate analysis of 15 populations by using data for 33polymorphic loci. These analyses have shown that the first major split ofthe phylogenetic tree separates Africans from non-Africans and that thissplit occurs with a 100% bootstrap probability. The second split separatesCaucasian populations from all other non-African populations, and thissplit is also supported by bootstrap tests. The third major split occursbetween Native American populations and the Greater Asians that includeEast Asians (mongoloids), Pacific Islanders, and Australopapuans (nativeAustralians and Papua New Guineans), but Australopapuans are geneticallyquite different from the rest of the Greater Asians. The second and thirdlevels of population splitting are quite different from those of thephylogenetic tree obtained by Cavalli- Sforza et al. (1988), whereCaucasians, Northeast Asians, and Ameridians from the Northeurasiansupercluster and the rest of non- Africans form the Southeast Asiansupercluster. One of the major factors that caused the difference betweenthe two trees is that Cavalli-Sforza et al. used unweighted pair-groupmethod with arithmetic mean (UPGMA) in phylogenetic inference, whereas weused the NJ method in which evolutionary rate is allowed to vary amongdifferent populations. Bootstrap tests have shown that the UPGMA treereceives poor statistical support whereas the NJ tree is well supported.Implications that the phylogenetic tree obtained has on the currentcontroversy over the out-of-Africa and the multiregional theories of humanorigins are discussed.