A back migration from Asia to sub-Saharan Africa is supported by high-resolution analysis of human Y-chromosome haplotypes

Abstract

The variation of 77 biallelic sites located in the nonrecombining portion of the Y chromosome was examined in 608 male subjects from 22 African populations. This survey revealed a total of 37 binary haplotypes, which were combined with microsatellite polymorphism data to evaluate internal diversities and to estimate coalescence ages of the binary haplotypes. The majority of binary haplotypes showed a nonuniform distribution across the continent. Analysis of molecular variance detected a high level of interpopulation diversity (PhiST=0.342), which appears to be partially related to the geography (PhiCT=0.230). In sub-Saharan Africa, the recent spread of a set of haplotypes partially erased pre-existing diversity, but a high level of population (PhiST=0.332) and geographic (PhiCT=0.179) structuring persists. Correspondence analysis shows that three main clusters of populations can be identified: northern, eastern, and sub-Saharan Africans. Among the latter, the Khoisan, the Pygmies, and the northern Cameroonians are clearly distinct from a tight cluster formed by the Niger-Congo-speaking populations from western, central western, and southern Africa. Phylogeographic analyses suggest that a large component of the present Khoisan gene pool is eastern African in origin and that Asia was the source of a back migration to sub-Saharan Africa. Haplogroup IX Y chromosomes appear to have been involved in such a migration, the traces of which can now be observed mostly in northern Cameroon.

Figure 1

Maximum-parsimony phylogeny of NRY biallelic markers. Markers are indicated on the branches. Italicized markers were not typed in the present study, but their allelic state could be deduced unequivocally from their position in the genealogy given by Underhill et al. (2001b). Group IX markers defining the 117b derivative haplotypes 110–112, 114–116, and 118 (Underhill et al. ; P.A.U., unpublished data) have not been typed in the present study. Haplogroup and haplotype designations are according to Underhill et al. (2001b), with some modifications (see the “Results” section). Only the haplotypes observed in the present study are numbered. Underlined haplotypes are those described here for the first time. Phylogenetic relationships between group IX markers M207 and M173 (Karafet et al. 2001) and M17 and SRY₁₀₈₃₁ (Weale et al. 2001) have recently been resolved.

Figure 2

Distribution of major Y-chromosome haplotypes in 11 geographic areas of Africa. Some evolutionarily related haplotypes were pooled according to the simplified phylogeny shown at the bottom of the figure. The 12 most common haplotypes or groups of related haplotypes are represented by different colors. The remaining haplotypes are represented in the white sector of the relevant pie charts. Geographic groups are as follows: 1, Morocco; 2, Mali; 3, Sudan; 4, Ethiopia; 5, Burkina Faso; 6, northern Cameroon; 7, southern Cameroon; 8, Central African Republic (Biaka Pygmies); 9, Democratic Republic of Congo (Mbuti Pygmies); 10, southern Africa Bantu; 11, southern Africa Khoisan. Data for Morocco, Burkina Faso, Cameroon, Central African Republic, and Democratic Republic of Congo are from the present study; the frequencies for Ethiopia and the southern Africa Khoisan have been obtained by pooling data from the present study and the study by Underhill et al. (2000); data for Mali, Sudan, and southern Africa Bantu are from Underhill et al. (2000).

Figure 3

Correspondence analysis scores; plot of populations and haplotypes in the space of the first (X-axis) and second (Y-axis) dimension. Only the haplotypes (represented by asterisks [*]) most contributing to the inertia of the first and/or second dimension are represented. A, 31 African populations, including the 22 populations listed in table 1, 5 populations from Underhill et al. (2000), and 4 Moroccan populations from Bosch et al. (2001). Symbols used to identify groups of populations are as follows: diamonds, sub-Saharan Africans; circles, northwestern Africans; squares, eastern Africans. B, Correspondence analysis for the 20 African populations having negative scores for the first dimension in the previous analysis. The following symbols have been used to identify groups of populations: triangles, Khoisan; circles, Pygmies; squares, northern Cameroonians; crosses, other sub-Saharan Africans (populations from Burkina Faso and Bantu-speaking populations from southern Cameroon and southern Africa).

Figure 4

MJ microsatellite networks of six different binary haplotypes. Microsatellite haplotypes are represented by circles, with areas proportional to the number of individuals harboring the haplotype. A, haplotype 2 (M13); B, haplotype 15 (M112); C, haplotype 22 (M191); D, haplotype 24 (DYS271); E, haplotype 35 (M35); and F, haplotype 41 (M85). For each network, the smallest circles represent a count of one individual, with the exception of the network depicted in D (haplotype 24), where only haplotypes having an absolute count of two or more are represented. Branch lengths are proportional to the number of one-repeat mutations separating two haplotypes. Branch length of the network B (haplotype 15) is one-half that in the other networks. Network B also includes three haplotype 15 chromosomes reported by Underhill et al. (2000). Network E (haplotype 35) also includes 16 haplotype 35 chromosomes from southern Ethiopia (R.S., unpublished data). Pygmy chromosomes in network F also carry the M200 derived allele (haplotype 40).