Haplotypes in the dystrophin DNA segment point to a mosaic origin of modern human diversity

Abstract

Although Africa has played a central role in human evolutionary history, certain studies have suggested that not all contemporary human genetic diversity is of recent African origin. We investigated 35 simple polymorphic sites and one T(n) microsatellite in an 8-kb segment of the dystrophin gene. We found 86 haplotypes in 1,343 chromosomes from around the world. Although a classical out-of-Africa topology was observed in trees based on the variant frequencies, the tree of haplotype sequences reveals three lineages accounting for present-day diversity. The proportion of new recombinants and the diversity of the T(n) microsatellite were used to estimate the age of haplotype lineages and the time of colonization events. The lineage that underwent the great expansion originated in Africa prior to the Upper Paleolithic (27,000-56,000 years ago). A second group, of structurally distinct haplotypes that occupy a central position on the tree, has never left Africa. The third lineage is represented by the haplotype that lies closest to the root, is virtually absent in Africa, and appears older than the recent out-of-Africa expansion. We propose that this lineage could have left Africa before the expansion (as early as 160,000 years ago) and admixed, outside of Africa, with the expanding lineage. Contemporary human diversity, although dominated by the recently expanded African lineage, thus represents a mosaic of different contributions.

Figure 1

Allelic structure of dys44 haplotypes. Position IDs and haplotype names are arbitrary (consistent with Ziętkiewicz et al. , ; Labuda et al. 2000). For clarity, only new (derived) alleles are shown, and an empty space implies identity with the ancestral allele (the same as that found in nonhuman primates), as shown at the top of the figure. The new alleles of the worldwide polymorphic loci (21 sites) are highlighted in color; those of the continentally restricted polymorphisms (15 sites) are indicated by bold letters, whereas the numbers in the column separating sites 72 and 85 indicate the associated length alleles of the T_n microsatellite. Geographic affiliations are indicated on the left (Af = African; Am = American; As = Asiatic; Eu = European; In = Indonesian/PNG; continentally shared haplotypes are left blank); haplotype names are color coded to indicate their structural clustering into haplogroups (see text and fig. 2), reflected by similarity to the dominant most frequent haplotype; and boxes indicate the shortest putative recombination/conversion sites relating the rare recombinant to the dominant haplotype of the group (only for non-African haplotypes).

Figure 2

NJ tree (A) of haplotype sequences and the corresponding haplotype frequencies (counts) (B) in five continental regions. Structural affiliations of haplotypes use the same color code as in figure 1. Asterisks (*) indicate haplotypes that occur in a single continental group.

Figure 3

NJ tree of populations from the distance matrix calculated from haplotype frequencies. The bootstrapping values show how many times in 100 runs the cluster to the right was observed.

Figure 4

Population frequencies of the frequent, continentally shared haplotypes. Different colors, if present, indicate different T_n alleles shared by the same B haplotype. Note that B006 is not shown in the African American sample, because of the strong evidence that it is a result of recent admixture with the Amerindian population (see text).

Figure 5

Frequency distribution of length alleles of the T_n microsatellite worldwide (A), in non-Africans (B), and in sub-Saharan Africans (C), subdivided further into chromosomes carrying African-specific haplotypes (D) and those shared with other continents (E).

Figure A1

Frequencies of 86 observed haplotypes (A) compared with their individual probabilities of back recombination (B). C, Probability distribution of new recombinants resulting from pairwise recombination of the observed haplotypes from the whole data set. Note that the scale of the ordinate in C is reduced by a factor of 10 compared with A and B, that the total count of new haplotypes is 6,857 (C), and that the left and right parts of chart C depict the 86 most frequent novel haplotypes versus the remaining 6,771 relatively infrequent novel haplotypes.