REVIEWS chromosome Alu polymorphism; MI in FIG. 1)5.Therefore, most East Asian Y chromosomes (H2-H17)can be traced back to an African ancestor. However, as the East Asian individuals that carry H1 (about 10% )are ancestral to all the other haplotypes. and are therefore indistinguishable from those in Africans, these could conceivably be carrying Y-chro M130 M1 mosomal lineages that represent a local origin. Recently, up to 98 new Y-chromosomal simple uence polymorphisms(point mutations and short the flanking non-coding regions of four Y-chromosome ●M122·M119◆M95●M5●M45 genes.0.They reveal an extreme geographic structure, with the oldest clades representing African populations and the younger ones representing some African and all African Y chromosomes seem to be derived from a small number of Africans Thomson et al.39 estimated ●M0●M11 the age of the common ancestor of the y chromosomes to be 59,000 years (95% confidence interval: African ancestors H1 H2 H3 H4 H7 H6 H8 H9 H10 H11 H12 H5 H17 H13 H14 H15 H16 40,000-140,000 years) if based on the assumption of recent population growth, or 84,000 years(95% confi- Figure 1 I The phylogeny of Y-chromosome haplotypes in East Asians(H1-H17) The Y- dence interval: 55,000-149,000 years)if a constant pop- chromosome haplotypes H1-H17 are observed in East Asian populations2. 22.East Asian- ulation size is assumed. The recent common african specific haplotypes(H6-H12)are indicated in red. The genetic loci associated with the ancestry is therefore a strong confirmation of the Out- plotypes are labelled next to the branches, and the polymorphisms at these loci are indicated in brackets. The shaded region represents the work in REFS 21, 22. On the basis of this of-Africa hypothesis In contrast, the clear geographic work, the ancestral haplotype. Hi, is indistinguishable from African haplotypes. Mutations structure was not observed in the mtDNA data owing and 2 are two ancient polymorphisms (point mutations). Mutations 1 and 2 establish that H to frequent recurrence of mutations, which tends to blur the geographic pattern, or possibly because of the larger EFFECTIVEPOPULATION SIZE of mt DNA compared with at of the y chromosome(BOX 2) reducing the bias that might be introduced by sampling Two critical point mutations, defined as mutations 1 errors or from geneticdrift and 2 are believed to be associated with the out-of Of the haplotypes derived from the 19 Y-chromo- Africa migratory event. When the East Asian data are some biallelic markers, 17 are present in more than incorporated into the global tree, all the haplotypes 1,000 East Asian individuals from different ethnic back- found in East Asia(including HI)fall into the younger grounds. No recurrent mutations were observed and the clade unified by mutation 2. This mutation is derived phylogenetic relationships among the haplotypes are from mutation 1(REE. 40), which is only polymorphic in unequivocal owing to the non-recombinant nature ofY Africa"(FIG. 1). Therefore all 1,046 Y chromosomes 122 chromosomes (FIG. 1). Therefore, the geographic distrib- sampled from East Asia are directly derived from the ution(see FIG. 2 for the geographic locations of rican li With these data, the replacement of GENETIC DRIFT tions sampled)of these 17 haplotypes reflects the demo- modern humans of African origin in East Asia is com- hanges in allele frequency that graphic history of the populations under study. HI is plete. This leaves the multiregional hypothesis onlya the ancestral haplotype and H2-H17 are the derived remote possibility (in East Asia, at least), although a appearing in offspring are not a haplotypes. The majority(about 70%)of East AsianY larger sample may be needed to establish this with perfectly representative sample chromosomes(H6-H12)are specific East Asians and stronger confidence example, in small populations). absent in other world populations, indicating a com mon origin of East Asian populations (FIG. 2). Those Genetic and archaeological data ALUSEQUENCE aplotypes specific to East Asians share a C-G muta- The Out-of-Africa hypothesis has been subjected to two tion at locus m9 that is also prevalent in other world main criticisms from the school of multiregionalists opulations, except those in Africa".H5 and H13-H17 The first is the claim that the multiregional hypothesis about 300,000 copies.The also belong to this lineage. H4 was derived from Hl, but has been misinterpreted, by suggesting that it demands sequence is about 300 base pairs is ancestral to the MgG haplotypes(H5-H17)and is that modern populations have strictly independent ori- defined by a C-T mutation at locus M89. This muta- gins. According to the network(trellis)model, frequent (Alud) that cleaves it. tion has a non-trivial contribution (roughly 4% on gene exchanges would have occurred between continen- average)to East Asians, especially to northerners(for tal populations since H erectus came out of Africa abou EFFECTIVE POPULATION SIZE Mongolians the h2 and h3 haplotypes whichare evolution act:图mm)pgw individuals who contribute to predominant in Tibetan and Japanese populations but claimed that the extensive genetic data supporting the the next generation, assuming generally absent in other East Asians, contain an Out-of-Africa hypothesis could also be explained under random mating insertion of a 300-base-pair ALU SEQUENCE, or YAP(Y- this version of the multiregional hypothesis. However, 128 NOVEMBER 2000 VOLUME 1 Nnature. com/reviews/genetics M@ 2000 Macmillan Magazines Ltd128 | NOVEMBER 2000 | VOLUME 1 www.nature.com/reviews/genetics REVIEWS chromosome Alu polymorphism; M1 in FIG. 1)5 .Therefore, most East Asian Y chromosomes (H2–H17) can be traced back to an African ancestor. However, as the East Asian individuals that carry H1 (about 10%) are ancestral to all the other haplotypes, and are therefore indistinguishable from those in Africans21, these could conceivably be carrying Y-chro￾mosomal lineages that represent a local origin. Recently, up to 98 new Y-chromosomal simple￾sequence polymorphisms (point mutations and short insertions or deletions) were reported in the coding and the flanking non-coding regions of four Y-chromosome genes39,40. They reveal an extreme geographic structure, with the oldest clades representing African populations and the younger ones representing some African and all non-African populations. In other words, the non￾African Y chromosomes seem to be derived from a small number of Africans. Thomson et al.39 estimated the age of the common ancestor of the Y chromosomes to be 59,000 years (95% confidence interval: 40,000–140,000 years) if based on the assumption of recent population growth, or 84,000 years (95% confi￾dence interval: 55,000–149,000 years) if a constant pop￾ulation size is assumed. The recent common African ancestry is therefore a strong confirmation of the Out￾of-Africa hypothesis. In contrast, the clear geographic structure was not observed in the mtDNA data1,3 owing to frequent recurrence of mutations, which tends to blur the geographic pattern, or possibly because of the larger EFFECTIVE POPULATION SIZE of mtDNA compared with that of the Y chromosome (BOX 2). Two critical point mutations, defined as mutations 1 and 2, are believed to be associated with the Out-of￾Africa migratory event. When the East Asian data are incorporated into the global tree, all the haplotypes found in East Asia (including H1) fall into the younger clade unified by mutation 2. This mutation is derived from mutation 1 (REF. 40), which is only polymorphic in Africa40 (FIG. 1). Therefore, all 1,046 Y chromosomes21,22 sampled from East Asia are directly derived from the African lineages. With these data, the replacement of modern humans of African origin in East Asia is com￾plete. This leaves the multiregional hypothesis only a remote possibility (in East Asia, at least), although a larger sample may be needed to establish this with stronger confidence. Genetic and archaeological data The Out-of-Africa hypothesis has been subjected to two main criticisms from the school of multiregionalists. The first is the claim that the multiregional hypothesis has been misinterpreted, by suggesting that it demands that modern populations have strictly independent ori￾gins. According to the network (trellis) model, frequent gene exchanges would have occurred between continen￾tal populations since H. erectus came out of Africa about one million years ago, causing shared multiregional evolution across the human range (BOX 1)42. It was claimed that the extensive genetic data supporting the Out-of-Africa hypothesis could also be explained under this version of the multiregional hypothesis. However, reducing the bias that might be introduced by sampling errors or from GENETIC DRIFT. Of the haplotypes derived from the 19 Y-chromo￾some biallelic markers, 17 are present in more than 1,000 East Asian individuals from different ethnic back￾grounds. No recurrent mutations were observed and the phylogenetic relationships among the haplotypes are unequivocal owing to the non-recombinant nature of Y chromosomes (FIG. 1). Therefore, the geographic distrib￾ution (see FIG. 2 for the geographic locations of popula￾tions sampled) of these 17 haplotypes reflects the demo￾graphic history of the populations under study. H1 is the ancestral haplotype and H2–H17 are the derived haplotypes. The majority (about 70%) of East Asian Y chromosomes (H6–H12) are specific to East Asians and absent in other world populations, indicating a com￾mon origin of East Asian populations (FIG. 2). Those haplotypes specific to East Asians share a C→G muta￾tion at locus M9 that is also prevalent in other world populations, except those in Africa36. H5 and H13–H17 also belong to this lineage. H4 was derived from H1, but is ancestral to the M9G haplotypes (H5–H17) and is defined by a C→T mutation at locus M89. This muta￾tion has a non-trivial contribution (roughly 4% on average) to East Asians, especially to northerners (for example, 20% in Hui, 11% in Manchurians and 8.3% in Mongolians). The H2 and H3 haplotypes, which are predominant in Tibetan and Japanese populations but are generally absent in other East Asians, contain an insertion of a 300-base-pair ALU SEQUENCE, or YAP (Y￾African ancestors H1 1 2 M130 (C→T) M1 (Alu insertion) M89 (C→T) M9 (C→G) M122 M119 M95 M5 M45 M15 M7 M134 M50 M103 M110 M111 M88 M120 M117 M3 H2 H3 H4 H7 H6 H8 H9 H10 H11 H12 H5 H17 H13 H14 H15 H16 Figure 1 | The phylogeny of Y-chromosome haplotypes in East Asians (H1–H17). The Y￾chromosome haplotypes H1–H17 are observed in East Asian populations21,22. East Asian￾specific haplotypes (H6–H12) are indicated in red. The genetic loci associated with the haplotypes are labelled next to the branches, and the polymorphisms at these loci are indicated in brackets. The shaded region represents the work in REFS 21,22. On the basis of this work, the ancestral haplotype, H1, is indistinguishable from African haplotypes. Mutations 1 and 2 are two ancient polymorphisms (point mutations)40. Mutations 1 and 2 establish that H1 is actually derived from African haplotypes. GENETIC DRIFT Changes in allele frequency that result because the genes appearing in offspring are not a perfectly representative sample of the parental genes (for example, in small populations). ALU SEQUENCE A dispersed, intermediately repetitive DNA sequence found in the human genome in about 300,000 copies. The sequence is about 300 base pairs long. The name Alu comes from the restriction endonuclease (AluI) that cleaves it. EFFECTIVE POPULATION SIZE The size of a population determined by the number of individuals who contribute to the next generation, assuming random mating. © 2000 Macmillan Magazines Ltd