Aa2000nAtureAmericaInc..http:/iGenetics.nature.com letter 161718静如02122224252627282加3343533839404144444546 146 495051525555655896日B26466b昭s77n7;4777}aB84B8a9制9ss的100101021104105105y10Bt10ttt21t31 ⅨX ig. 1 Maximum parsimony phylogeny of human NRY chromosome bi-allelic variation. The tree is rooted with respect to non-human primate sequences. The 116 M139. also share the only homopolymer-associated marker M91. All haplogroup I in in Africa, share the ancestral allele of M168. haplogroup Ill is generally the most frequent one in Africa. Its frequency decreases wit 2z898 oup ough M1 and M145, is found mainly in Ja endently resisted effectively subsequent gene flow 23. The distinction between Eurasians and East Asians was also observed with mtDNA (ref. 24)and autosomal genes2. Haplogroup x is common in the Americas, although its origin may have been in Central Asia where traces of it persist(Table 1). to an estimate of 46,000-91,000 years based on 8 Y chromosome The new levels of bi-allelic variation revealed here indicate a microsatellites 2 and, therefore, is considerably less than estimates recent ancestry of the paternal lineages of our species from Africa of greater than 100,000 years obtained previously. Of course, this and testify to the informativeness of the Y chromosome in deci- assumes that selection or population structure has not had a major phering the evolution of humankind. effect on NRY diversity, an assumption that may be wrong in light of our findings of significantly reduced variability on NRY. As the Methods number of mutations of all segments departing from the DNA samples. The ascertainment set consisted of the following 53 samples mately 6,900 years. This puts the age of M168, which marks the II, Ill: 2 KhoisanI, Ill; 1 Berta VI: 1 Surma L: 1 Mali Tuareg Ill; 1 Mali Bozo expansion of anatomically modern humans out of Africa, at Il; Europe: 1 Sardinian VI: 2 Italian VI IX; 1 German VI: 3 Basque VI, IX of 47,000 years with 95% probability intervals of 35,000-89.000 Taiwan Ami, vIL, 2 Cambodian VI, vIl: Pakistan: 2 Hunza vL Ix: 2 Pathan using the program GENETREE (ref. 11). This concurs with Arab lx: 1 Uzbek Ix: 1 Kazak V: Mideast: I druze vi:Pacific: 2 New recent archeological and mtDNA data, and is also consistent, Guinean V. vll: 2 Bougainville Islanders VIIl; 2 Australian VI, X: America: though at a compressed time scale, with the weak Garden-of-Eden 1 Brazil Surui. 1 Brazil Karatina, 1 Columbian, I Mayan all X. We yggmic orally modern humans peft africa and se parated into several region s n dhina for af chargers other than tntsn gn the oerapihac ly isolated groups represented today by the major haplogroups branches of the phylogeny. We genotyped the latter only in individuals Ill-X. Those groups remained small throughout the last glaciation from the haplogroup to which those markers belonged. This hierarchic efore they underwent roughly simultaneous expansions in size as genotyping protocol was necessitated by the limited amounts of genomic suggested by a star-like genealogy(Fig. 1) nature genetics.volume 26.november 2000letter nature genetics • volume 26 • november 2000 359 to an estimate of 46,000–91,000 years based on 8 Y chromosome microsatellites12 and, therefore, is considerably less than estimates of greater than 100,000 years obtained previously5. Of course, this assumes that selection or population structure has not had a major effect on NRY diversity, an assumption that may be wrong in light of our findings of significantly reduced variability on NRY. As the average number of mutations of all segments departing from the root is 8.60 (Table 2), and with a TMRCA value of 59,000 years, the average time for adding a new mutation to the tree is approxi￾mately 6,900 years. This puts the age of M168, which marks the expansion of anatomically modern humans out of Africa, at approximately 44,000 years, in agreement with a previous estimate of 47,000 years with 95% probability intervals of 35,000–89,000 years using the program GENETREE (ref. 11). This concurs with recent archeological13 and mtDNA data14, and is also consistent, though at a compressed time scale, with the weak Garden-of-Eden hypothesis15. Under this hypothesis, a small subgroup of behav￾iourally modern humans13 left Africa and separated into several fairly isolated groups represented today by the major haplogroups III–X. Those groups remained small throughout the last glaciation before they underwent roughly simultaneous expansions in size as suggested by a star-like genealogy (Fig. 1). The new levels of bi-allelic variation revealed here indicate a recent ancestry of the paternal lineages of our species from Africa and testify to the informativeness of the Y chromosome in deci￾phering the evolution of humankind. Methods DNA samples. The ascertainment set consisted of the following 53 samples with their subsequently determined haplogroup designations: Africa: 3 Central African Republic Biaka II, III (1); 2 Zaire Mbuti II, III; 2 Lissongo II, III; 2 Khoisan I, III; 1 Berta VI; 1 Surma I; 1 Mali Tuareg III; 1 Mali Bozo III; Europe: 1 Sardinian VI; 2 Italian VI IX; 1 German VI; 3 Basque VI, IX (2); Asia: 3 Japanese IV, V, VII; 2 Han Chinese VII, 1 Taiwan Atayal VII, 1 Taiwan Ami, VII, 2 Cambodian VI, VII; Pakistan: 2 Hunza VI, IX; 2 Pathan VI, VII; 1 Brahui VIII; 1 Baloochi VI; 3 Sindhi III, VI, VIII; Central Asia: 2 Arab IX; 1 Uzbek IX; 1 Kazak V; MidEast: 1 Druze VI; Pacific: 2 New Guinean V, VIII; 2 Bougainville Islanders VIII; 2 Australian VI, X: America: 1 Brazil Surui, 1 Brazil Karatina, 1 Columbian, 1 Mayan all X. We geno￾typed an additional 1,009 chromosomes, representing 21 geographic regions, by DHPLC for all markers other than those on the terminal branches of the phylogeny. We genotyped the latter only in individuals from the haplogroup to which those markers belonged. This hierarchic genotyping protocol was necessitated by the limited amounts of genomic DNA available for most samples. 118 127 171 63 13 51 144 59 28 32 14 06 31 114 141 135 71 49 29 23 146 60 152 109 150 108 43 129 169 30 112 115 108 91 42 94 139 168 116.1 02 155 10 149 58 154 66 156 41 54 75 96 85 90 98 132 33 44 123 136 34 148 78 35 107 81 165 40 15 55 57 64 116.2 125 151 01 145 38 48 93 08 131 105 86 77 130 72 26 161 21 170 89 12 102 99 172 92 67 166 163 137 69 68 47 158 82 36 97 52 39 138 62 113 110 162 133 117 134 122 07 88 164 159 121 101 119 50 103 95 111 09 70 147 11 61 22 20 76 27 46 128 106 104 16 05 83 04 157 37 173 160 126 18 65 153 167 87 17 56 64 73 74 45 120 19 03 143 124 25 174 175 1 14 2 3 4 5 6 7 8 9 10 11 12 13 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 100 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 101102103104105106107108109110111112113114115116 I VI VII VIII IX II III IV V X Fig. 1 Maximum parsimony phylogeny of human NRY chromosome bi-allelic variation. The tree is rooted with respect to non-human primate sequences. The 116 numbered compound haplotypes were constructed from 167 mutations, of which 160 were discovered by DHPLC. The remaining seven were taken from the litera￾ture and included YAP (M1)17, DYS271 (M2)18, PN3 (M29)19, SRY 4064 (M40)5, TAT (M46)20, RPS4YC711T (M130)21 and SRY 2627 (M167)22. Marker numbers indicated on the segments are discontinuous because of the removal of all but one polymorphism associated with tandem repeats and homopolymer tracts whose ancestral state is uncertain. Haplotypes are assorted into 10 haplogroups (I–X) using criteria given in Table 2. Haplogroup I members, ancestral for M42, M94 and M139, also share the only homopolymer-associated marker M91. All haplogroup I individuals have an 8-T length variant, whereas 1,009 men in haplogroups II–X have 9 and in 2 cases 10-T length variants (not shown). Only one inconsistent haplogroup X individual had an 8-T length variant (not shown). Haplogroups I and II, both of which are almost exclusively represented in Africa, share the ancestral allele of M168. haplogroup III is generally the most frequent one in Africa. Its frequency decreases with increasing distance from Africa, from 27% in the Mid-East to a few per cent in Northern Europe and South and Central Asia. Haplogroup IV, related to the former through M1 and M145, is found mainly in Japan. Haplogroups V and VIII are prevalent in New Guinea and Australia, but they are also found at varying though smaller frequencies throughout Asia. Haplogroup VIII represents the relevant source of Haplogroups VII, IX and X. Haplogroups VI and IX are found mostly in Europe and the Indus Valley. They are not observed in East Asia, where haplogroup VII dominates, suggesting that this part of the world where agriculture developed inde￾pendently resisted effectively subsequent gene flow23. The distinction between Eurasians and East Asians was also observed with mtDNA (ref. 24) and autosomal genes25. Haplogroup X is common in the Americas, although its origin may have been in Central Asia where traces of it persist (Table 1). © 2000 Nature America Inc. • http://genetics.nature.com © 2000 Nature America Inc. • http://genetics.nature.com