|
|
Post by Admin on Mar 28, 2020 22:51:16 GMT
The Genomic Formation of Human Populations in East Asia Chuan-Chao Wang1,2,3,4,*, Hui-Yuan Yeh5,*, Alexander N Popov6,*, Hu-Qin 3 Zhang7,*, Hirofumi Matsumura8, Kendra Sirak2,9, Olivia Cheronet10, Alexey 4 Kovalev11, Nadin Rohland2, Alexander M. Kim2,12, Rebecca Bernardos2
88 The deep population history of East Asia remains poorly understood due to a
89 lack of ancient DNA data and sparse sampling of present-day people. We report
90 genome-wide data from 191 individuals from Mongolia, northern China,
91 Taiwan, the Amur River Basin and Japan dating to 6000 BCE - 1000 CE, many
92 from contexts never previously analyzed with ancient DNA. We also report 383
93 present-day individuals from 46 groups mostly from the Tibetan Plateau and
94 southern China. We document how 6000-3600 BCE people of Mongolia and the
95 Amur River Basin were from populations that expanded over Northeast Asia,
96 likely dispersing the ancestors of Mongolic and Tungusic languages. In a time
97 transect of 89 Mongolians, we reveal how Yamnaya steppe pastoralist spread
98 from the west by 3300-2900 BCE in association with the Afanasievo culture,
99 although we also document a boy buried in an Afanasievo barrow with ancestry
100 entirely from local Mongolian hunter-gatherers, representing a unique case of
101 someone of entirely non-Yamnaya ancestry interred in this way. The second
102 spread of Yamnaya-derived ancestry came via groups that harbored about a
103 third of their ancestry from European farmers, which nearly completely
104 displaced unmixed Yamnaya-related lineages in Mongolia in the second
105 millennium BCE, but did not replace Afanasievo lineages in western China
106 where Afanasievo ancestry persisted, plausibly acting as the source of the early
107 splitting Tocharian branch of Indo-European languages. Analyzing 20 Yellow
108 River Basin farmers dating to ~3000 BCE, we document a population that was a
109 plausible vector for the spread of Sino-Tibetan languages both to the Tibetan
110 Plateau and to the central plain where they mixed with southern agriculturalists
111 to form the ancestors of Han Chinese. We show that the individuals in a time
112 transect of 52 ancient Taiwan individuals spanning at least 1400 BCE to 600 CE
113 were consistent with being nearly direct descendants of Yangtze Valley first
114 farmers who likely spread Austronesian, Tai-Kadai and Austroasiatic languages
115 across Southeast and South Asia and mixing with the people they encountered,
116 contributing to a four-fold reduction of genetic differentiation during the
117 emergence of complex societies. We finally report data from Jomon hunter
118 gatherers from Japan who harbored one of the earliest splitting branches of East
119 Eurasian variation, and show an affinity among Jomon, Amur River Basin,
120 ancient Taiwan, and Austronesian-speakers, as expected for ancestry if they all
121 had contributions from a Late Pleistocene coastal route migration to East Asia.
 Figure 1: Geographical locations of newly reported ancient individuals.
123 Main text
124 East Asia, one of the oldest centers of animal and plant domestication, today harbors
125 more than a fifth of the world’s human population, with present-day groups speaking
126 languages representing eleven major families: Sino-Tibetan, Tai-Kadai, Austronesian,
127 Austroasiatic, Hmong-Mien, Indo-European, Altaic (Mongolic, Turkic, and
128 Tungusic), Koreanic, Japonic, Yukgahiric, and Chukotko-Kanchatkan1. The past
129 10,000 years have been a period of profound economic and cultural change in East
130 Asia, but our current understanding of the genetic diversity, major mixture events, and
131 population movements and turnovers during the transition from foraging to
132 agriculture remains poor due to minimal sampling of the diversity of present-day
133 people on the Tibetan Plateau and southern China2. A particular limitation has been a
134 deficiency in ancient DNA data, which has been a powerful tool for discerning the
135 deep history of populations in Western and Central Eurasia3-8.
137 We genotyped 383 present-day individuals from 46 populations indigenous to China
138 (n=337) and Nepal (n=46) using the Affymetrix Human Origins array (Table S1 and
139 Supplementary Information section 1). We also report genome-wide data from 191
140 ancient East Asians, many from cultural contexts for which there is no published
141 ancient DNA data. From Mongolia we report 89 individuals from 52 sites dating
142 between ~6000 BCE to ~1000 CE. From China we report 20 individuals from the
143 ~3000 BCE Neolithic site of Wuzhuangguoliang. From Japan we report 7 Jomon
144 hunter-gatherers from 3500-1500 BCE. From the Russian Far East we report 23
145 individuals: 18 from the Neolithic Boisman-2 cemetery at ~5000 BCE, 1 from the
146 Iron Age Yankovsky culture at ~1000 BCE, 3 from the Medieval Heishui Mohe and
147 Bohai Mohe culture at ~1000 CE; and 1 historic period hunter-gatherer from Sakhalin
148 Island. From archaeological sites in Eastern Taiwan—the Bilhun site at Hanben on the
149 main island and the Gongguan site on Green Island—we report 52 individuals from
150 the Late Neolithic through the Iron Age spanning at least 1400 BCE - 600 CE.
152 For all but the Chinese samples we enriched the ancient DNA for a targeted set of
153 about 1.2 million single nucleotide polymorphisms (SNPs)4,9 , while for the
154 Wuzhuangguoliang samples from China we used exome capture (18 individuals) or
155 shotgun sequencing (2 individuals) (Figure 1, Supplementary Data files 1 and 2 and
156 Supplementary Information section 1). We performed quality control to test for
157 contamination by other human sequences, assessed by the rate of cytosine to thymine
158 substitution in the terminal nucleotide and polymorphism in mitochondrial DNA
159 as well as X chromosome sequences in males, and restricted analysis to
160 (Online Table 1). We detected close kinship
161 between individuals at the same site, including a Boisman nuclear family with 2
162 parents and 4 children (Table S2). We merged the new data with previously reported
163 data: 4 Jomon individuals, 8 Amur River Basin Neolithic individuals from the Devil’s
164 Gate site, 72 individuals from the Neolithic to the Iron Age in Southeast Asia, and 8
165 from Nepal7,12-20. We assembled 123 radiocarbon dates using bone from the
166 individuals, of which 94 are newly reported (Online Table 3), and clustered
167 individuals based on time period and cultural associations, then further by genetic
168 cluster which in the Mongolian samples we designated by number (our group names
169 thus have the format “<Country>_<Time Period>_<Genetic Cluster>_<Cultural
170 Association If Any>”) (Supplementary Note, Table S1 and Online Table 1). We
171 merged the data with previously reported data (Online Table 4).
|
|
|
|
Post by Admin on Mar 29, 2020 5:22:28 GMT
173 We carried out Principal Component Analysis (PCA) using smartpca21, projecting the
174 ancient samples onto axes computed using present-day people. The analysis shows
175 that population structure in East Asia is correlated with geographic and linguistic
176 categories, albeit with important exceptions. Groups in Northwest China, Nepal, and
177 Siberia deviate towards West Eurasians in the PCA (Supplementary Information
178 section 2, Figure 2), reflecting multiple episodes of West Eurasian-related admixture
179 that we estimate occurred 5 to 70 generations ago based on the decay of linkage disequilibrium22
180 (Table S3 and Table S4). East Asians with minimal proportions of
181 West Eurasian-related ancestry fall along a gradient with three clusters at their poles.
182 The “Amur Basin Cluster” correlates geographically with ancient and present-day
183 populations living in the Amur River Basin, and linguistically with present-day
184 indigenous people speaking Tungusic languages and the Nivkh. The “Tibetan Plateau
185 Cluster” is most strongly represented in ancient Chokhopani, Mebrak, and Samzdong individuals
186 from Nepal15 and in present-day people speaking Tibetan-Burman
187 languages and living on the Tibetan Plateau. The “Southeast Asian Cluster” is
188 maximized in ancient Taiwan groups and present-day people in Southeast Asia and
189 southern parts of China speaking Austroasiatic, Tai-Kadai and Austronesian
190 languages (Figure S1, Figure S2). Han are intermediate among these clusters, with
191 northern Han projecting close to the Neolithic Wuzhuangguoliang individuals from
192 northern China (Figure 2). We observe two genetic clusters within Mongolia: one falls
193 closer to ancient individuals from the Amur Basin Cluster (‘East’ based on their
 Figure 2: Principal Component Analysis (PCA). (A) Projection of ancient samples onto PCA dimensions 1 and 2 defined by East Asians, Europeans, Siberians and Native Americans. (B) Projection onto groups with the little West Eurasian mixture.
194 geography), and the second clusters toward ancient individuals of the Afanasievo
195 culture ( ‘West’), while a few individuals take intermediate positions between the two
196 (Supplementary Information section 2).
197
198 The three most ancient individuals of the Mongolia ‘East’ cluster are from the
199 Kherlen River region of eastern Mongolia (Tamsag-Bulag culture) and date to 6000-
200 4300 BCE (this places them in the Early Neolithic period, which in Northeast Asia is
201 defined by the use of pottery and not by agriculture23). These individuals are
202 genetically similar to previously reported Neolithic individuals from the cis-Baikal
203 region and have minimal evidence of West Eurasian-related admixture as shown in
204 PCA (Figure 2), f4-statistics and qpAdm (Table S5, Online Table 5, labeled as
205 Mongolia_East_N). The other seven Neolithic hunter-gatherers from northern
206 Mongolia (labeled as Mongolia_North_N) can be modeled as having 5.4% ±1.1%
207 ancestry from a source related to previously reported West Siberian Hunter-gatherers(WSHG)8
208 (Online Table 5), consistent with the PCA where they are part of an east
209 west Neolithic admixture cline in Eurasia with increasing proximity to West Eurasians
210 in groups further west. Because of this ancestry complexity, we use the
211 Mongolia_East_N individuals without significant evidence of West Eurasian-related
212 admixture as reference points for modeling the East Asian-related ancestry in later
213 groups (Online Table 5). The two oldest individuals from the Mongolia ‘West’ cluster
214 have very different ancestry: they are from the Shatar Chuluu kurgan site associated
215 with the Afanasievo culture, with one directly dated to 3316-2918 calBCE (we quote
216 a 95% confidence interval here and in what follows whenever we mention a direct
217 date), and are indistinguishable in ancestry from previously published ancient
218 Afanasievo individuals from the Altai region of present-day Russia, who in turn are
219 similar to previously reported Yamnaya culture individuals supporting findings that
220 eastward Yamnaya migration had a major impact on people of the Afansievo culture5,8
221 . All the later Mongolian individuals in our time transect were mixtures of
222 Mongolian Neolithic groups and more western steppe-related sources, as reflected by
223 statistics of the form f3 (X, Y; Later Mongolian Groups), which resulted in
224 significantly negative Z scores (Z<−3) when Mongolia_East_N was used as X, and
225 when Yamnaya-related Steppe populations, AfontovaGora3, WSHG, or European
226 Middle/Late Neolithic or Bronze Age populations were used as Y (Table S6).
|
|
|
|
Post by Admin on Mar 29, 2020 21:05:59 GMT
228 To quantify the admixture history of the later Mongolians, we again used qpAdm. A
229 large number of groups could be modeled as simple two-way admixtures of
230 Mongolia_East_N as one source (in proportions of 65-100%) and WSHG as the other
231 source (in proportions of 0-35%), with negligible contribution from Yamnaya-related
232 sources as confirmed by including Russia_Afanasievo and Russia_Sintashta groups in
233 the outgroup set (Figure 3). The groups that fit this model were not only the two
234 Neolithic groups (0-5% WSHG), but also the Early Bronze Age people from the
235 Afanasievo Kurgak govi site (15%), the Ulgii group (28%), the main grouping of
236 individuals from the Middle Bronze Age Munkhkhairkhan culture (33%), Late Bronze
237 Age burials of the Ulaanzuukh type (6%), a combined group from the Center-West
238 region (27%), the Mongun Taiga type from Khukh tolgoi (35%), and people of the
239 Iron Age Slab Grave culture (9%). A striking finding in light of previous
240 archaeological and genetic data is that the male child from Kurgak govi (individual
241 I13957, skeletal code AT_629) has no evidence of Yamnaya-related ancestry despite
242 his association with Afanasievo material culture (for example, he was buried in a
243 barrow in the form of circular platform edged by vertical stone slabs, in stretched
244 position on the back on the bottom of deep rectangular pit and with a typical
245 Afanasievo egg-shaped vessel (Supplementary Note); his late Afanasievo chronology
246 is confirmed by a direct radiocarbon date of 2858-2505 BCE24). This is the first
247 known case of an individual buried with Afanasievo cultural traditions who is not
248 overwhelmingly Yamnaya-related, and he also shows genetic continuity with an
249 individual buried at the same site Kurgak govi 2 in a square barrow (individual I6361,
250 skeletal code AT_635, direct radiocarbon date 2618-2487 BCE). We label this second
251 individuals as having an Ulgii cultural association, although a different archaeological
252 assessment associates this individual to the Afanasievo or Chemurchek cultures25, so
253 it is possible that this provides a second example of Afanasievo material culture being
254 adopted by individuals without any Yamnaya ancestry. The legacy of the Yamnaya
255 era spread into Mongolia continued in two individuals from the Chemurchek culture
256 whose ancestry can be only modeled by using Afanasievo as one of the sources
257 (49.0%±2.6%, Online Table 5). This model fits even when ancient European farmers
258 are included in the outgroups, showning that if the long-distance transfer of West
259 European megalithic cultural traditions to people of the Chemurchek culture that has
260 been suggested in the archaeological literature occurred,26 it must have been through
261 spread of ideas rather than through movement of people.
 Figure 3: qpAdm modeling of ancestry change over time in Mongolia. We use Mongolia_East_N, Afanasievo, WSHG, and Sintashta_MLBA as sources, and for each combined archaeological and genetic grouping identify maximally parsimonious models (fewest numbers of sources) that fit with P>0.05 (Online Table 5). We plot results for groupings that give a unique parsimonious model, and include at least one individual with data that “PASS” at high quality and with a confident chronological assignment (Online Table 1). The bars show proportions of each ancestry source, and we also include time spans for the individuals in the cluster. Groupings that include more eastern individuals (longitude >102.7 degrees) are indicated in green and typically have very little Yamnaya-related admixture even at late dates.
263 Beginning in the Middle Bronze Age in Mongolia, there is no compelling evidence
264 for a persistence of the Yamnaya-derivd lineages originally spread into the region
265 with Afanasievo. Instead in the Late Bronze Age and Iron Age and afterward we have
266 data from multiple Mongolian groups whose Yamnaya-related ancestry can only be
267 modeled as deriving not from the initial Afanasievo migration but instead from a later
268 eastward spread into Mongolia related to people of the Middle to Late Bronze Age
269 Sintashta and Andronovo horizons who were themselves a mixture of ~2/3 Yamnaya-related
270 and 1/3 European farmer-related ancestry5,7,8. The Sintashta-related ancestry is
271 detected in proportions of 5% to 57% in individuals from the
272 Mongolia_LBA_6_Khovsgol (a culturally mixed group from the literature14),
273 Mongolia_LBA_3_MongunTaiga, Mongolia_LBA_5_CenterWest,
274 Mongolia_EIA_4_Sagly, Mongolia_EIA_6_Pazyryk, and Mongolia_Mongol groups,
275 with the most substantial proportions of Sintashta-related ancestry always coming
276 from western Mongolia (Figure 3, Online Table 5). For all these groups, the qpAdm
277 ancestry models pass when Afanasievo is included in the outgroups while models
278 with Afanasievo treated as the source with Sintashta more distantly related outgroups
279 are all rejected (Figure 3, Online Table 5). Starting from the Early Iron Age, we
280 finally detect evidence of gene flow in Mongolia from groups related to Han Chinese.
281 Specifically, when Han are included in the outgroups, our models of mixtures in
282 different proportions of Mongolia_East_N, Russia_Afanasievo, Russia_Sintashta, and
283 WSHG continue to work for all Bronze Age and Neolithic groups, but fail for an
284 Early Iron Age individual from Tsengel sum (Mongolia_EIA_5), and for Xiongnu and
285 Mongols. When we include Han Chinese as a possible source, we estimate ancestry
286 proportions of 20-40% in Xiongnu and Mongols (Online Table 5).
288 While the Afanasievo-derived lineages are consistent with having largely disappeared
289 in Mongolia by the Late Bronze Age when our data showed that later groups with
290 Steppe pastoralist ancestry made an impact, we confirm and strengthen previous
291 ancient DNA analysis suggesting that the legacy of this expansion persisted in
292 western China into the Iron Age Shirenzigou culture (410-190 BCE)27. The only
293 parsimonious model for this group that fits according to our criteria is a 3-way
294 mixture of groups related to Mongolia_N_East, Russia_Afanasievo, WSHG. The only
295 other remotely plausible model (although not formally a good fit) also requires
296 Russia_Afanasievo as a source (Figure 3, Online Table 5). The findings of the original
297 study that reported evidence that the Afanasievo spread was the source of Steppe
298 ancestry in the Iron Age Shirenzigou have been questioned with the proposal of
299 alternative models that use ancient Kazakh Steppe Herders from the site of Botai,
300 Wusun, Saka and ancient Tibetans from the site of Mebrak15 in present-day Nepal as
301 major sources for Steppe and East Asian-related ancestry28. However, when we fit
302 these models with Russia_Afanasievo and Mongolian_East_N added to the outgroups,
303 the proposed models are rejected (P-values between 10-7 and 10-2), except in a model
304 involving a single low coverage Saka individual from Kazakhstan as a source
305 (P=0.17, likely reflecting the limited power to reject models with this low coverage).
306 Repeating the modeling using other ancient Nepalese with very similar genetic
307 ancestry to that in Mebrak results in uniformly poor fits (Online Table 5). Thus,
308 ancestry typical of the Afanasievo culture and Mongolian Neolithic contributed to the
309 Shirenzigou individuals, supporting the theory that the Tocharian languages of the
310 Tarim Basin—from the second-oldest-known branch of the Indo-European language
311 family—spread eastward through the migration of Yamnaya steppe pastoralists to the
312 Altai Mountains and Mongolia in the guise of the Afansievo culture, from where they
313 spread further to Xinjiang5,7,8,27,29,30. These results are significant for theories of Indo
314 European language diversification, as they increase the evidence in favor of the
315 hypothesis the branch time of the second-oldest branch in the Indo-European language
tree occurred at the end of the fourth millennium BCE27,29,3
|
|
|
|
Post by Admin on Mar 30, 2020 4:50:58 GMT
318 The individuals from the ~5000 BCE Neolithic Boisman culture and the ~1000 BCE
319 Iron Age Yankovsky culture together with the previously published ~6000 BCE data
320 from Devil’s Gate cave19 are genetically very similar, documenting a continuous
321 presence of this ancestry profile in the Amur River Basin stretching back at least to
322 eight thousand years ago (Figure 2 and Figure S2). The genetic continuity is also
323 evident in the prevailing Y chromosomal haplogroup C2b-F1396 and mitochondrial
324 haplogroups D4 and C5 of the Boisman individuals, which are predominant lineages
325 in present-day Tungusic, Mongolic, and some Turkic-speakers. The Neolithic
326 Boisman individuals shared an affinity with Jomon as suggested by their intermediate
327 positions between Mongolia_East_N and Jomon in the PCA and confirmed by the
328 significantly positive statistic f4 (Mongolia_East_N, Boisman; Mbuti, Jomon).
329 Statistics such as f4 (Native American, Mbuti; Test East Asian,
330 Boisman/Mongolia_East_N) show that Native Americans share more alleles with
331 Boisman and Mongolia_East_N than they do with the great majority of other East
332 Asians in our dataset (Table S5). It is unlikely that these statistics are explained by
333 back-flow from Native Americans since Boisman and other East Asians share alleles
334 at an equal rate with the ~24,000-year-old Ancient North Eurasian MA1 who was
335 from a population that contributed about 1/3 of all Native American ancestry31. A
336 plausible explanation for this observation is that the Boisman/Mongolia Neolithic
337 ancestry was linked (deeply) to the source of the East Asian-related ancestry in Native
338 31. We can also model published data from Neolithic and Early Bronze Age
339 individuals around Lake Baikal7 as sharing substantial ancestry (77-94%) with
340 the lineage represented by Mongolia_East_N, revealing that this type of ancestry was
341 once spread over a wide region spanning across Lake Baikal, eastern Mongolia, and
342 the Amur River Basin (Table S7). Some present-day populations around the Amur
343 River Basin harbor large fractions of ancestry consistent with deriving from more
344 southern East Asian populations related to Han Chinese (but not necessarily Han
345 themselves) in proportions of 13-50%. We can show that this admixture occurred at
346 least by the Early Medieval period because one Heishui_Mohe individual (I3358,
347 directly dated to 1050-1220 CE) is estimated to have harbored more than 50%
348 ancestry from Han or related groups (Table S8).
350 The Tibetan Plateau, with an average elevation of more than 4,000 meters, is one of
351 the most extreme environments in which humans live. Archaeological evidence
352 suggests two main phases for modern human peopling of the Tibetan Plateau. The
353 first can be traced back to at least ~160,000 years ago probably by Denisovans32 and
354 then to 40,000-30,000 years ago as reflected in abundant blade tool assemblages
355 However, it is only in the last ~3,600 years that there is evidence for continuous
356 permanent occupation of this region with the advent of agriculture34. We grouped 17
357 present-day populations from the highlands into three categories based on genetic
358 clustering patterns (Figure S3): “Core Tibetans” who are closely related to the ancient
359 Nepal individuals such as Chokopani with a minimal amount of admixture with
360 groups related to West Eurasians and lowland East Asians in the last dozens of
361 generations, “northern Tibetans” who are admixed between lineages related to Core
362 Tibetans and West Eurasians, and “Tibeto-Yi Corridor” populations (the eastern edge
363 of the Tibetan Plateau connecting the highlands to the lowlands) that includes not just
364 Tibetan speakers but also Qiang and Lolo-Burmese speakers who we estimate using
365 qpAdm4,35 have 30-70% Southeast Asian Cluster-related ancestry (Table S9). We
366 computed f3 (Mbuti; Core Tibetan, non-Tibetan East Asian) to search for non-Tibetans
367 that share the most genetic drift with Tibetans. Neolithic Wuzhuangguoliang, Han and
368 Qiang appear at the top of the list (Table S10), suggesting that Tibetans harbor
369 ancestry from a population closely related to Wuzhuangguoliang that also contributed
370 more to Qiang and Han than to other present-day East Asian groups. We estimate that
371 the mixture occurred 60-80 generations ago (2240-1680 years ago assuming 28 years
372 per generation36 under a model of a single pulse of admixture (Table S11). This
373 represents an average date and so only provides a lower bound on when these two
374 populations began to mix; the start of their period of admixture could plausibly be as
375 old as the ~3,600-year-old date for the spread of agriculture onto the Tibetan plateau.
376 These findings are therefore consistent with archaeological evidence that expansions
377 of farmers from the Upper and Middle Yellow River Basin influenced populations of
378 the Tibetan Plateau from the Neolithic to the Bronze Age as they spread across the
379 China Central plain37,38, and with Y chromosome evidence that the shared common
380 haplogroup Oα-F5 between Han and Tibetans coalesced to a common ancestry less
381 than 5,800 years ago39.
|
|
|
|
Post by Admin on Mar 30, 2020 23:21:30 GMT
383 In the south, we find that the ancient Taiwan Hanben and Gongguan culture
384 individuals dating from at least a span of 1400 BCE - 600 CE are genetically most
385 similar to present-day Austronesian speakers and ancient Lapita individuals from
386 Vanuatu as shown in outgroup f3-statistics and significantly positive f4-statistics
387 (Taiwan_Hanben/Gongguan, Mbuti; Ami/Atayal/Lapita, other Asians) (Table S8).
388 The similarity to Austronesian-speakers is also evident in the Iron Age dominant
389 paternal Y chromosome lineage O3a2c2-N6 and maternal mtDNA lineages E1a,
390 B4a1a, F3b1, and F4b, which are widespread lineages among Austronesianspeakers40,41
391 . We compared the present-day Austronesian-speaking Ami and Atayal of
392 Taiwan with diverse Asian populations using statistics like f4 (Taiwan Iron
393 Age/Austronesian, Mbuti; Asian1, Asian2). Ancient Taiwan groups and Austronesian
394 China and in Hainan Island42 speakers share significantly more alleles with Tai-Kadai speakers
395 in southern mainland than they do with other East Asians (Table S8),
396 consistent with the hypothesis that ancient populations related to present-day Tai
397 Kadai speakers are the source for the spread of agriculture to Taiwan island around
398 5000 years ago43. The Jomon share alleles at an elevated rate with ancient Taiwan
399 individuals and Ami/Atayal as measured by statistics of the form f4 (Jomon, Mbuti;
400 Ancient Taiwan/Austronesian-speaker, other Asians) compared with other East Asian
401 groups, with the exception of groups in the Amur Basin Cluster (Table S8)44.
 Figure 4: qpAdm modeling of Han Chinese cline. We used the ancient Wuzhuangguoliang as a proxy for Yellow River Farmers and Taiwan_Hanben as a proxy for Yangtze River Farmers related ancestry.
403 The Han Chinese are the world’s largest ethnic group. It has been hypothesized based
404 on the archaeologically documented spread of material culture and farming
405 technology, as well as the linguistic evidence of links among Sino-Tibetan languages,
406 that one of the ancestral populations of the Han might have consisted of early farmers
407 along the Upper and Middle Yellow River in northern China, some of whose
408 descendants also may have spread to the Tibetan Plateau and contributed to presentday
409 Tibeto-Burmans. Archaeological and historical evidence document how during
410 the past two millennia, the Han expanded south into regions inhabited by previously
411 established agriculturalists46. Analysis of genome-wide variation among present-day
412 populations has revealed that the Han Chinese are characterized by a “North-South” cline47,48
413 , which is confirmed by our analysis. The Neolithic Wuzhuangguoliang,
414 present-day Tibetans, and Amur River Basin populations, share significantly more
415 alleles with Han Chinese compared with the Southeast Asian Cluster, while the
416 Southeast Asian Cluster groups share significantly more alleles with the majority of
417 Han Chinese groups when compared with the Neolithic Wuzhuangguoliang (Table
418 S12, Table S13). These findings suggest that Han Chinese may be admixed in variable
419 proportions between groups related to Neolithic Wuzhuangguoliang and people
420 related to those of the Southeast Asian Cluster. To determine the minimum number of
421 source populations needed to explain the ancestry of the Han, we used qpWave4,49 to
422 study the matrix of all possible statistics of the form f4 (Han1, Han2; O1, O2), where
423 “O1” and “O2” are outgroups that are unlikely to have been affected by recent gene
424 flow from Han Chinese. This analysis confirms that two source populations are
425 consistent with all of the ancestry in most Han Chinese groups (with the exception of
426 some West Eurasian-related admixture that affects some northern Han Chinese in
427 proportions of 2-4% among the groups we sampled; Table S14 and Table S15).
428 Specifically, we can model almost all present-day Han Chinese as mixtures of two
429 ancestral populations, in a variety of proportions, with 77-93% related to Neolithic
430 Wuzhuangguoliang from the Yellow River basin, and the remainder from a
431 population related to ancient Taiwan that we hypothesize was closely related to the
432 rice farmers of the Yangtze River Basin. This is also consistent with our inference that
433 the Yangtze River farmer related ancestry contributed nearly all the ancestry of
434 Austronesian speakers and Tai-Kadai speakers and about 2/3 of some Austroasiatic
435 speakers17,20 (Figure 4). A caveat is that there is a modest level of modern
436 contamination in the Wuzhuangguoliang we use as a source population for this
437 analysis (Online Table 1), but this would not bias admixture estimates by more than
438 the contamination estimate of 3-4%. The average dates of West Eurasian-related
439 admixture in northern Han Chinese populations Han_NChina and Han_Shanxi are 32-
440 45 generations ago, suggesting that mixture was continuing at the time of the Tang
441 Dynasty (618-907 CE) and Song Dynasty (960-1279 BCE) during which time there
442 are historical records of integration of Han Chinese amd western ethnic groups, but
443 this date is an average so the mixture between groups could have begun earlier.
|
|
