|
Post by Admin on Dec 8, 2020 5:25:21 GMT
The ancestry of late Levantine Bronze Age populations It was striking to us that previously published Bronze Age Levantine samples from the sites of 'Ain Ghazal in present-day Jordan (Levant_BA_South) and Sidon in present-day Lebanon (Levant_BA_North) can be modeled as two-way admixtures, without the Anatolia_N contribution that is required to model the Levant_ChL population24,26. This suggests that the Levant_ChL population may not be directly ancestral to these later Bronze Age Levantine populations, because if it were, we would also expect to detect an Anatolia_N component of ancestry. In what follows, we treat Levant_BA_South and Levant_BA_North as separate populations for analysis, since the symmetry statistic f4(Levant_BA_North, Levant_BA_South; A, Chimp) is significant for a number test populations A (|Z| ≥ 3) (Supplementary Data 5), consistent with the different estimated proportions of Levant_N and Iran_ChL ancestry reported in24,26.
To test the hypothesis that Levant_ChL may be directly ancestral to the Bronze Age Levantine populations, we attempted to model both Levant_BA_South and Levant_BA_North as two-way admixtures between Levant_ChL and every other ancient population in our dataset, using the base 09NW set of populations as the “Right” outgroups. We also compared these models to the previously published models that used the Levant_N and Iran_ChL populations as sources (Table 2; Supplementary Figure 5; Supplementary Data 6). In the case of Levant_BA_South from 'Ain Ghazal, Jordan, multiple models were plausible, and thus we returned to the strategy of adding additional “Right” population outgroups that are differentially related to one or more of the “Left” populations (specifically, we added various combinations of Armenia_EBA, Steppe_EMBA, Switzerland_HG, Iran_LN, and Iran_N). Only the model including Levant_N and Iran_ChL remains plausible under all conditions. Thus, we can conclude that groups related to Levant_ChL contributed little ancestry to Levant_BA_South.
Source left populations Admixture proportions Target A B Outgroup right pops p Value rank = 2 A B Standard error Levant_BA_South Levant_N Iran_ChL 09NW 9.88E−01 0.549 0.451 0.031 Levant_BA_South Levant_N Iran_ChL 09NWFPY 5.14E−01 0.571 0.429 0.026 Levant_BA_South Levant_N Iran_ChL 09NWFPSD 1.95E−01 0.582 0.418 0.025 Levant_BA_South Levant_N Iran_ChL 09NWA 9.94E−01 0.55 0.45 0.027 Levant_BA_South Levant_N Iran_ChL 09NWAZ 1.39E−02 0.601 0.399 0.026 Levant_BA_South Levant_ChL CHG 09NW 5.97E−02 0.788 0.212 0.032 Levant_BA_South Levant_ChL CHG 09NWFPY 1.82E−03 0.812 0.188 0.024 Levant_BA_South Levant_ChL Iran_ChL 09NW 2.00E−01 0.714 0.286 0.04 Levant_BA_South Levant_ChL Iran_ChL 09NWFPY 3.06E−02 0.723 0.277 0.033 Levant_BA_South Levant_ChL Iran_LN 09NW 3.53E−01 0.717 0.283 0.039 Levant_BA_South Levant_ChL Iran_LN 09NWFPY 1.22E−02 0.779 0.221 0.026 Levant_BA_South Levant_ChL Iran_HotuIIIb 09NW 2.43E−01 0.556 0.444 0.051 Levant_BA_South Levant_ChL Iran_HotuIIIb 09NWFPSD 3.79E−02 0.585 0.415 0.047 Levant_BA_South Levant_ChL Iran_N 09NW 4.41E−01 0.797 0.203 0.028 Levant_BA_South Levant_ChL Iran_N 09NWFPSD 8.00E−04 0.853 0.147 0.075 Levant_BA_North Levant_N Iran_ChL 09NW 0.003804 0.348 0.652 0.028 Levant_BA_North Levant_N Iran_ChL Haber 0.222705 0.518 0.482 0.04 Levant_BA_North Levant_N Iran_ChL Haber + A 0.002457 0.394 0.606 0.025 Levant_BA_North Levant_ChL Iran_LN 09NW 0.267145 0.532 0.468 0.031 Levant_BA_North Levant_ChL Iran_LN Haber 0.398822 0.555 0.445 0.04 Levant_BA_North Levant_ChL Iran_LN Haber + A 0.455948 0.535 0.465 0.019 Levant_BA_North Levant_ChL Iran_N 09NW 0.401157 0.63 0.37 0.024 Levant_BA_North Levant_ChL Iran_N Haber 0.638884 0.655 0.345 0.035 Levant_BA_North Levant_ChL Iran_N Haber + A 0.693801 0.638 0.362 0.015 Levant_BA_North Levant_ChL Iran_HotuIIIb 09NW 0.216066 0.377 0.623 0.033 Levant_BA_North Levant_ChL Iran_HotuIIIb Haber 0.03318 0.299 0.701 0.047 Levant_BA_North Levant_ChL Iran_HotuIIIb Haber + A 0.007102 0.399 0.601 0.019 Note: Populations that produce p values greater than 0.05 with plausible admixture proportions (between 0 and 1) are highlighted in italic. Models that are robust to the maximum number of outgroups are shown in bold
We observe a qualitatively different pattern in the Levant_BA_North samples from Sidon, Lebanon, where models including Levant_ChL paired with either Iran_N, Iran_LN, or Iran_HotuIIIb populations appear to be a significantly better fit than those including Levant_N + Iran_ChL. We largely confirm this result using the “Right” population outgroups defined in Haber et al.26 (abb. Haber: Ust_Ishim, Kostenki14, MA1, Han, Papuan, Ami, Chuckhi, Karitiana, Mbuti, Switzerland_HG, EHG, WHG, and CHG), although we find that the specific model involving Iran_HotuIIIb no longer works with this “Right” set of populations. Investigating this further, we find that the addition of Anatolia_N in the “Right” outgroup set excludes the model of Levant_N + Iran_ChL favored by Haber et al.26. These results imply that a population that harbored ancestry more closely related to Levant_ChL than to Levant_N contributed to the Levant_BA_North population, even if it did not contribute detectably to the Levant_BA_South population.
We obtained additional insight by running qpAdm with Levant_BA_South as a target of two-way admixture between Levant_N and Iran_ChL, but now adding Levant_ChL and Anatolia_N to the basic 09NW “Right” set of 11 outgroups. The addition of the Levant_ChL causes the model to fail, indicating that Levant_BA_South and Levant_ChL share ancestry following the separation of both of them from the ancestors of Levant_N and Iran_ChL. Thus, in the past there existed an unsampled population that contributed both to Levant_ChL and to Levant_BA_South, even though Levant_ChL cannot be the direct ancestor of Levant_BA_South because, as described above, it harbors Anatolia_N-related ancestry not present in Levant_BA_South.
|
|
|
Post by Admin on Dec 8, 2020 23:45:28 GMT
Genetic heterogeneity in the Levantine Bronze Age We were concerned that our finding that the Levant_ChL population was a mixture of at least three groups might be an artifact of not having access to samples closely related to the true ancestral populations. One specific possibility we considered is that a single ancestral population admixed into the Levant to contribute to both the Levant_ChL and the Levant_BA_South populations, and that this was an unsampled population on an admixture cline between Anatolia_N and Iran_ChL, explaining why qpAdm requires three source populations to model it. To formally test this hypothesis, we used qpWave36,37,38, which determines the minimum number of source populations required to model the relationship between “Left” populations relative to “Right” outgroup populations. Unlike qpAdm, qpWave does not require that populations closely related to the true source populations are available for analysis. Instead it treats all “Left” populations equally, and attempts to determine the minimum number of theoretical source populations required to model the “Left” population set, relative to the “Right” population outgroups. Therefore, we model the relationship between Levant_N, Levant_ChL, and Levant_BA_South as “Left” populations, relative to the 09NW “Right” outgroup populations (Table 3). We find that a minimum of three source populations continues to be required to model the ancestry of these Levantine populations, supporting a model in which at least three separate sources of ancestry are present in the Levant between the Neolithic, Chalcolithic, and Bronze Age.
Table 3 Determining the number of streams of ancestry in the Levant From: Ancient DNA from Chalcolithic Israel reveals the role of population mixture in cultural transformation
Left pops Right pops Rank Degrees of freedom Chi squared p Value Levant_N 0 20 190.024 1.047e−29 Levant_ChL 09NW 1 9 32.641 1.541e−4 Levant_BA_South 2 0 0.000 1.000 Levant_N 0 20 399.438 2.673e−72 Levant_BA_South 09NW 1 9 6.574 0.681 Levant_BA_North 2 0 0.000 1.000 Levant_N 0 20 706.552 3.221e−135 Levant_BA_South 09NWZ 1 9 28.050 1.772e−3 Levant_BA_North 2 0 0.000 1.000 Note: Models that have a rank that is plausible (i.e., p value of greater than 0.05) are shown in bold. Rank is equal to the minimum number of source populations required to model the “Left” population group relative to the “Right” population group, minus 1 (thus, Rank 2, which is the only working solution for all sets of three “Left” populations, reflects three admixing populations)
We applied qpWave again, replacing Levant_ChL with Levant_BA_North, and found that the minimum number of source populations is only two. However, when we include the Levant_ChL population as an additional outgroup, three source populations are again required. This suggests that in the absence of the data from Levant_ChL there is insufficient statistical leverage to detect Anatolian-related ancestry that is truly present in admixed form in the Levant_BA_North population (data from the Levant_ChL population makes it possible to detect this ancestry). This may explain why Haber et al.26 did not detect the Anatolian Neolithic-related admixture in Levant_BA_North.
Biologically important mutations in the Peqi’in population This study nearly doubles the number of individuals with genome-wide data from the ancient Levant. Measured in terms of the average coverage at SNPs, the increase is even more pronounced due to the higher quality of the data reported here than in previous studies of ancient Near Easterners24,26. Thus, the present study substantially increases the power to analyze the change in frequencies of alleles known to be biologically important.
We leveraged our data to examine the change in frequency of SNP alleles known to be related to metabolism, pigmentation, disease susceptibility, immunity, and inflammation in the Levant_ChL population, considered in relation to allele frequencies in the Levant_N, Levant_BA_North, Levant_BA_South, Anatolia_N and Iran_ChL populations and present-day pools of African (AFR), East Asian (EAS), and European (EUR) ancestry in the 1000 Genomes Project Phase 3 dataset39 (Supplementary Data 7).
We highlight three findings of interest. First, an allele (G) at rs12913832 near the OCA2 gene, with a proven association to blue eye color in individuals of European descent40, has an estimated alternative allele frequency of 49% in the Levant_ChL population, suggesting that the blue-eyed phenotype was common in the Levant_ChL.
Second, an allele at rs1426654 in the SLC24A5 gene which is one of the most important determinants of light pigmentation in West Eurasians41 is fixed for the derived allele (A) in the Levant_ChL population suggesting that a light skinned phenotype may have been common in this population, although any inferences about skin pigmentation based on allele frequencies observed at a single site need to be viewed with caution42.
Third, an allele (G) at rs6903823 in the ZKSCAN3 and ZSCAN31 genes which is absent in all early agriculturalists reported to date (Levant_N, Anatolia_N, Iran_N) and that has been argued to have been under positive selection by Mathieson et al.31, occurs with an estimated frequency of 20% in the Levant_ChL, 17% in the Levant_BA_South, and 15% in the Iran_ChL populations, while it is absent in all other populations. This suggests that the allele was rising in frequency in Chalcolithic and Bronze Age Near Eastern populations at the same time as it was rising in frequency in Europe.
|
|
|
Post by Admin on Dec 9, 2020 19:38:56 GMT
Discussion The Chalcolithic period in the Levant witnessed major cultural transformations in virtually all areas of culture, including craft production, mortuary and ritual practices, settlement patterns, and iconographic and symbolic expression43. The current study provides insight into a long-standing debate in the prehistory of the Levant, implying that the emergence of the Chalcolithic material culture was associated with population movement and turnover.
The quality of ancient DNA obtained from the Peqi’in Cave samples is excellent relative to other sites in the Near East. We hypothesize that the exceptional preservation is due to two factors. First, the targeted sampling of ancient DNA from the petrous portion of the temporal bone makes it possible to obtain high-quality ancient DNA from previously inaccessible geographic regions24,27,44,45. Secondly, the environment of Peqi’in Cave is likely to be favorable for DNA preservation. The skeletal remains—either stored in ossuaries or laid in the ground—were quickly covered by a limestone crust, isolating them from their immediate surroundings and protecting them from acidic conditions that are known to be damaging to DNA.
We find that the individuals buried in Peqi’in Cave represent a relatively genetically homogenous population. This homogeneity is evident not only in the genome-wide analyses but also in the fact that most of the male individuals (nine out of ten) belong to the Y-chromosome haplogroup T (see Supplementary Table 1), a lineage thought to have diversified in the Near East46. This finding contrasts with both earlier (Neolithic and Epipaleolithic) Levantine populations, which were dominated by haplogroup E24, and later Bronze Age individuals, all of whom belonged to haplogroup J24,26.
Our finding that the Levant_ChL population can be well-modeled as a three-way admixture between Levant_N (57%), Anatolia_N (26%), and Iran_ChL (17%), while the Levant_BA_South can be modeled as a mixture of Levant_N (58%) and Iran_ChL (42%), but has little if any additional Anatolia_N-related ancestry, can only be explained by multiple episodes of population movement. The presence of Iran_ChL-related ancestry in both populations – but not in the earlier Levant_N – suggests a history of spread into the Levant of peoples related to Iranian agriculturalists, which must have occurred at least by the time of the Chalcolithic. The Anatolian_N component present in the Levant_ChL but not in the Levant_BA_South sample suggests that there was also a separate spread of Anatolian-related people into the region. The Levant_BA_South population may thus represent a remnant of a population that formed after an initial spread of Iran_ChL-related ancestry into the Levant that was not affected by the spread of an Anatolia_N-related population, or perhaps a reintroduction of a population without Anatolia_N-related ancestry to the region. We additionally find that the Levant_ChL population does not serve as a likely source of the Levantine-related ancestry in present-day East African populations (see Supplementary Note 4)24.
These genetic results have striking correlates to material culture changes in the archaeological record. The archaeological finds at Peqi’in Cave share distinctive characteristics with other Chalcolithic sites, both to the north and south, including secondary burial in ossuaries with iconographic and geometric designs. It has been suggested that some Late Chalcolithic burial customs, artifacts and motifs may have had their origin in earlier Neolithic traditions in Anatolia and northern Mesopotamia8,13,47. Some of the artistic expressions have been related to finds and ideas and to later religious concepts such as the gods Inanna and Dumuzi from these more northern regions6,8,47,48,49,50. The knowledge and resources required to produce metallurgical artifacts in the Levant have also been hypothesized to come from the north11,51.
Our finding of genetic discontinuity between the Chalcolithic and Early Bronze Age periods also resonates with aspects of the archeological record marked by dramatic changes in settlement patterns43, large-scale abandonment of sites52,53,54,55, many fewer items with symbolic meaning, and shifts in burial practices, including the disappearance of secondary burial in ossuaries56,57,58,59. This supports the view that profound cultural upheaval, leading to the extinction of populations, was associated with the collapse of the Chalcolithic culture in this region18,60,61,62,63,64.
These ancient DNA results reveal a relatively genetically homogeneous population in Peqi’in. We show that the movements of people within the region of the southern Levant were remarkably dynamic, with some populations, such as the one buried at Peqi’in, being formed in part by exogenous influences. This study also provides a case-study relevant beyond the Levant, showing how combined analysis of genetic and archaeological data can provide rich information about the mechanism of change in past societies.
|
|
|
Post by Admin on Feb 2, 2022 19:56:59 GMT
A genomic snapshot of demographic and cultural dynamism in Upper Mesopotamia during the Neolithic Transition
Abstract Upper Mesopotamia played a key role in the Neolithic Transition in Southwest Asia through marked innovations in symbolism, technology, and foodways. We present thirteen ancient genomes (c.8500-7500 calBCE) from Pre-Pottery Neolithic Çayönü in the Tigris basin together with bioarchaeological and material culture data. Our findings reveal that Çayönü was a genetically diverse population, carrying a mixed ancestry from western and eastern Fertile Crescent, and that the community received immigrants. Our results further suggest that the community was organised along biological family lines. We document bodily interventions such as head-shaping and cauterization among the individuals examined, reflecting Çayönü’s cultural ingenuity. Finally, we identify Upper Mesopotamia as the likely source of eastern gene flow into Neolithic Anatolia, in line with material culture evidence. We hypothesise that Upper Mesopotamia’s cultural dynamism during the Neolithic Transition was the product not only of its fertile lands but also of its interregional demographic connections.
Introduction Located between the Euphrates and Tigris rivers, the hilly flanks of Upper Mesopotamia were home to the earliest sedentary hunter-gatherers who built the first monumental structures at Göbekli Tepe (1) and domesticated numerous local plant and animal species, including einkorn, emmer, sheep, goat, pig, and cattle (2–5). The innovative spirit and cultural dynamism of these societies during the Neolithic Transition in Southwest Asia (c. 9800-6500 BC) is well documented in the archaeological record, but their demographic history and social structures has remained unknown owing to the lack of genomes from North Mesopotamia. This stands in contrast with a significant number of recent archaeogenomic studies that focused on the three most distant corners of Neolithic Southwest Asia, namely South Levant, Central Zagros, and Central Anatolia (Fig. 1A, B) (6–12). This body of work has together revealed (a) genetically distinct populations in all three regions, (b) a dominant trend of population continuity between pre-Neolithic, Pre-Pottery Neolithic (PPN) and Pottery Neolithic (PN) communities, (c) an overlay of interregional gene flow through time, such as inferred “southern” and “eastern” gene flow events into Central Anatolia between the Early and Late Neolithic. Meanwhile, key questions about the possible roles of Upper Mesopotamia in interregional demographic and cultural change, e.g., whether Upper Mesopotamia influenced Late Neolithic Central Anatolia and whether it was the source of the post-Neolithic gene flow into Anatolia (6, 13), have remained open. With the exception of a single ancient DNA study reporting 15 mitochondrial DNA sequences from the Upper Euphrates (14), Upper Mesopotamia has remained genomically unexplored, mostly owing to low DNA preservation in the region. Here we address this gap by studying genomic data from Çayönü Tepesi (hereon Çayönü) of the Upper Tigris area (Fig. 1A), a settlement that presents one of the best examples of the transition from foraging to food production in Southwest Asia (15). Firstly, Çayönü’s uninterrupted stratigraphy extending from the Pre-Pottery Neolithic A (PPNA) (c. 9500 cal BCE) to the final PPN (c. 7000 cal BCE) is unparalleled in the region. Secondly, the Çayönü Neolithic community is recognized for its marked cultural dynamism, which is reflected (a) in evidence for intense plant cultivation (16) and animal management (pig, cattle, sheep and goat) (17), (b) in continuous innovation in architectural styles (Fig. 1C), (c) in technological experimentation, from pioneering lime burning techniques (15, 18) to the production of copper beads and reamer-like objects (19). Finally, both western (Levant-Euphrates) and eastern (Tigris-Zagros) influences and parallel developments are traceable in Çayönü’s material culture (20) (Supplementary Table 1). These observations suggest that Çayönü and contemporaneous Upper Mesopotamian communities could have acted as hubs of cultural interaction and innovation in Neolithic Southwest Asia. Our study presents genomic data from Çayönü, which we then use to describe (i) the structure of Fertile Crescent populations in comparison with interregional material culture affinities, (ii) the Neolithic demographic transition reflected in intraregional genomic diversity, (iii) genetic kinship among co-burials in domestic structures at Çayönü, and (iv) the potential role of Upper Mesopotamia in Neolithic and post-Neolithic human movements influencing Anatolia. We also detail the curious case of a Çayönü infant, whom we infer to be a migrant offspring, and whose skeletal material presents the earliest known examples of cauterization and head-shaping in the region.
Fig. 1. Spatio-temporal distribution of the samples and the population structure of Neolithic Southwest Asia. (A) Timeline of ancient Southwest Asian individuals used in the analyses. Colored rectangles at the bottom represent the sub-periods of the Neolithic Era in Southwest Asia. (B) The map shows Epi-Paleolithic (EP) and Neolithic populations from Southwest Asia. Shaded areas mark Pre-Pottery Neolithic (PPN) cultural zones. (C) Çayönü building types and their approximate dates of use, considered as evidence for Çayönü’s cultural openness and ingenuity. Adapted from (103). (D) First two dimensions of multidimensional scaling (MDS) plot of genetic distances. The MDS summarises the genetic distance matrix among ancient genomes calculated as (1 - outgroup f3) values. Outgroup f3-statistics were calculated as f3(Yoruba; individual1, individual2). The labels represent the following sites: Anatolia EP: Pınarbaşı; Anatolia PPN: Boncuklu and Aşıklı Höyük; Anatolia PN: Çatalhöyük and Barcın Höyük; Levant EP: Natufian; Levant PPN: Ain’ Ghazal, Kfar HaHoresh, Motza and Ba’ja; C Zagros N (Central Zagros Neolithic): Ganj Dareh, Tepe Abdul and Wezmeh Cave; S Caucasus EP (South Caucasus EP): Kotias and Satsurblia. We use “Anatolia” here following the traditional geographic definition, referring to the west of the Anatolian Diagonal.
|
|
|
Post by Admin on Feb 2, 2022 21:15:37 GMT
Results and Discussion We studied a total of 33 human remains from Çayönü (Fig. 1A, B, Supplementary Table 2). These were mainly found as subfloor burials located inside or within the proximity of six PrePottery Neolithic B (PPNB) buildings (Table S1). We screened 33 aDNA libraries by shotgun sequencing, which revealed endogenous DNA proportions varying between 0.04% and 5% (median = 0.2%, Fig. S1). This was lower than aDNA preservation in a contemporaneous Central Anatolian settlement, Aşıklı (median = 1.4%, Wilcoxon rank sum test p < 0.05), but comparable to another Central Anatolian site, Boncuklu (median = 0.1%, Wilcoxon rank sum test p > 0.05, Fig. S2). Libraries from 14 individuals were chosen for deeper sequencing (Methods), from which we generated shotgun genomes with depths ranging from 0.016x to 0.49x. (Fig. S1, Supplementary Table 2). High rates of post-mortem damage (PMD) accumulation at read ends, short average fragment sizes (49-60 bps, median = 51.4 bps), and mitochondrial haplotype-based estimates suggested authenticity of all 14 libraries (Methods) (Supplementary Table 2). With this data we first estimated genetic kinship among all individual pairs (Methods). Two samples, both identified as female infants (cay018 and cay020), were genetically inferred to either belong to the same individual or to be identical twins. Anthropological evaluation also showed that both petrouses could belong to the same individual. We therefore merged their genomic data and treated this merged data as representing a single individual, reducing our sample size to 13 individuals (6 adult females, 2 adult males, 3 sub-adult females, 2 sub-adult males). We further identified four related pairs of 1st to 3rd degree (see below) and removed all but one individual among sets of closely related individuals in population genetic analyses (Methods). The east-west genetic structure of Neolithic Southwest Asia To obtain an overview of genetic affinities among human populations in Neolithic Southwest Asia we compared the 13 Çayönü genomes with published ancient genomes dating to c.15,000- 5,500 BCE from the Fertile Crescent and neighbouring regions (Supplementary Table 3) (6–8, 11, 12, 21–24) using multidimensional scaling (MDS) of pairwise f3 results, D-statistics, and qpAdm analyses (25). These revealed a number of observations. In the MDS analysis, the Çayönü group occupied a distinct and intermediate position within the space of Southwest Asian genetic diversity bordered by early Holocene South Levant, Central Zagros and South Caucasus, and Central Anatolia (Fig. 1D, Fig. S3). Our sample of Çayönü genomes was internally homogeneous within this space, with the exception of an “outlier” individual, cay008, who appeared relatively closer to Zagros/Caucasus individuals. D-statistics likewise showed that the Çayönü group was genetically closer to western Southwest Asia (early Holocene Central Anatolia and South Levant) than to eastern Southwest Asia (Central Zagros) (Fig. 2A, Supplementary Table 4). At the same time, Central Zagros genomes showed higher genetic affinity to our Çayönü sample than to Central Anatolia or South Levant (Fig. 2B). Finally, we found that cay008 harbours higher Zagros contribution than other Çayönü individuals (Fig. 1D, Fig. 2C, Supplementary Table 5). Given these observations, we first investigated the origins of the genetic structure in Neolithic Southwest Asia. The higher genetic affinity among Upper Mesopotamia (represented by Çayönü), Central Anatolia, and South Levant populations relative to Central Zagros (Fig. 1D, Fig. S3) was intriguing, which led us to ask whether this affinity could be explained by an isolation-by-distance process (26, 27). We computed shared genetic drift between each pair of individuals in our Southwest Asia sample and compared these to geodesic geographic distance among settlements (Methods). To eliminate the effect of temporal genetic changes, we only included individual pairs separated by <1,000 years. We found a general correlation between spatial and genetic distances, as expected (Fig. 3A). However, we also found that Central Zagros genomes were significantly more differentiated compared to that expected from a linear isolation-by-distance model (Fig. 3B). We therefore infer a stark west-east genetic structure in the Fertile Crescent, where the lowest effective migration (28) appears to lie between Upper Mesopotamia and Central Zagros. At face value, this result may seem to imply resistance to gene flow between Upper Mesopotamia and Central Zagros during the Neolithic. However, such resistance does not align with observed material culture affinities between the two regions (e.g., (29) Supplementary Table 1). We therefore suggest an alternative scenario to explain the observed genetic structure. During Last Glacial Maximum (LGM), the eastern regions of Southwest Asia (the ancestors of Central Zagros / South Caucasus) could have been isolated from the western regions (the ancestors of Central Anatolians/Levantines), with the east and west populations differentiating through drift or by admixture with third populations. Sometime after the LGM, these eastern and western regional populations could have re-expanded and partly admixed within Southwest Asia. It is plausible that east-west admixture occurred in Upper Mesopotamia, giving rise to Çayönü’s gene pool, and may have also influenced Central Anatolia by the PN (6). We note that the duration and timing of this putative admixture process remain unclear, and that alternative scenarios are also conceivable (30). Irrespective of the demographic mechanisms, though, Central Zagros appears to have been genetically the most distinct group in early Holocene Southwest Asia.
Fig. 2. Genetic affinities of Çayönü population with the neighbouring populations. Formal tests computed in the form of (A) D(Yoruba, Çayönü/cay008; pop2, test) (B) D(Yoruba, pop1; Çayönü/cay008, Anatolia EP/PPN/PN). Z-scores were corrected with the Benjamini-Hochberg multiple testing correction (82). Horizontal bars represent ±2 standard errors. (C) qpAdm modelling of the Çayönü group and cay008. The “local” Çayönü group or outlier cay008 individual was the “target”; Central Anatolia Epi-Paleolithic, Central Zagros Neolithic and South Levant Neolithic samples were sources for both targets. The “local” Çayönü group was also used as “source” for modelling of cay008. Horizontal bars represent standard errors of the coefficients. All three models yielded p-values > 0.05. We also cannot reject a three-way model of Central Anatolia PPN, Central Zagros and South Levant at >0.01 p-value threshold (Supplemental Table 5). In all analyses shown in the figure, “Çayönü” represents the 9 genomes listed in Table 1, excluding relatives and cay008.
|
|