Tweet
Ancient DNA from Mesopotamia suggests distinct Pre-Pottery and Pottery Neolithic migrations into Anatolia
We present the first ancient DNA data from the Pre-Pottery Neolithic of Mesopotamia (Southeastern Turkey and Northern Iraq), Cyprus, and the Northwestern Zagros, along with the first data from Neolithic Armenia. We show that these and neighboring populations were formed through admixture of pre-Neolithic sources related to Anatolian, Caucasus, and Levantine hunter-gatherers, forming a Neolithic continuum of ancestry mirroring the geography of West Asia. By analyzing Pre-Pottery and Pottery Neolithic populations of Anatolia, we show that the former were derived from admixture between Mesopotamian-related and local Epipaleolithic-related sources, but the latter experienced additional Levantine-related gene flow, thus documenting at least two pulses of migration from the Fertile Crescent heartland to the early farmers of Anatolia.
[...]
Pre-Pottery Neolithic Cyprus (CYP_PPNB; this study)
The PPN population of Cyprus can be modeled as a clade with that of NW and Central Anatolia (Table S1), but with a positive coefficient of ancestry from Levantine (PPN or Natufian) sources when considering two-way models. This pattern is also evident in the main 5- way model of (12). Thus, while Cyprus is geographically intermediate between the Levant and Anatolia, its PPN population is more related to that of Anatolia than the Levant, although the pattern seems possibly consistent with ancestry from both sources. However, when modeling the PPN of Cyprus as a mixture of the AşıklıHöyük PPN population from Central Anatolia and Levantine PPN, only a non-significant 6.8±4.2% of the latter is estimated, so overall, we can say that the PPN of Cyprus resembled that of the PPN of the Anatolian mainland to the north.
[...]
Levant_PPN(1)
A single source for the Levantine Neolithic (from Jordan and Israel) can be firmly rejected (p<1e-12), but a total of six 2-way models work well, all of which involve an Anatolian Neolithic or related source and the Epipaleolithic Natufians.(1) Interestingly, the model in which the source for the Levantine Neolithic is Mesopotamian rather Anatolian is firmly rejected as well (p=5e-6). In a previous publication(1), the Levantine PPN population was modeled in a similar 2-way fashion but there was an open question whether a more intermediate population (such as Mesopotamian) might be contributing to the Levant. Our results prove that this was not the case by rejecting a contribution from Mesopotamian populations of the Tigris region. As we will now see in the modeling of the population of Mesopotamia, this can be explained by the presence of CHG-related ancestry in Mesopotamia.
[...]
Neolithic North Mesopotamia (Mesopotamia_PPN; this study)
The Northern Mesopotamian population includes two individuals from Nemrik9 in Iraq and an individual from Mardin (Boncuklu Tarla) in SE Turkey; the two are co-analyzed given their geographical and temporal proximity and clustering in PCA. A one-way model fits them with Aknashen (p=0.894), Masis Blur (p=0.103), or Azerbaijan Neolithic (p=0.340)as a source (Table S1). Given that the Mesopotamian samples predate these 6th millennium BCE ones, the Mesopotamian samples are unlikely to be derived from those of the South Caucasus, and these models merely point to the overall similarity of North Mesopotamian with South Caucasus populations. To determine the origins of the Mesopotamian PPN population we also examined 2-way models (Table S2) and find that a model in which Mesopotamian samples are a mix of Anatolian PPN Neolithic (AşıklıHöyük) and Ganj Dareh(p=0.16) and also as a mix of Cyprus PPN and Ganj Dareh (p=0.194); we remarked above on the similarity between the Cyprus PPN and the Central Anatolian PPN, so it makes sense that its population can also be used as a genetic source stand-in although this scenario is less plausible geographically. North Mesopotamia is geographically intermediate between Anatolia and Iran, so this model is more plausible, although by no means definite, given the uncertainty about the genetic composition of PPN populations elsewhere in the ancient Near East between Central Anatolia and the Zagros. We note that an Anatolian Neolithic+CHG model fails (p<1e-9) as does a Levantine+Ganj Dareh one (p<1e-9), as does a Barcın Anatolian Neolithic+Ganj Dareh one (p=0.002), as does a Levantine Neolithic+CHG one (p=0.008) thus excluding these other possible pairwise combinations for which Mesopotamia is geographically intermediate. The LBA individual from Nemrik9 fits as a clade with the Mesopotamian PPN (p=0.27) and with Azerbaijan Neolithic (p=0.25) and with Masis Blur (p=0.026) providing some evidence for long-term continuity of the population of Mesopotamia and its environs, a pattern that must be verified with more individuals at both ends of the time span between the PPN period and the Late Bronze Age. Given that it is of higher data quality than the other samples from Mesopotamia we group it with them in a Mesopotamian meta-population which we use in our analyses of the rest of the paper.
[...]
Neolithic Anatolia (TUR_Marmara_Barcın_N(2), TUR_C_Boncuklu_PPN_Feldman(6), TUR_C_AşıklıHöyük_PPN(8), TUR_C_Çatalhöyük_N(8))
Finally, we turn to Neolithic Central/Western Anatolia; we remarked above on the similarity between it and the PPN population of Cyprus, and this is the only population that can be used as a source for the Neolithic NW Anatolia; however, it would seem unlikely that the Neolithic population of the geographically large region of Anatolia would be descended from that of the small island to its south, especially since the 2-way models involve the Epipaleolithic sample from Pınarbaşıas one of the two sources, suggesting a rooting of Neolithic Anatolian populations on local (pre-Neolithic) hunter-gatherers of the region. The pottery Neolithic from Çatalhöyük are marginally consistent with forming a clade with the PPN of AşıklıHöyük (p=0.03), but other than that, no pairs of Anatolian Neolithic populations form a clade, suggesting that the PPN and pottery Neolithic groups inhabiting the large Anatolian peninsula did not form a single homogeneous population. Moreover, while Pınarbaşı features prominently in 2-way models for Anatolian Neolithic populations, it does not form a clade with any of them by itself: thus, a general conclusion can be drawn that Anatolian Neolithic populations were descended from local hunter-gatherers but not exclusively so. For the two PPN populations from Boncuklu and Aşıklı Höyük who both inhabited Central Anatolia during, models of Pınarbaşı+ Mesopotamian Neolithic fit with contrasting amounts of Mesopotamian ancestry in the two sites (~26% low at Boncuklu vs. ~70% high at Aşıklı Höyük). The results for point to gene flow from the east (represented by Mesopotamia_PPN) into Anatolia, a conclusion that makes sense given the earlier appearance of agriculture in SE Anatolia and N Mesopotamia and its subsequent spread westward. But, gene flow was potentially bidirectional, as we have seen that Neolithic populations of the South Caucasus can be modeled as admixtures of Anatolian populations with CHG and the Mesopotamian population could be modeled as a mixture of Ganj Dareh +AşıklıHöyük. Thus, in both locales (Central Anatolia and the South Caucasus) there is persistence of the local hunter-gatherer ancestry (Pınarbaşıand CHG respectively) but with some influence from the other (CHG/Mesopotamian in Anatolia and Anatolian in the South Caucasus). The results on the Anatolian PPN add another cline to the picture: between Anatolia and Mesopotamia in which PPN samples from Boncuklu in Anatolia and Nemrik 9 and Mardin in Mesopotamia occupied opposite ends, with Aşıklı Höyük being intermediate. We cannot speak of persistence of hunter-gatherer ancestry in Mesopotamia itself, as we lack pre-Neolithic individuals; it is possible that the population there was also a Neolithic mixture (as suggested by the 2-way admixture models that fit it), but admixture in this population may also precede the advent of the Neolithic. For the two pottery Neolithic populations, we observe that the Çatalhöyük population from Central Anatolia can be modeled as mostly (~96%) of Aşıklı Höyük origin (+Pınarbaşı) suggesting continuity with the pre-pottery Neolithic population of the same region. We note, however, that while we were able to model the Aşıklı Höyük population as a mixture of Pınarbaşı and Mesopotamian PPN (p=0.25), the same model fails for Çatalhöyük(p=0.002); we investigate the cause of this disparity below. As for the other pottery Neolithic population (Barcın from NW Anatolia), we can model it well as a mixture of about half Çatalhöyük and half Pınarbaşı ancestry. Thus, the data are consistent with this population from the NW end of the Anatolian peninsula being derived from pottery Neolithic populations of Central Anatolia (closer to the centers of domestication of the Fertile Crescent) but with absorption of Epipaleolithic hunter-gatherer ancestry.
A common 3-way model for the Neolithic Near East
To better understand the origin of Neolithic Near Eastern populations, many of which could be modeled as mixtures of each other we considered them in terms of the 3-way model (CHG+Pınarbaşı+ Natufians) that allow us to compare them using a common set of sources representing the earliest (and pre-Neolithic) populations in the South Caucasus, Anatolia, and the Levant. Thus this model describes the studied populations in terms of ancient Near Eastern populations predating the emergence of agriculture. The results of this model are shown in Table S3. We summarize them below:
(1)Many of the populations of Table S3require ancestry from all 3 sources of the model. Exceptions are the Levantine PPN for which the contribution of the CHG is consistent with zero and the farmers from the Zagros (Bestansur, Shanidar, and Ganj Dareh) for which the contribution of Pınarbaşı is consistent with zero. The PPN farmers from Boncuklu inCentral Anatolia have a small and not significantly positive fraction of Natufian ancestry (4.8±6.8%), and this jumps to 40.3±12.1% at Aşıklı Höyük and 51.3±5% at Çatalhöyük. Importantly, none of them have ancestry only from a single pre-Neolithic source.(2) Anatolian Epipaleolithic ancestry (represented by Pınarbaşı) makes up a major source of ancestry of Neolithic farmers from Central/NW Anatolia; Levantine Epipaleolithic ancestry (represented by Natufians) makes up the major source of ancestry of PPN farmers from the Levant; CHG ancestry is represented across the West Asian highlands but diminishes towards the west (in Anatolia) and the South (into the Levant) (3)The PPN population of Mesopotamia is intermediate in the context of these three Epipaleolithic sources of ancestry and is modeled as deriving its ancestry from all three. Two possibilities are raised: first, that the PPN population was indeed deeply admixed as its geographically intermediate position would suggest, or, second, that there is another unsampled hunter-gatherer population in the Mesopotamian region from which the PPN population is descended. As an analogy, prior to the sampling of Pınarbaşı(6)the Anatolian Neolithic population could be modeled as a 3-way mixture involving Levantine Neolithic, WHG, and CHG sources(1), but now can be modeled with Pınarbaşı as its major source on top of which Mesopotamian Neolithic ancestry was added to various degrees (4)The PPN population of Cyprus and the Neolithic populations of the South Caucasus (Armenia and Azerbaijan) represent similar mixtures as the PPN population of Mesopotamia but with varying proportions of the three components. Cyprus is similar to Anatolia in its lack of substantial CHG ancestry, but its Pınarbaşı/Natufian balance tilted towards the latter, explaining the results of 2-way modeling above. The South Caucasus is similar to Mesopotamia, but with an excess of Pınarbaşı-related ancestry.These observations are presented visually in Fig.2.
Caucasus hunter-gatherer vs. Iranian Neolithic ancestry
CHG and IRN_Ganj_Dareh_Noften appear as interchangeable sources for Neolithic populations and we wanted to determine if we can differentiate between the two. We thus fit the two 3-way models: (CHG or IRN_Ganj_Dareh_N) +Pınarbaşı+Natufians, i.e., setting one of the two populations as a source and including the other in the right set of outgroups (Table S4). For many Test populations both models cannot be rejected, and so for them the “Caucasus-Iran” influence does not clearly stem from either the CHG or Neolithic Iran. For two populations (AZE_N and Mesopotamia_PPN) which neighbor both Iran and the South Caucasus both models are rejected, suggesting that while these populations clearly have some “Caucasus-Iran”-related ancestry as suggested by the modeling of Table S2, they cannot be modeled with only one of the two to the exclusion of the other: a possible interpretation is that these geographically intermediate populations possess ancestry related to both their Caucasus and Iran neighbors. For the Barcın Neolithic, the CHG model is rejected (1.13E-03) while the Iran one narrowly accepted (0.0142). This population can be modeled to derive some of its ancestry from the east of NW Anatolia, and thus potentially from both CHG-and Iran-related sources. Finally, for the South Caucasus Neolithic at Aknashen, the CHG model is not rejected
(p=0.46) while that with Iran as a source is (p=3.23E-04), suggesting that in the South Caucasus the Neolithic population can be modeled with CHG ancestry alone.
Levantine vs. Mesopotamian influence in Anatolia
Anatolian Neolithic populations cannot be modeled with only ancestry from Pınarbaşı (Epipaleolithic Central Anatolia), but also require Natufian and CHG ancestry (pre-Neolithic sources from the Levant and Caucasus) (Table S3). In order to identify the proximate source of this ancestry we examined a 3-way model Levant_PPN+Mesopotamia_PPN+Pınarbaşı (Table S5). This model reveals that for the PPN populations from Central Anatolia (Boncuklu and AşıklıHöyük) Mesopotamian PPN ancestry alone can be added to the Pınarbaşı substratum, with the two sites contrasting in their proportions of the two components, with Boncuklu derived more from the substratum and AşıklıHöyük more from a Mesopotamian source. By contrast, both pottery Neolithic populations (Barcınand Çatalhöyük) from NW and Central Anatolia are modeled not only with Mesopotamian ancestry but with a statistically significant proportions of~13-17% Levantine PPN ancestry. (One of the individuals from AşıklıHöyük (Ash133.SG) is a PCA outlier, Fig. 2, clustering with Çatalhöyük individuals. Out of caution, we refit the model of Table S5excluding this individual and obtained an estimate of-1.0±9.8% Levant_PPN, 68.1±13.9% Mesopotamia_PPN, and 32.9±12.5% CHG ancestry, similar to those of Table S5.) An interpretation of these results is that the population of Anatolia experienced an influx from North Mesopotamia during the spread of the Pre-Pottery Neolithic while the later spread of the pottery Neolithic was derived from a slightly different population that was intermediate between the sampled Levantine and Mesopotamian PPN populations. Future sampling of the first pottery Neolithic populations of the Near East may reveal whether these could be a source that could account for the westward spread of the extra Levantine ancestry westward into Anatolia.
https://www.science.org/doi/10.1126/science.abq0762
file:///C:/Users/18509/Downloads/science.abq0762_sm.pdf
We present the first ancient DNA data from the Pre-Pottery Neolithic of Mesopotamia (Southeastern Turkey and Northern Iraq), Cyprus, and the Northwestern Zagros, along with the first data from Neolithic Armenia. We show that these and neighboring populations were formed through admixture of pre-Neolithic sources related to Anatolian, Caucasus, and Levantine hunter-gatherers, forming a Neolithic continuum of ancestry mirroring the geography of West Asia. By analyzing Pre-Pottery and Pottery Neolithic populations of Anatolia, we show that the former were derived from admixture between Mesopotamian-related and local Epipaleolithic-related sources, but the latter experienced additional Levantine-related gene flow, thus documenting at least two pulses of migration from the Fertile Crescent heartland to the early farmers of Anatolia.
[...]
Pre-Pottery Neolithic Cyprus (CYP_PPNB; this study)
The PPN population of Cyprus can be modeled as a clade with that of NW and Central Anatolia (Table S1), but with a positive coefficient of ancestry from Levantine (PPN or Natufian) sources when considering two-way models. This pattern is also evident in the main 5- way model of (12). Thus, while Cyprus is geographically intermediate between the Levant and Anatolia, its PPN population is more related to that of Anatolia than the Levant, although the pattern seems possibly consistent with ancestry from both sources. However, when modeling the PPN of Cyprus as a mixture of the AşıklıHöyük PPN population from Central Anatolia and Levantine PPN, only a non-significant 6.8±4.2% of the latter is estimated, so overall, we can say that the PPN of Cyprus resembled that of the PPN of the Anatolian mainland to the north.
[...]
Levant_PPN(1)
A single source for the Levantine Neolithic (from Jordan and Israel) can be firmly rejected (p<1e-12), but a total of six 2-way models work well, all of which involve an Anatolian Neolithic or related source and the Epipaleolithic Natufians.(1) Interestingly, the model in which the source for the Levantine Neolithic is Mesopotamian rather Anatolian is firmly rejected as well (p=5e-6). In a previous publication(1), the Levantine PPN population was modeled in a similar 2-way fashion but there was an open question whether a more intermediate population (such as Mesopotamian) might be contributing to the Levant. Our results prove that this was not the case by rejecting a contribution from Mesopotamian populations of the Tigris region. As we will now see in the modeling of the population of Mesopotamia, this can be explained by the presence of CHG-related ancestry in Mesopotamia.
[...]
Neolithic North Mesopotamia (Mesopotamia_PPN; this study)
The Northern Mesopotamian population includes two individuals from Nemrik9 in Iraq and an individual from Mardin (Boncuklu Tarla) in SE Turkey; the two are co-analyzed given their geographical and temporal proximity and clustering in PCA. A one-way model fits them with Aknashen (p=0.894), Masis Blur (p=0.103), or Azerbaijan Neolithic (p=0.340)as a source (Table S1). Given that the Mesopotamian samples predate these 6th millennium BCE ones, the Mesopotamian samples are unlikely to be derived from those of the South Caucasus, and these models merely point to the overall similarity of North Mesopotamian with South Caucasus populations. To determine the origins of the Mesopotamian PPN population we also examined 2-way models (Table S2) and find that a model in which Mesopotamian samples are a mix of Anatolian PPN Neolithic (AşıklıHöyük) and Ganj Dareh(p=0.16) and also as a mix of Cyprus PPN and Ganj Dareh (p=0.194); we remarked above on the similarity between the Cyprus PPN and the Central Anatolian PPN, so it makes sense that its population can also be used as a genetic source stand-in although this scenario is less plausible geographically. North Mesopotamia is geographically intermediate between Anatolia and Iran, so this model is more plausible, although by no means definite, given the uncertainty about the genetic composition of PPN populations elsewhere in the ancient Near East between Central Anatolia and the Zagros. We note that an Anatolian Neolithic+CHG model fails (p<1e-9) as does a Levantine+Ganj Dareh one (p<1e-9), as does a Barcın Anatolian Neolithic+Ganj Dareh one (p=0.002), as does a Levantine Neolithic+CHG one (p=0.008) thus excluding these other possible pairwise combinations for which Mesopotamia is geographically intermediate. The LBA individual from Nemrik9 fits as a clade with the Mesopotamian PPN (p=0.27) and with Azerbaijan Neolithic (p=0.25) and with Masis Blur (p=0.026) providing some evidence for long-term continuity of the population of Mesopotamia and its environs, a pattern that must be verified with more individuals at both ends of the time span between the PPN period and the Late Bronze Age. Given that it is of higher data quality than the other samples from Mesopotamia we group it with them in a Mesopotamian meta-population which we use in our analyses of the rest of the paper.
[...]
Neolithic Anatolia (TUR_Marmara_Barcın_N(2), TUR_C_Boncuklu_PPN_Feldman(6), TUR_C_AşıklıHöyük_PPN(8), TUR_C_Çatalhöyük_N(8))
Finally, we turn to Neolithic Central/Western Anatolia; we remarked above on the similarity between it and the PPN population of Cyprus, and this is the only population that can be used as a source for the Neolithic NW Anatolia; however, it would seem unlikely that the Neolithic population of the geographically large region of Anatolia would be descended from that of the small island to its south, especially since the 2-way models involve the Epipaleolithic sample from Pınarbaşıas one of the two sources, suggesting a rooting of Neolithic Anatolian populations on local (pre-Neolithic) hunter-gatherers of the region. The pottery Neolithic from Çatalhöyük are marginally consistent with forming a clade with the PPN of AşıklıHöyük (p=0.03), but other than that, no pairs of Anatolian Neolithic populations form a clade, suggesting that the PPN and pottery Neolithic groups inhabiting the large Anatolian peninsula did not form a single homogeneous population. Moreover, while Pınarbaşı features prominently in 2-way models for Anatolian Neolithic populations, it does not form a clade with any of them by itself: thus, a general conclusion can be drawn that Anatolian Neolithic populations were descended from local hunter-gatherers but not exclusively so. For the two PPN populations from Boncuklu and Aşıklı Höyük who both inhabited Central Anatolia during, models of Pınarbaşı+ Mesopotamian Neolithic fit with contrasting amounts of Mesopotamian ancestry in the two sites (~26% low at Boncuklu vs. ~70% high at Aşıklı Höyük). The results for point to gene flow from the east (represented by Mesopotamia_PPN) into Anatolia, a conclusion that makes sense given the earlier appearance of agriculture in SE Anatolia and N Mesopotamia and its subsequent spread westward. But, gene flow was potentially bidirectional, as we have seen that Neolithic populations of the South Caucasus can be modeled as admixtures of Anatolian populations with CHG and the Mesopotamian population could be modeled as a mixture of Ganj Dareh +AşıklıHöyük. Thus, in both locales (Central Anatolia and the South Caucasus) there is persistence of the local hunter-gatherer ancestry (Pınarbaşıand CHG respectively) but with some influence from the other (CHG/Mesopotamian in Anatolia and Anatolian in the South Caucasus). The results on the Anatolian PPN add another cline to the picture: between Anatolia and Mesopotamia in which PPN samples from Boncuklu in Anatolia and Nemrik 9 and Mardin in Mesopotamia occupied opposite ends, with Aşıklı Höyük being intermediate. We cannot speak of persistence of hunter-gatherer ancestry in Mesopotamia itself, as we lack pre-Neolithic individuals; it is possible that the population there was also a Neolithic mixture (as suggested by the 2-way admixture models that fit it), but admixture in this population may also precede the advent of the Neolithic. For the two pottery Neolithic populations, we observe that the Çatalhöyük population from Central Anatolia can be modeled as mostly (~96%) of Aşıklı Höyük origin (+Pınarbaşı) suggesting continuity with the pre-pottery Neolithic population of the same region. We note, however, that while we were able to model the Aşıklı Höyük population as a mixture of Pınarbaşı and Mesopotamian PPN (p=0.25), the same model fails for Çatalhöyük(p=0.002); we investigate the cause of this disparity below. As for the other pottery Neolithic population (Barcın from NW Anatolia), we can model it well as a mixture of about half Çatalhöyük and half Pınarbaşı ancestry. Thus, the data are consistent with this population from the NW end of the Anatolian peninsula being derived from pottery Neolithic populations of Central Anatolia (closer to the centers of domestication of the Fertile Crescent) but with absorption of Epipaleolithic hunter-gatherer ancestry.
A common 3-way model for the Neolithic Near East
To better understand the origin of Neolithic Near Eastern populations, many of which could be modeled as mixtures of each other we considered them in terms of the 3-way model (CHG+Pınarbaşı+ Natufians) that allow us to compare them using a common set of sources representing the earliest (and pre-Neolithic) populations in the South Caucasus, Anatolia, and the Levant. Thus this model describes the studied populations in terms of ancient Near Eastern populations predating the emergence of agriculture. The results of this model are shown in Table S3. We summarize them below:
(1)Many of the populations of Table S3require ancestry from all 3 sources of the model. Exceptions are the Levantine PPN for which the contribution of the CHG is consistent with zero and the farmers from the Zagros (Bestansur, Shanidar, and Ganj Dareh) for which the contribution of Pınarbaşı is consistent with zero. The PPN farmers from Boncuklu inCentral Anatolia have a small and not significantly positive fraction of Natufian ancestry (4.8±6.8%), and this jumps to 40.3±12.1% at Aşıklı Höyük and 51.3±5% at Çatalhöyük. Importantly, none of them have ancestry only from a single pre-Neolithic source.(2) Anatolian Epipaleolithic ancestry (represented by Pınarbaşı) makes up a major source of ancestry of Neolithic farmers from Central/NW Anatolia; Levantine Epipaleolithic ancestry (represented by Natufians) makes up the major source of ancestry of PPN farmers from the Levant; CHG ancestry is represented across the West Asian highlands but diminishes towards the west (in Anatolia) and the South (into the Levant) (3)The PPN population of Mesopotamia is intermediate in the context of these three Epipaleolithic sources of ancestry and is modeled as deriving its ancestry from all three. Two possibilities are raised: first, that the PPN population was indeed deeply admixed as its geographically intermediate position would suggest, or, second, that there is another unsampled hunter-gatherer population in the Mesopotamian region from which the PPN population is descended. As an analogy, prior to the sampling of Pınarbaşı(6)the Anatolian Neolithic population could be modeled as a 3-way mixture involving Levantine Neolithic, WHG, and CHG sources(1), but now can be modeled with Pınarbaşı as its major source on top of which Mesopotamian Neolithic ancestry was added to various degrees (4)The PPN population of Cyprus and the Neolithic populations of the South Caucasus (Armenia and Azerbaijan) represent similar mixtures as the PPN population of Mesopotamia but with varying proportions of the three components. Cyprus is similar to Anatolia in its lack of substantial CHG ancestry, but its Pınarbaşı/Natufian balance tilted towards the latter, explaining the results of 2-way modeling above. The South Caucasus is similar to Mesopotamia, but with an excess of Pınarbaşı-related ancestry.These observations are presented visually in Fig.2.
Caucasus hunter-gatherer vs. Iranian Neolithic ancestry
CHG and IRN_Ganj_Dareh_Noften appear as interchangeable sources for Neolithic populations and we wanted to determine if we can differentiate between the two. We thus fit the two 3-way models: (CHG or IRN_Ganj_Dareh_N) +Pınarbaşı+Natufians, i.e., setting one of the two populations as a source and including the other in the right set of outgroups (Table S4). For many Test populations both models cannot be rejected, and so for them the “Caucasus-Iran” influence does not clearly stem from either the CHG or Neolithic Iran. For two populations (AZE_N and Mesopotamia_PPN) which neighbor both Iran and the South Caucasus both models are rejected, suggesting that while these populations clearly have some “Caucasus-Iran”-related ancestry as suggested by the modeling of Table S2, they cannot be modeled with only one of the two to the exclusion of the other: a possible interpretation is that these geographically intermediate populations possess ancestry related to both their Caucasus and Iran neighbors. For the Barcın Neolithic, the CHG model is rejected (1.13E-03) while the Iran one narrowly accepted (0.0142). This population can be modeled to derive some of its ancestry from the east of NW Anatolia, and thus potentially from both CHG-and Iran-related sources. Finally, for the South Caucasus Neolithic at Aknashen, the CHG model is not rejected
(p=0.46) while that with Iran as a source is (p=3.23E-04), suggesting that in the South Caucasus the Neolithic population can be modeled with CHG ancestry alone.
Levantine vs. Mesopotamian influence in Anatolia
Anatolian Neolithic populations cannot be modeled with only ancestry from Pınarbaşı (Epipaleolithic Central Anatolia), but also require Natufian and CHG ancestry (pre-Neolithic sources from the Levant and Caucasus) (Table S3). In order to identify the proximate source of this ancestry we examined a 3-way model Levant_PPN+Mesopotamia_PPN+Pınarbaşı (Table S5). This model reveals that for the PPN populations from Central Anatolia (Boncuklu and AşıklıHöyük) Mesopotamian PPN ancestry alone can be added to the Pınarbaşı substratum, with the two sites contrasting in their proportions of the two components, with Boncuklu derived more from the substratum and AşıklıHöyük more from a Mesopotamian source. By contrast, both pottery Neolithic populations (Barcınand Çatalhöyük) from NW and Central Anatolia are modeled not only with Mesopotamian ancestry but with a statistically significant proportions of~13-17% Levantine PPN ancestry. (One of the individuals from AşıklıHöyük (Ash133.SG) is a PCA outlier, Fig. 2, clustering with Çatalhöyük individuals. Out of caution, we refit the model of Table S5excluding this individual and obtained an estimate of-1.0±9.8% Levant_PPN, 68.1±13.9% Mesopotamia_PPN, and 32.9±12.5% CHG ancestry, similar to those of Table S5.) An interpretation of these results is that the population of Anatolia experienced an influx from North Mesopotamia during the spread of the Pre-Pottery Neolithic while the later spread of the pottery Neolithic was derived from a slightly different population that was intermediate between the sampled Levantine and Mesopotamian PPN populations. Future sampling of the first pottery Neolithic populations of the Near East may reveal whether these could be a source that could account for the westward spread of the extra Levantine ancestry westward into Anatolia.
https://www.science.org/doi/10.1126/science.abq0762
file:///C:/Users/18509/Downloads/science.abq0762_sm.pdf