Abstract
The Caucasus and surrounding areas, with their rich metal resources, became a crucible of the Bronze Age1 and the birthplace of the earliest steppe pastoralist societies2. Yet, despite this region having a large influence on the subsequent development of Europe and Asia, questions remain regarding its hunter-gatherer past and its formation of expansionist mobile steppe societies3,4,5. Here we present new genome-wide data for 131 individuals from 38 archaeological sites spanning 6,000 years. We find a strong genetic differentiation between populations north and south of the Caucasus mountains during the Mesolithic, with Eastern hunter-gatherer ancestry4,6 in the north, and a distinct Caucasus hunter-gatherer ancestry7 with increasing East Anatolian farmer admixture in the south. During the subsequent Eneolithic period, we observe the formation of the characteristic West Eurasian steppe ancestry and heightened interaction between the mountain and steppe regions, facilitated by technological developments of the Maykop cultural complex8. By contrast, the peak of pastoralist activities and territorial expansions during the Early and Middle Bronze Age is characterized by long-term genetic stability. The Late Bronze Age marks another period of gene flow from multiple distinct sources that coincides with a decline of steppe cultures, followed by a transformation and absorption of the steppe ancestry into highland populations.
Similar content being viewed by others
Main
The Caucasus region and surrounding areas lie at the interface of Europe and Asia. By the mid-Holocene, the Greater Caucasus Mountain range functioned as a semipermeable barrier through which ideas, technologies, languages and people moved1. The wide variety of climate zones in the topographically complex South Caucasus supported a high level of biodiversity, whereas the mountain highlands and hilly piedmont zones in the North Caucasus transitioned into the flat open grasslands of the West Eurasian steppe belt9. With its diverse ecologies and rich metal resources, the Caucasus region became a crucible of the Bronze Age (BA) and the birthplace of the earliest steppe pastoralist societies during the fourth millenniumâbc (ref.â2). The subsequent continental expansions of these steppe pastoralist groups over the next two millennia ultimately reshaped the genetic make-up, languages and cultural trajectories of much of Eurasia10,11. However, their emergence out of local hunter-gatherer groups and connections to nascent farming communities in the Fertile Crescent remain poorly understood, as does their ultimate disappearance in the second millenniumâbc.
Genetic structure
We report new genome-wide data for 131 individuals from 38 archaeological sites and 84 new radiocarbon dates across and around the Caucasus region, including the piedmont and steppe zones, tripling the available genomic data (Fig. 1 and Supplementary Tables 1 and 2). The genetic time transect covers about 6,000 years, ranging from the Mesolithic and Neolithic (seventh and sixth millenniaâbc, nâ=â7), Eneolithic (fifth millenniumâbc, nâ=â11), Late Eneolithic and Early BA (EBA; fourth millenniumâbc, nâ=â20), EBA and Middle BA (MBA; third millenniumâbc, nâ=â51), to the final MBA and Late BA (LBA; second millenniumâbc, nâ=â42; Supplementary Table 1). Individuals (nâ=â26) who did not meet the quality criteria were excluded (Methods). The final dataset for population genetic analyses included 102 unrelated individuals, who were combined with published ancient and modern-day individuals (Supplementary Tables 3â5).
a, Map of the wider Caucasus region showing the locations of the 38 sampled sites and 131 individuals from this study and published data5 (Supplementary Tables 1 and 5). The number in parentheses indicates the studied individuals per site. The inset shows a magnified view of the area marked by the rectangle in the main image. Scale bars, 100âkm (main image) and 10âkm (inset). b, Timeline of the median 14C ages of newly reported (black outline) and published individuals (no outline), separated on the xâaxis into main genetic and geographic clusters (Steppe, Intermediate and Caucasus), as well as other published individuals from south of the Greater Caucasus for comparison. Colours in the piecharts (a), and colours and symbols in the timeline (b) represent sampled individuals associated with different archaeological cultural complexes, and these are used consistently across all figures (see Figs. 2 and 3, Extended Data Fig. 1, Supplementary Table 1 and Supplementary Fig. 1 for full site names and individual symbols). Grey bars correspond to the chronological chapters in the results section. The map was generated using Base Relief: Mapzen, OpenStreetMap, and rivers, lakes and borders were added using free vector and raster map data from Natural Earth (https://www.naturalearthdata.com). OpenStreetMap is open data, licensed under the Open Data Commons Open Database Licence by the OpenStreetMap Foundation.
First we carried out principal component analysis (PCA) and ADMIXTURE analysis to qualitatively assess the genetic affinities of the ancient individuals (Methods and Extended Data Figs. 1 and 2). We substantiate the genetic differentiation observed between the steppe and mountain groups5, hereafter termed the Steppe and Caucasus clusters, and describe the formation and persistence of ancestries, including mixed Intermediate groups, which reflect dynamic phases of biological and cultural interaction resulting in the establishment and spread of pastoralism, first in the PonticâCaspian and subsequently in the entire Eurasian steppe zone (Fig. 1). Finally, during the LBA period, we observe the dissolution of the main BA ancestry clusters and the formation of the ancestry found today in the people(s) of the North Caucasus region12,13.
The MesolithicâNeolithic transition
The oldest individuals in this study are from Satanaj cave in Russia (SJG001, 6221â6082âcal bc), and from the early Neolithic site of Arukhlo in Georgia (5885â5476âcal bc, nâ=â4; Fig. 1). SJG001 predates the arrival of the Near Eastern Neolithic into the Caucasus and overlaps with Eastern European hunter-gatherers (EHGs) in PC space, despite being geographically close (about 50âkm) to Caucasus hunter-gatherer (CHG) sites in the South Caucasus, whose individuals carry a different genetic ancestry profile7 (Fig. 2a). We find that SJG001 and EHG individuals from Karelia form a clade with respect to CHG individuals and other test populations (Extended Data Fig. 3a and Supplementary Tables 6 and 7). Using formal ancestry modelling with qpAdm, we were able to successfully model SJG001 with either Karelia_EHG or Sidelkino_EHG ancestry as a single source (Fig. 2c and Supplementary Table 8).
a,b, PCA of newly produced ancient individuals (with outline) and individuals from previous publications (no outline) from the seventh and fifth millenniumâbc (a) and from the fourth millenniumâbc (b), projected onto 102 modern-day populations (grey dots). The dashed arrows in a represent the observed admixture clines between central Anatolian Neolithic and CHG (brown), between CHG and EHG (light green), and between Eneolithic_Caucasus and EHG ancestry-carrying Steppe groups (dark green), respectively. The dashed pink arrow in b represents an observed cline of mixture between Maykop-associated Caucasus groups and those carrying WSHG ancestry in the steppe. The corresponding labels and groupings are listed in Supplementary Table 5. c,d, Sankey diagram of genetic ancestry modelling for the seventh to fifth millenniumâbc (c) and fourth millenniumâbc (d) individuals from the Caucasus region with temporally and geographically proximal sources. The admixture proportions (as percentages) are indicated on each ancestry flow, with sources on the left and target populations on the right, and P values for each model in brackets under the population names (Supplementary Tables 8, 10 and 13). The suffixes in the group labels present archaeological time periods and geographical regions: WHG, Western hunter-gatherer; SHG, Scandinavian hunter-gatherer; MN, Middle Neolithic; C, Chalcolithic; EC, Early Chalcolithic; LC, Late Chalcolithic. Ãatalhöyük_N is shown as Anatolia Neolithic, Tell Kurdu is shown as Anatolia EC and Jordan_PPNB and Jordan_PPNC are shown as Levant PPN.
By contrast, the four Neolithic individuals from Arukhlo (ARO and AO2, Georgia_Neolithic) together with Armenia_N (with N denoting Neolithic) and Azerbaijan_LN (with LN denoting Late Neolithic) form a genetic cline between CHGs and central Anatolian Neolithic individuals (for example, Ãatalhöyük), who themselves fall on a cline between Anatolian and Levantine ancestries14,15,16 (Fig. 2a and Extended Data Figs. 3b and 4). Thus, we modelled all Neolithic groups from the region using Anatolia_PPN (with PPN denoting pre-pottery Neolithic), Levant_PPN and CHG-Iran_N as distal ancestry sources (Fig. 2c, Extended Data Fig. 5 and Supplementary Table 8), and find that Georgia_Neolithic and Armenia_Aknashen_N carry the highest proportion of CHG-like ancestry, whereas Armenia_MasisBlur_N and Azerbaijan_LN carry more Levant_PPN-related ancestry. However, we also tested two-way mixture models between Anatolian or Levantine Neolithic groups and CHG and find that Georgia_Neolithic can be modelled as a two-way mixture between CHG and Ãatalhöyük_N (shown as Anatolia Neolithic in Fig. 2a) or Tell Kurdu (Fig. 2c), but not with PPN groups from central Anatolia, Levant and Mesopotamia or Neolithic Northwest Anatolia.
Eneolithic
During the Eneolithic period (4900/4700â3900 bc), a settled Near Eastern Neolithic lifestyle was introduced on the northern flanks of the Caucasus in association with the DarkvetiâMeshoko culture (4500â4000âbc)5. Subsequently, the steppe zone further north was populated by transitional forager-pastoralist groups of the lower Volga Khvalynsk Eneolithic culture17. Adding genome-wide data from eight new individuals allows us to describe the formation of the Eneolithic groups in the North Caucasus steppe zone. We observe the earliest formation of Steppe ancestry, resulting from the gradual mixture of EHG-like ancestry with CHG-like ancestry from the south. Together with published Eneolithic and Khvalynsk individuals5,18, the newly reported Steppe_Eneolithic individuals form a genetic cline in PC space between EHG and CHG (Fig. 2a). f4-statistics show that Steppe_Eneolithic individuals from the North Caucasus have a higher affinity to CHG than individuals from Khvalynsk (Extended Data Figs. 4 and 6 and Supplementary Table 9), and can be modelled as 55% CHG-like and 45% EHG-like ancestry (Supplementary Table 10).
We thus confirm the emergence of two distinct genetic clusters on the basis of PCA and ADMIXTURE, the Steppe cluster along the EHGâCHG cline and the mountain-oriented Caucasus cluster on the cline from Anatolia Neolithic to CHG-Iran_N (Fig. 2a, Extended Data Figs. 1 and 2 and ref.â5). Using the new Georgia_Neolithic data, we reassess the genetic ancestry of agropastoral DarkvetiâMeshoko Caucasus_Eneolithic individuals from the Northwest Caucasus, which we model as a two-way mixture of Georgia_Neolithic (51%â±â6) and CHG (49%â±â6) ancestries (Fig. 2c, Extended Data Fig. 4 and Supplementary Table 10).
The individuals from Nalchik (NCK001 4531â4359âcal bc; NCK002 4930â4686âcal bc), labelled as Eneolithic_intermediate, fall between the Steppe and Caucasus groups in the PCA, suggesting gene flow between these groups. Indeed, ADMIXTURE and f4-statistics show that Nalchik individuals carry Caucasus ancestry, but also EHG-like ancestry (Fig. 2d, Extended Data Fig. 2 and Supplementary Tables 9 and 10). This means that the EHGâCHG cline and Caucasus ancestry16 must have already been formed by the time the ancestors of Nalchik met. Further, it limits the time frame of this mixture to the early fifth millenniumâbc and anticipates an axis of interaction that intensifies in the fourth millenniumâbc. Using DATES19, we estimate admixture dates for groups of both clusters and find that the Anatolia Neolithic to CHG-Iran_N cline formed around 6300â6000âbc, consistent with previous estimates16, and the EHGâCHG cline formed around 5800â5300âbc (Extended Data Fig. 5c and Supplementary Table 11). Of note, KHB003 (4318â4057âcal bc) from the western-most site has a higher genetic affinity to Western hunter-gatherer (WHG) and Anatolia Neolithic-like ancestry (|Z| > 3), and can be modelled as a two-way mixture between CHG and Ukraine_Neolithic (Fig. 2d and Supplementary Tables 9 and 10).
Late Eneolithic and EBA
The fourth millenniumâbc signifies a time period of dynamic population interaction and cultural transitions, which are visible in the material culture associated with Maykop traditions. Among the set of Late Eneolithic and EBA Maykop individuals, to which we added 20 new individuals, we confirm three previously defined genetic groups5: Maykop_main, Steppe_Maykop and Steppe_Maykop_outlier1, but also identify three new groups: Late_Steppe_Eneolithic, Late_Steppe_Eneolithic_outlier and Steppe_Maykop_outlier2. The genetic profiles of Maykop_main individuals can be further distinguished as three subgroups: Maykop, Late_Maykop and Maykop_Novosvobodnaya5.
The individuals KST001 and NV3003 (3781â3652âcal bc), labelled as the Late_Steppe_Eneolithic group, fall also on the EHGâCHG cline (Fig. 2a,b). However, we find significantly higher affinity of Late_Steppe_Eneolithic individuals to EHG than Steppe_Eneolithic (Supplementary Table 12), although qpAdm reveals a similar ancestry profile with 52% SJG001 and 48% CHG ancestry (Fig. 2d and Supplementary Table 13). The individuals ZO1002 and ZO1004 (3953â3713âcal bc) are shifted towards the Caucasus cluster in PCA and ADMIXTURE (Fig. 2b and Extended Data Fig. 2), and carry less EHG-related ancestry compared to the Late_Steppe_Eneolithic individuals, and are thus labelled Late_Steppe_Eneolithic_outlier. Using proximal sources, we could model both as a two-way mixture of Caucasus_Eneolithic (55%â±â6.4) and Steppe_Eneolithic (45%â±â6.4) ancestries (Fig. 2d and Supplementary Table 13). Together with the Nalchik and Steppe_Maykop_outlier1 individuals, this reflects gene flow between Eneolithic groups living in the steppe and the Caucasus foothills (Fig. 1b).
The Maykop-associated individuals form a tight third Caucasus Maykop_main cluster, which resembles the preceding Caucasus_Eneolithic individuals, suggesting genetic continuity between these groups (Supplementary Table 12). Two-way mixture models between the preceding Caucaus_Eneolithic and Armenia_C (with C denoting Chalcolithic) or Anatolia_C lack statistical support (Pâ<â0.05), but the addition of Iran_C groups as a third source resulted in well-fitted models for all Maykop_main individuals (Fig. 2d and Supplementary Table 13). Late_Maykop individuals can be modelled with earlier Maykop ancestry as a single, locally preceding source, suggesting that Iran_C-related gene flow had occurred during the early Maykop phase.
The remaining two groups fall along a different genetic cline between West Siberian hunter-gatherers (WSHGs) and Caucasus_Eneolithic individuals. The first group (AY2004, IV3005 and KUG001) from sites in the PonticâCaspian steppe dates to the Maykop period, but is genetically positioned between Steppe_Eneolithic and individuals from Botai in Central Asia and WSHG east of the Ural Mountains, who carry increased Ancestral North Eurasian (ANE) ancestry (Fig. 2b and Extended Data Fig. 2). As this group shares archaeological features attributed to the Maykop culture, it was originally described as Steppe_Maykop5. We show that WSHG-like ancestry contributes up to 48% to the genetic make-up of Steppe_Maykop individuals, arguing for gene flow from regions further northeast, whereas this component is absent from all other contemporaneous groups in the Caucasus and Steppe clusters (Fig. 2d, Supplementary Tables 10 and 13 and Extended Data Fig. 2).
The last group (nâ=â6; KUG002-005, IV3010 and AY2001) falls in the space of preceding Steppe_Eneolithic groups (Fig. 2b), but f4-statistics and single-source qpAdm models reject direct population continuity with the preceding Steppe_Eneolithic individuals (Supplementary Tables 12 and 13). However, as this group post-dates the horizon of Maykop and Steppe_Maykop interaction, and three out of four male individuals carry the Y-chromosome haplogroup Q1b-M346 more commonly found in Steppe_Maykop and North Siberian populations, we explored alternative models involving Steppe_Maykop ancestry. This group can indeed be modelled as a two-way mixture of Steppe_Maykop (62â±â1.6%) and Maykop_Novosvobodnaya (38â±â1.6%) ancestries, and was consequently labelled Steppe_Maykop_outlier2.
From EBA to MBA
We report new genome-wide data from ten individuals associated with the Yamnaya cultural complex that we refer to as Yamnaya_North_Caucasus (Yamnaya_NC; 3300â2800âbc; Supplementary Table 1). They broadly fall on the EHGâCHG cline of the Steppe groups in PC space (Fig. 3a, Extended Data Fig. 1 and Supplementary Table 14), forming a tight cluster with published data from the Black Sea, Samara and North Caucasus regions. f4-statistics confirm their close genetic similarity, albeit with subtle geographic differences. Responding to previous studies that explored the regions of contact between Yamnaya pastoralists and farming groups5,20,21, we tested a series of possible two-way qpAdm models. Using various Steppe groups as the baseline ancestry and CucuteniâTrypillia and Globular Amphora (CTC-GAC) or Maykop_main individuals as a second âfarming-associatedâ source, only western Ukraine_Yamnaya can be modelled as a two-way mixture of Steppe_Eneolithic and CTC-GAC, whereas these models are rejected for Yamnaya_NC and Yamnaya_Samara individuals (Supplementary Table 15). However, adding Ukraine_Neolithic and Mesolithic as a third source improved the model fit for almost all groups. Thus, we can model Yamnaya_NC as a three-way mixture of the local proximal sources Steppe_Eneolithic, Maykop and Ukraine_Neolithic (Fig. 3c), although models with CTC-GAC as alternative source(s) are also supported.
a,b, PCA of newly produced ancient individuals (black outline) and individuals from previous publications (no outline) from the third millenniumâbc (a), and the second millenniumâbc (b), projected onto 102 modern-day populations (grey dots). The dashed arrows represent observed mixture clines between the Caucasus and Steppe groups and re-emerging gene flow from the northeast. The correspondent labels and groupings are listed in Supplementary Table 5. c,d, Sankey diagram of genetic ancestry modelling for third millenniumâbc (c) and second millenniumâbc (d) individuals from this study based on temporally and geographically proximal sources. The admixture proportions (as percentages) are indicated on each ancestry flow, with sources on the left and target populations on the right, and P values for each model in brackets under the population names (Supplementary Tables 15 and 17). The suffixes in the group labels present archaeological time periods and geographical regions: MLBA, MiddleâLate BA; BIA, BronzeâIron Age; IA, Iron Age.
Later MBA individuals associated with the North Caucasus culture (NCC; 2800â2400âcal bc, nâ=â12) and Catacomb culture (2800â2200âcal bc, nâ=â8) also fall within the cluster of Yamnaya-associated individuals in PCA (Fig. 3a), and f4-symmetry tests reveal only subtle genetic differences among them (Supplementary Table 14), which suggests genetic continuity. Indeed, Catacomb individuals can be modelled with Yamnaya_NC as a single, locally preceding source, whereas NCC individuals can be modelled only with Ukraine_Yamnaya instead (Supplementary Table 15). All Steppe groups from the third millenniumâbc overlap in their admixture date estimates, ranging from about 4800âbc to 4000âbc, and thus differ from the early Eneolithic formation of steppe-related ancestry, but coincide with the presence and early interaction of both Steppe and Caucasus Eneolithic groups north of the Caucasus (Supplementary Table 11).
A similar pattern of genetic continuity and homogenization is also observed in the Caucasus cluster. Individuals associated with the KuraâAraxes culture in Georgia (3600/3300â2400âcal bc, nâ=â6) fall close in PC space and ADMIXTURE with published KuraâAraxes individuals from Armenia and Dagestan, as well as Maykop individuals (Figs. 2b and 3a and Extended Data Fig. 2), suggesting continuity of the Caucasus ancestry profile during the MBA, but with heterogeneity20 among different KuraâAraxes groups (Supplementary Table 14). Using Maykop groups as a single source results in well-fitted models (from Pâ=â0.09 to Pâ=â0.7; Supplementary Table 15), with Maykop_Novosvobodnaya as the best source for KuraâAraxes individuals from Georgia and Armenia (Berkaber, Kalavan, Karnut and Shengavit), and Maykop as the best source for Talin (Pâ=â0.2). By contrast, individuals from Kaps in Armenia or Velikent in Dagestan require additional ancestry from either Armenia_C or Iran_C, or both.
The Iran_BA individuals BOE001 (2861â2489âcal bc) and BOE003 (2881â2623âcal bc) fall close to those from the nearby Chalcolithic site Tepe Hissar11 on a Southwest Asian cline. f4-statistics indicate that Iran_BA has a higher genetic affinity with Chalcolithic groups from Anatolia and KuraâAraxes individuals from Karnut (Armenia), and ancestry modelling using qpAdm resulted in well-fitted models with Iran_TepeHissar_C and KuraâAraxes groups as sources (Supplementary Tables 14 and 15).
Final MBA and LBA
The final MBA (2200â1650âbc), represented by the post-Catacomb cultural horizon22 and LBA (1800â1200/1000âbc)23 phases, marks another period of increased population interaction and transformation, as evidenced by 33 new individuals from this period. The PC space previously occupied by Steppe cluster individuals is now largely void, with only four individuals of different genetic ancestries (KVO009, KNK006, ESY007 and ESY009) falling in this position. All other kurgan burials in the central steppe zone are shifted towards the Caucasus cluster owing to significantly increased affinity to South Caucasus populations (Zâ=â3.352; Fig. 3b and Supplementary Table 16). Consequently, these post-Catacomb individuals can be modelled successfully as a mixture of preceding Catacomb (79%) and KuraâAraxes (21%) groups (Fig. 3d and Supplementary Table 17).
Among contemporaneous Steppe populations, individuals of the Lola culture (nâ=â9) represent the predominant ancestry pattern, which falls close in PC space to earlier Steppe_Maykop individuals5 (Fig. 3b and Extended Data Fig. 2). We separate them into two groups (Lola_1 and Lola_2), on the basis of differing amounts of ANE ancestry indicated by PCA and f4-statistics (Supplementary Table 16). We next tested whether Lola represented a continuation of Steppe_Maykop ancestry that had returned to the North Caucasus, but find that models with Steppe_Maykop as a single ancestry source are rejected, whereas two-way mixtures of Steppe_Maykop and either NCC or Catacomb-associated ancestry are supported (Supplementary Table 17). Considering the 2,000-year time gap between Steppe_Maykop and Lola, we also tested multiple qpAdm models with North Caucasus MBA Steppe groups (for example, Catacomb or NCC) as the local substrate and central steppe Early and Middle BA11 groups as a non-local source of ANE ancestry, and find well-fitted models with Kazakhstan Kumsay EBA. The individual KVO013 from the Northwest Caucasus falls close in PC space to BA Srubnaya individuals from the eastern European forest steppe (Fig. 3b) and can be modelled as a mixture of preceding BA Sintashta11 and Steppe groups (Fig. 3d and Supplementary Table 17). The LBA Prescythian_steppe individuals ESY006, ESY007, ESY009 and KNK006 represent the last signal of local Steppe ancestry in the North Caucasus (Fig. 3b). However, f4-symmetry tests show that these individuals carry traces of ANE ancestry similar to Lola and Steppe_Maykop, and can be modelled as a two-way mixture of Srubnaya and Lola_1 ancestry (Supplementary Tables 16 and 17).
The final MBA and LBA individuals from the Caucasus cluster are markedly shifted upwards on PC2 towards the Steppe cluster (Fig. 3b). This marks the first time in which ancient individuals fall within the same PC space as present-day populations. Individuals from the western and eastern Caucasus highlands are spread along PC2, implying varying levels of admixture with the Steppe cluster (Fig. 3b and Extended Data Fig. 2). The individuals KVO008 (MBA_Caucasus) and PUT001 (Arkhon) carry similar ancestry profiles, suggesting that a convergence of both clusters might not have been restricted to the Caucasus highlands. Using f4-statistics, we find that most MBA and LBA individuals show an affinity to Steppe groups (Supplementary Table 16). Further, compared to the MBA_Highland_east group, KVO008 and PUT001 show an even higher genetic affinity to EHGâWSHG and BA Steppe groups, and both can be modelled successfully with KuraâAraxes as a proxy for Caucasus ancestry and BA Steppe groups as sources (Supplementary Tables 16 and 17). This two-way model is also supported for MBA and LBA eastern highland individuals, albeit with a higher proportion of KuraâAraxes-related ancestry, but can also be modelled with the preceding MBA_Highland_east group as a single source (Supplementary Table 17), implying that this gene flow had occurred already before the final MBA. Moreover, we observe a genetic cline from east to west, suggesting geographic structure in LBA highlander ancestry (Fig. 3b). Here, individuals from Shushuk and Marchenkova Gora in the northwest share more genetic drift with BA Srubnaya and Steppe groups compared to eastern individuals from Ginchi and Gatyn-Kale, which is consistent with the finding of the Srubnaya-associated individual KVO013 from the Northwest Caucasus (Figs. 1a and 3b). Western highlanders also show a greater genetic affinity with Maykop individuals, whereas eastern highlanders are more similar to KuraâAraxes individuals (Supplementary Table 16). Using Maykop and KuraâAraxes groups as the respective locally preceding ancestry source and Srubnaya as the second source resulted in well-fitted models (Supplementary Table 17). The five LBA individuals from the site of Lernakert in Armenia (1411â1266âcal bc) occupy a similar position in the PCA as the published MBA and LBA individuals from other sites in Armenia, suggesting local genetic continuity20,24. Using Armenia_MBA as the local baseline in f4-statistics and qpAdm, we find support for a single local source.
Time transects and demographic snapshots
The most conspicuous mortuary features on the Eurasian steppe are kurgans, earthen mounds that marked graves in the North Caucasus since the fifth millenniumâbc, throughout the BA and beyond. Kurgans were often built incrementally, spanning many centuries and cultural periods25, and thus provide a perspective on genetic continuity or discontinuity at individual sites. In many mounds, the deceased were buried in a non-continuous series of events. Previous research has assumed close genetic or genealogical relations between individuals buried in such mounds26. To test this, we estimated genetic relatedness between pairs of individuals. Among 105 individuals from 21 multi-burial kurgans5, we find 15 first- or second-degree relationships among all possible pairs (nâ=â5,460, 0.27%). Even when filtering for pairs from the same site (nâ=â272; 5.5%) or chrono-cultural overlap (nâ=â1,147; 1.3%), we find predominantly unrelated or only distantly related pairs (Supplementary Tables 3, 4 and 18). Moreover, we observe a significant sex ratio bias towards male burials in the Steppe (Pâ=â0.035) but not in the Caucasus (Pâ=â0.850; Extended Data Fig. 7) cluster. Overall, this suggests that kurgans were generally not pedigree- or lineage-based burial grounds.
The multiphase mounds of Komsomolec 1-Marfa (KMM) and Marinskaya 5 (MK5) represent two well-dated examples. Focusing on the primary EBA and MBA occupation phases, with 12 and 7 individuals, respectively, we observe shifting cultural and genetic affinities among individuals in chronological succession (Extended Data Fig. 8a,b). We find a Late_Maykop brotherâsister pair at KMM, and a grandfatherâgrandson pair at MK55, whereas all other individuals associated with later cultures are unrelated.
Using ancIBD27, we estimated genetic relationships up to the sixth degree between all individuals from the Caucasus. We confirm the first-degree relationship between the Maykop individuals VIN001 and VS5001 from neighbouring mounds, and also identify more distant relationships (for example, between the grandfatherâgrandson pair at MK5 and the individual ESY005, buried 60âkm apart; Extended Data Fig. 9 and Supplementary Table 18). However, we observe no shared identity by descent between Steppe and Caucasus cluster individuals in the entire dataset. Examining runs of homozygosity (ROH)28, we observe that Steppe cluster individuals have a higher number of short ROH tracts (4â8âcM and 8â12âcM) compared to Caucasus cluster individuals (Extended Data Fig. 10a,c), indicating a smaller effective population size of Steppe communities. We also detect five cases of consanguinity (ROHâ>20âcM) in Maykop_main individuals, including offspring of second-cousin unions (AY2001 and AY2003), a first-cousin union (ESY005 and SIJ003) and a full- or half-sibling union (VS5001; Extended Data Fig. 10b,d).
Discussion
We observe two genetically distinct populations in the Caucasus region before the Neolithic transition. SJG001 shares a close genetic affinity to hunter-gatherer groups from Karelia and the Samara region, as opposed to the geographically closer Ukraine_Mesolithic and Neolithic individuals, attesting to a lasting legacy of EHG ancestry across a large area of eastern Europe6. The lack of genetic admixture from the south opposes ideas of immigration of Epipalaeolithic groups from the Fertile Crescent, as proposed on the basis of similar lithic industries29,30. By contrast, groups carrying CHG-related ancestry must have persisted in the South Caucasus, as this ancestry is a source for Georgia_Neolithic. This cline between Anatolian Neolithic and CHG-like groups echoes the purported origin of the Neolithic expansion from the Fertile Crescent to the intermontane valleys in the Lesser Caucasus31,32 and shows that expanding Neolithic groups interacted early and intensively with local groups. The earliest sites in the Kura and Araxes valleys date to 6000/5900âbc. The individuals from Arukhlo reflect the rapid assimilation of the initial immigrants, which contrasts with the limited interaction between expanding farmers and WHG in Europe. The local CHG-like ancestry is also still detectable in individuals from Menteshtepe (Azerbaijan)33 and Akanashen (Armenia)15.
Neolithic lifeways emerged on the northern flanks of the Greater Caucasus among Eneolithic DarkvetiâMeshoko agropastoral pioneers who share ancestry with South Caucasus populations. About 4300âbc, a different population, culturally related to the Khvalynsk Eneolithic17,34, arose in the PonticâCaspian steppe further to the north and built the first burial mounds. This marks the first appearance of Steppe ancestry in the region, formed through pre-Neolithic hunter-gatherer interactions before the emergence of the DarkvetiâMeshoko groups, whose ancestry contains Anatolian Neolithic ancestry not found in Steppe groups. The initial formation of Steppe ancestry dates to the mid-sixth millenniumâbc and is thus consistent with the described sequence of events. KHB003 from Khutor Belyy reflects genetic interaction of Steppe groups with peoples along the eastern parts of the Black Sea35, a trajectory that later became more important during the Yamnaya period. The oldest Eneolithic and genetically intermediate individuals from the cemetery of Nalchik provide a temporal constraint for the preceding CHG-related gene flow to the north. However, placing these Eneolithic events in sequence is challenged by radiocarbon dating reservoir effects, a problem that to some degree also affects BA pastoralist groups36,37. The earliest interaction between Steppe and Caucasus groups also extended to the south (for example, Areni 1 in Armenia)24. The consumption of dairy products by the Steppe_Eneolithic individuals KUG007 and PG2001 (ref.â2) and the presence of caprine teeth ornaments at Nalchik graves34 suggest that incipient pastoral economies may have facilitated these contacts in the Eurasian steppe2,38. Future genetic studies of domesticated animals will probably clarify the origins of the animals and related pastoral technologies that had already advanced to the mountains of Central Asia and Mongolia in the early third millenniumâbc (refs.â39,40).
The fourth millenniumâbc emergence of Maykop traditions in the Caucasus mountains and piedmonts signifies a clear cultural transition, whereas Eneolithic traditions persisted in the neighbouring steppe41. We observe five genetically distinguishable and largely contemporaneous groups. All piedmont-associated Maykop individuals carry Caucasus ancestry inherited from southern Neolithic and Eneolithic groups, which was probably maintained by keeping close kinship ties in cohesive communities, also reflected in shared architectural construction features in some Maykop mounds. By contrast, the genetic variability and scarcity of close biological relationships in the four Steppe Eneolithic groups suggest different and more flexible kin structures, which echo the persistence of varying cultural practices41.
For all groups, however, the archaeological record supports the transfer of Maykop material culture and social practices from the piedmonts into the steppe25,38, and the alternating use of the same mounds by different genetic groups suggests cultural interconnectedness. As sheep dairying practices became more prominent during this period2, isotope data from about 3500âbc onwards show a separation of grazing lands between communities in piedmont and steppe environments42. Maykop contexts show that the cultural interactions persisted for about 1,000 years, whereas the actual Caucasus cluster population (Maykop_main) did not spread and largely avoided intermarriage with Steppe groups. Innovations such as cattle-drawn wheeled transport and initial steps towards horse domestication gradually boosted mobility and herd management43,44. The clockwise tilt in the genetic cline of Late_Steppe_Eneolithic individuals from an EHGâCHG axis to an WSHGâMaykop axis is thus notable, as it encompasses additional WSHG ancestry, which is also found at Botai in Central Asia, another area of incipient equid domestication11,44,45. The genetic affinity of the Steppe_Maykop to eastern groups reflects the opening of the Eurasian steppe, even though this link is enigmatic and not yet related to any known archaeological phenomenon. Other technological innovations also started to spread, such as grassland-adapted sheep for dairying2, wheels and wagons43,46, and possibly wool as a material for insulating clothes and mobile architecture47. The display of wealth in graves was also soon transferred westwards, as indicated by close cultural links to Usatovo groups in the northwestern Black Sea area48. Indeed, the genetic profiles of several Steppe_Eneolithic individuals attest to contact with groups at the regionâs western periphery and form the genetic substrate from which the third millenniumâbc PonticâCaspian steppe pastoralists later emerged21. The horizon of combined innovations in the North Caucasus enabled the emergence of pastoralism and the dynamic modes of interaction, connectivity and mobility that subsequently spread across larger geographic regions, bridging the Caucasus region, the PonticâCaspian steppe and Europe with lands in Central and Inner Asia10.
Genetic and archaeological evidence combined offer a perspective on the consolidation of pastoral economies in the third millenniumâbc, including homogenization of the Steppe ancestry profile and the emergence of Yamnaya groups. Notably, members of culturally distinct NCC and Catacomb communities25 also fall into this homogeneous genetic group. Moreover, we find individuals carrying Steppe ancestry at sites in the Caucasus mountains5, which suggests that groups with Caucasus ancestry had retreated higher into the mountains. The scarcity of closely related Steppe individuals is remarkable, given that some stem from narrow burial sequences within one mound. Combined with an elevated parental background relatedness, this suggests a form of social organization and kinship that regulates exogamy within a relatively small effective population. The Western Eurasian steppe pastoralists, best represented by the Yamnaya culture, stabilized and expanded their economy based on multispecies dairy products2,49,50 and wheeled vehicle mobility46, and spread their sustainable, permanent and self-supporting mobile lifestyle across the Eurasian steppe39,40,51.
By contrast, the contemporaneous Caucasus population is represented only by individuals of the KuraâAraxes culture at present. In parallel to the pastoralist expansions across the Eurasian steppe, KuraâAraxes groups expanded into the Levant and todayâs Iran52, which is reflected in the genetic shift towards Iranian BA populations. We observe no genetically intermediate individuals, but also acknowledge sampling gaps for this epoch.
The second millenniumâbc is marked by a decline of activity and population density in the North Caucasus steppe, and a total abandonment by about 1700âbc (ref.â2), possibly brought about by climatic shifts and overexploitation of ecologically fragile steppe habitats following the 4.2âkyrâbp environmental crisis22,53. The homogeneous Steppe ancestry cluster of the preceding third millenniumâbc dissolves into several post-Catacomb groups, who show genetic affinities to Central Asia, as seen among the Lola, or to the northwestern Srubnaya. Interaction between mountain and steppe cultures also intensified during this period, evident in material culture and the genetic shift of the Caucasus groups towards the Steppe cluster and vice versa. This was previously interpreted as an expansion of mountain populations into the PonticâCaspian steppe between 2200âbc and 1700âbc (ref.â22), but we observe instead an absorption of Steppe groups into Caucasus groups, probably driven by the steppe habitats becoming increasingly inhospitable.
From these developments, a pan-Caucasian mountain interaction sphere from dolmens in the northwest (Shushuk) to Dagestan (Ginchi) emerged that culturally represents the LBA23 and also a genetic ancestry profile that still persists in the North Caucasus today12,13. This process anticipated the integration of the mountain populations in the succeeding LBA and Early Iron Age cultures23, and marked the transition from a mobile BA steppe pastoralist economy to a more sedentary and complex agropastoral mountain economy54.
For two millennia, mobile pastoralism dominated lifeways on the great expanses of steppe extending northwards from the Caucasus mountains. Fuelled by technological innovations such as wheeled transport and dairy pastoralism, as well as emerging horse husbandry, steppe populations from the CaucasusâSteppe interface exerted a large influence on the Eurasian landmass, leaving far-flung genetic and cultural footprints that remain even today. Understanding the dynamic and complex population interactions that shaped the regionâs most influential BA groups, such as the Maykop, Yamnaya and KuraâAraxes, is key to reconstructing the population history of both Europe and Asia. Here we reveal the genetic events that led to the formation of these groups and trace the regionâs history through its ultimate decline and abandonment about 1700âbc, even as the regionâs mobile pastoralism legacy continued to spread and flourish elsewhere55.
Methods
Permission statement
Permission to work on the archaeological samples was granted by the respective excavators, archaeologists, curators and museum directors of the sites, who are co-authoring the study. Excavation licence numbers are provided in the Supplementary Information.
Radiocarbon dating
We obtained new direct 14C dates for 84 individuals. Radiocarbon dating was carried out using accelerated mass spectrometry at the Curt-Engelhorn-Zentrum Archäometrie in Mannheim, Germany (Fig. 1a and Supplementary Table 1). All new and published dates from the Caucasus were calibrated on the basis of the IntCal20 database using OxCal v4.4.2.
Ancient DNA laboratory work
All of the laboratory work was carried out in dedicated ancient DNA facilities of the Archaeogenetics Department of the Max Planck Institute for the Science of Human History in Jena and that for Evolutionary Anthropology in Leipzig, Germany, and at the University of California, Santa Cruz. We have mainly sampled petrous bones and teeth (Supplementary Table 1) using a minimally invasive method described in the archived protocols https://doi.org/10.17504/protocols.io.bdyvi7w6 and https://doi.org/10.17504/protocols.io.bqebmtan. DNA extraction for all samples was carried out with a modified protocol56, and the details of each step are described at https://doi.org/10.17504/protocols.io.baksicwe. DNA double-stranded libraries57 were prepared from some of these extracts using a partial uracil-DNA glycosylase58 (UDG-half) treatment, followed by Illumina dual indexing59. A detailed description of the steps is available at https://doi.org/10.17504/protocols.io.bmh6k39e and https://doi.org/10.17504/protocols.io.4r3l287x3l1y/v3. For a proportion of the samples, we used an automated protocol for producing single-stranded libraries60,61, to improve sequence retrieval from ancient DNA. The single-stranded libraries for seven individuals (Dzedzvebi, Georgia) was prepared with the Santa Cruz reaction protocol62 and UDG-half treatment. We initially screened 253 samples from 211 individuals for ancient human DNA preservation by preparing double-stranded genetic libraries for 112 samples, single-stranded libraries for 115 samples, and both library types for 26 samples. Genomic libraries passing quality control thresholds (>0.1% endogenous DNA and >3% damage pattern) were enriched for about 1.24 million targeted single nucleotide polymorphisms (SNPs) across the human genome18. Details of the libraries produced for each individual are reported in Supplementary Table 2. All prepared libraries were initially sequenced on an Illumina HiSeq 4000 platform to an average of 5 million reads. Raw FastQC files were processed through the EAGER pipeline63, for assessment of human DNA content and DNA damage profiles. After quality assessment, all of the libraries with 0.1% human endogenous DNA or more were enriched for around 1.2 million SNPs in a targeted in-solution capture (1240k SNP capture)18. Additional in-house capture assays were prepared for the complete mitogenome64, for which the sequence reads were not sufficient for mitochondrial haplogroup calls, and for mappable regions of the Y chromosome65 (YMCA), for selected male individuals (Extended Data Fig. 7b). Captured libraries were sequenced for 20 to 40 million reads, using either a single-end (1âÃâ75âbase pair (bp)) or paired-end (2âÃâ50âbp) configuration.
Data processing and genotyping
The captured sequences were demultiplexed, and then further processed using EAGER63 (v2.4.0) and nf-core/eager (v2.3.2). AdapterRemoval (v2.3.2) was used to remove Illumina adaptors. Subsequently, BWA (v0.7.17) was used to map reads to the human reference genome hs37d5, and duplicates were removed using MarkDuplicates (v2.26.0). mapDamage (v2.2.1) was used to determine the deamination rate pattern (G to A and C to T substitutions) in the libraries. For the UDG-half-treated double-stranded libraries, the trimbam function of bamUtils v1.0.1366 was used to trim 2âbp from terminal ends of reads. Thereafter, we used the pileupCaller (v1.5.2) (https://github.com/stschiff/sequenceTools) tool for genotyping, which generates pseudo-haploid genotypes by randomly choosing one of the alleles at every SNP position. For single-stranded libraries we used the --singleStrandMode parameter to remove post-mortem ancient DNA damage (at C>T SNPs, we discard forward-mapping reads, and at G>A SNPs, we discard reverse-mapping reads). In cases in which both single-stranded and double-stranded libraries were prepared for the same individual, the libraries were genotyped separately as described above. Subsequently, the genotypes from these different libraries were merged using a custom script by randomly picking alleles from available genotype calls.
Ancient DNA authentication
The genetic sex of the individuals was determined with the Sex.DetERRmine tool (v1.1.2) through EAGER63. The genetic sex of the 120 individuals could be determined confidently, of which 71 were male and 49 female. Following this, the ANGSD67 (v0.935) tool was used to calculate the rate of heterozygosity on the Xâchromosome to determine contamination in genetically male individuals, applying a contamination threshold of 5% in individuals with at least 100 X-SNP positions covered twice. Furthermore, we used Schmutzi68 and ContamMix69 to quantify heterozygosity on the individual mitochondrial reads. In cases in which the coverage is not sufficient or the mitochondrial to nuclear DNA ratios are very high (>200), contamination estimates are not reliable70. Hence, when appropriate, we depended on alternative methods and/or the behaviour of these samples in population genetic analyses. In cases in which there are multiple individuals available from the same genetic group, the downstream analyses were carried out on non-post mortem damage (PMD)-filtered genotypes as minor levels of individual contamination would be diluted within the group. In total, 26 individuals who did not meet the quality and authentication criteria (<5% nuclear contamination, <10% mitochondrial DNA contamination and >30,000 SNPs covered on the 1240k panel) were excluded.
Uniparentally inherited markers
Sequences from mitogenome capture were aligned to the complete human mitochondrial genome, and Schmutzi68 was used to infer the consensus sequence for each individual. The mitochondrial haplogroups were assigned to each consensus sequence using HaploGrep2 (v2.4.0)71. The Y-chromosome reads from the 1240k SNP capture and YMCA were genotyped according to a SNP list from the International Society of Genetic Genealogy dataset65. This allows for the manual inspection of assigned ancestral and derived alleles and their correction in cases in which, owing to residual ancient damage (C to T or G to A mismatches), a more derived haplogroup was called.
Genetic relatedness
We used BREADR72Â and READv273 to calculate the pairwise mismatch rate and to estimate genetic relatedness between pairs of individuals belonging to the same site or the same genetic grouping. For the cases with few individuals, an unrelatedness baseline was provided based on the value observed in groups with similar genetic ancestry and/or chrono-cultural group. In the event of first-degree relative pairs, we excluded the individual with the lower number of SNPs from downstream population genetic analyses. For summary statistics, we refer to the chrono-cultural group-based estimates from READv2. Genetic data for samples that were identified as identical (that is, coming from the same individual; nâ=â6) were merged for downstream analysis. Among the newly reported individuals, there were three pairs from Essentukskiy 1, as well as one pair and one triplet from Balitshi-DzedzvebiâII. In addition, sample NV3002 was found to be from the same published individual NV3001 from the same site5.
Sex bias significance tests
To test for a significantly different sex bias ratio in Steppe and Caucasus cluster burials compared to expectation, we applied a one-sample test for proportions. We compared the proportion of reliably assigned genetically male individuals in each group to a male-to-female birth ratio of 1.06:1 (ref.â74), and calculated a P value using a two-sided test, and adjusted the P values using Bonferroni75 correction to correct for multiple hypothesis tests.
Power analysis for biological relatedness in multiphase kurgans
To investigate whether we had the statistical power required to carry out rigorous hypothesis testing for a change in relatedness within multiphase kurgans (Supplementary Table 3), we calculated the statistical power of a β-regression using the pwr package76. Using the standard significance level of αâ=â0.05, and power of βâ=â0.8, we can reliably detect only an effect size of fâ2â=â1.64 (greater than the possible effect size of 1.0 for β-regression). Similarly, if we use a Ï2 test, we can reliably detect only an effect size77 of greater than 0.305, which is considered greater than a medium effect size.
Reference dataset
The genotyped data were merged with the Human Origins78,79 and published ancient data3,5,6,7,11,14,16,18,20,24,45,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107 using AADR108 and Poseidon databases109 (Supplementary Table 5). The Human Origins dataset with 597,573 SNPs was used for analysis comparing ancient to modern groups, such as PCA and ADMIXTURE, whereas the 1240k dataset of 1,233,013 SNPs was used for comparison between ancient groups in f-statistics and ancestry modelling.
PCA
We used smartpca (v16000) from the EIGENSOFT (7.2.1) software package110 to carry out PCA with the âlsqproject: YESâ, âshrinkmode: YESâ and ânumoutlieriter: 0â parameters to project ancient individuals onto PCs calculated on different subsets of the genotype data of modern individuals from the Human Origins dataset. We used genotype data of 1,243 individuals from 82 modern populations for the West Eurasian PCA (Extended Data Fig. 1a), and an extended West Eurasian PCA with genotype data of 1,522 individuals from 102 populations (Extended Data Fig. 1b and Supplementary Table 20), including 20 populations from Tajikistan, Pakistan and India that enhance the geographic resolution of the studied genetic groups.
ADMIXTURE analysis
We carried out an unsupervised ADMIXTURE111 (v1.3.0) analysis together with groups from the Human Origins dataset. Before analysis, we filtered the dataset for linkage disequilibrium with the parameters --indep-pairwise 200 25 0.4. Additionally, we also removed any low-coverage individual with a 97% missing rate with the parameter --mind 0.97. In the end, 239,791 SNPs passed these quality filters. We used the default fivefold cross-validation (--cv=5), ranging the number of ancestral components from Kâ=â1 to Kâ=â17.
f-statistics
Both the ADMIXTOOLS (7.0.2) package79 and its wrapper program admixr (v0.9.1)112 were used to calculate all of the f3- and f4-statistics analyses with âinbreed: YESâ and âf4mode: YESâ parameters settings, respectively. The default block jackknife approach was used to calculate standard errors (Supplementary Tables 6, 7, 9, 12, 14 and 16).
Ancestry modelling and admixture proportion estimation
We used the qpWave and qpAdm (v1520) software packages available in the ADMIXTOOLS4,113 package to test various scenarios of continuity and admixture by identifying possible ancestry streams and to quantify these admixture proportions, with the âallsnps: NO, details: YESâ options. We used the âallsnps: NOâ parameter so that all underlying f-statistics were calculated with the same set of overlapping SNPs between all groups. In the cases in which the target individual or groups had <200,000 overlapping SNPs, we used the âallsnps: YESâ parameter to test the validity of the models. For all qpAdm models, we started with a base set of outgroups (Mbuti.DG, Russia_MA1_HG.SG, Italy_North_Villabruna_HG, Anatolia_Epipalaeolithic, Levant_PPN) and added certain groups or individuals as we progressed in time with our modelling. In other words, some of the groups that were used as a source in previous models were added to the outgroup set later. We are interested in possible proximal sources that contributed to the target groups under investigation. We used the available high-coverage individuals with ancestries relevant for our targeted regions and time periods for the outgroup set. However, we tried to keep the number of outgroup sets minimal, as large numbers eventually result in a reduction of P values to below the threshold113. In addition, we did not use anachronistic sources in our qpAdm models. Our threshold for feasible models was Pâ>â0.05. In the case of multiple feasible models, we report the most parsimonious models (that is, with the closest spatially and temporally sources and with fewer source groups). Details on the outgroups used in each model and results of all attempted models are reported in Supplementary Tables 8, 10, 13, 15 and 17.
Admixture date estimation
We used the software DATES (v4010)19 to estimate the date of admixture events. DATES measures decay of ancestry covariance patterns in target ancient groups, given two ancestral sources. The rate of decay is informative about the time of admixture. We used the following parameters settings: binsize 0.001; maxdis 1; qbin 10; lovalfit 0.45. To convert the estimated dates in generations to calendar years bc, we assumed 28 years per generation (Supplementary Table 11).
Imputation
We used GLIMPSE (v1.0.1) with the default parameters to impute the individuals in this study together with published data from the Caucasus. Genotype likelihoods were determined from trimmed bam files (2âbp) using bcftools with the 1000 Genome Phase 3 release as a reference. Imputation was carried out using GLIMPSE_impute on genomic chunks of 2,000,000âbp with a buffer size of 200,000âbp. The chunks were subsequently ligated using GLIMPSE_ligate, and the most likely haplotypes were identified with GLIMPSE_sample. Individuals with >500,000 SNPs (0.5Ã coverage on the 1240k positions) post-imputation were included in the identity-by-descent (IBD) analysis.
ROH
To estimate parental relatedness from regions with long ROHs for which an individual is homozygous, that is, has identical haplotypes inherited from each parent, we used hapROH (v1.0)28 for individuals with >400,000 SNPs.
IBD sharing
We used ancIBD (v0.4)27 software to analyse IBD sharing between individuals with more than 500,000 SNPs and genotype probabilities higher than 0.99 after imputation using GLIMPSE114,115. The HapBLOCK function was used to estimate IBD sharing. After merging imputed samples, we used the vcf_to_1240K_hdf command to convert the vcf files to the recommended hdf5 format. Using the default parameters of the hapBLOCK_chroms command, we carried out IBD sharing analysis for each chromosome. For each pair, a summary statistics table was generated, keeping only the pairs with at least one IBD recorded in 20âcM for plotting.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The DNA sequences reported in this paper have been deposited in the European Nucleotide Archive under the accession number PRJEB73987. The genotype data are available as a Poseidon package in the Poseidon Community Archive (https://www.poseidon-adna.org/#/archive_explorer). Maps were created using QGIS 3.36.0-Maidenhead, Adobe Adobe Illustrator 28.3 and Adobe Photoshop 25.5.
Code availability
All software used in this work is publicly available and versions are listed in the Nature Portfolio Reporting Summary. Corresponding publications are cited in the Article and Supplementary information. Custom code for the power analyses is available via Zenodo at https://doi.org/10.5281/zenodo.13709775 (ref.â116).
References
Giemsch, L. & Hansen, S. (eds) The Caucasus: Bridge Between the Urban Centres in Mesopotamia and the Pontic Steppes in the 4th and 3th millennium BC. Proceedings of the Caucasus Conference Frankfurt 2018 (Schnell & Steiner, 2021).
Scott, A. et al. Emergence and intensification of dairying in the Caucasus and Eurasian steppes. Nat. Ecol. Evol. 6, 813â822 (2022).
Allentoft, M. E. et al. Population genomics of Bronze Age Eurasia. Nature 522, 167â172 (2015).
Haak, W. et al. Massive migration from the steppe was a source for Indo-European languages in Europe. Nature 522, 207â211 (2015).
Wang, C.-C. et al. Ancient human genome-wide data from a 3000-year interval in the Caucasus corresponds with eco-geographic regions. Nat. Commun. 10, 590 (2019).
Posth, C. et al. Palaeogenomics of Upper Palaeolithic to Neolithic European hunter-gatherers. Nature 615, 117â126 (2023).
Jones, E. R. et al. Upper Palaeolithic genomes reveal deep roots of modern Eurasians. Nat. Commun. 6, 8912 (2015).
Hansen, S. in Western Anatolia before Troy: Proto-urbanisation in the 4th Millennium BC? (eds Horejs, B. & Mehofer, M.) 243â260 (Austrian Academy of Sciences Press, 2014).
Dinerstein, E. et al. An ecoregion-based approach to protecting half the terrestrial realm. Bioscience 67, 534â545 (2017).
Kohl, P. L. The Making of Bronze Age Eurasia (Cambridge Univ. Press, 2007).
Narasimhan, V. M. et al. The formation of human populations in South and Central Asia. Science 365, eaat7487 (2019).
Balanovsky, O. et al. Parallel evolution of genes and languages in the Caucasus region. Mol. Biol. Evol. 28, 2905â2920 (2011).
Yunusbayev, B. et al. The Caucasus as an asymmetric semipermeable barrier to ancient human migrations. Mol. Biol. Evol. 29, 359â365 (2012).
Yaka, R. et al. Variable kinship patterns in Neolithic Anatolia revealed by ancient genomes. Curr. Biol. 31, 2455â2468 (2021).
Lazaridis, I. et al. Ancient DNA from Mesopotamia suggests distinct Pre-Pottery and Pottery Neolithic migrations into Anatolia. Science 377, 982â987 (2022).
Skourtanioti, E. et al. Genomic history of Neolithic to Bronze Age Anatolia, Northern Levant, and Southern Caucasus. Cell 181, 1158â1175 (2020).
Anthony, D. W. et al. The Eneolithic cemetery at Khvalynsk on the Volga River. Praehist. Zeitschr. 97, 22â67 (2022).
Mathieson, I. et al. Genome-wide patterns of selection in 230 ancient Eurasians. Nature 528, 499â503 (2015).
Chintalapati, M., Patterson, N. & Moorjani, P. The spatiotemporal patterns of major human admixture events during the European Holocene. Elife 11, e77625 (2022).
Lazaridis, I. et al. The genetic history of the Southern Arc: a bridge between West Asia and Europe. Science 377, eabm4247 (2022).
Penske, S. et al. Early contact between late farming and pastoralist societies in southeastern Europe. Nature 620, 358â365 (2023).
Mimokhod, R. A., Gak, E. I., Khomutova, T. E., Ryabogina, N. E. & Borisov, A. E. Paleoekologiya â kulâtura â metalloproizvodstvo: printsipy i mekhanizmy epokhi v kul'turnom nasledii Vostochnoy Yevropy v sredniye veka - pervyye pozdney bronzy. Ross. Arkheol. 24â38 (2022).
Reinhold, S., Belinskiy, A. B. & Atabiev, B. C. in Der Kaukasus zwischen Osteuropa und Vorderem Orient in der Bronze- und Eisenzeit (eds KaÅ¡uba, M. T. et al.) 405â456 (Dietrich Reimer, 2020).
Lazaridis, I. et al. Genomic insights into the origin of farming in the ancient Near East. Nature 536, 419â424 (2016).
Shishlina, N. Reconstruction of the Bronze Age of the Caspian Steppes: Life Styles and Life Ways of Pastoral Nomads Vol. 1876 (Archaeopress, 2008).
Andreeva, M. V. Vostochnomanychskaya Kata-Kombnaya Kulâtura: Analiz Materialov Pogrebal'nykh Pamyatnikov (TAUS, 2014).
Ringbauer, H. et al. Accurate detection of identity-by-descent segments in human ancient DNA. Nat. Genet. 56, 143â151 (2023).
Ringbauer, H., Novembre, J. & Steinrücken, M. Parental relatedness through time revealed by runs of homozygosity in ancient DNA. Nat. Commun. 12, 5425 (2021).
Golovanova, L. V. et al. The Epipaleolithic of the Caucasus after the Last Glacial Maximum. Quat. Int. 337, 189â224 (2014).
Manko, V. O. & Chkhatarashvili, G. L. in Aere Perennius. More Lasting than Bronze. Essays in Honour of Valentin Dergachev on the Occasion of his 80th Birthday (ed. Dergacheva, L. V.) 101â118 (Stratum Library, 2023).
Helwing, B. et al. The Kura Projects: New Research on the Later Prehistory of the Southern Caucasus Vol. 16 (Dietrich Reimer, 2017).
Baudouin, E. Lâarchitecture de Mésopotamie et Du Caucase de La Fin Du 7e á La Fin Du 5e Millénaire Vol. 2 (Brepols, 2021).
Guarino-Vignon, P. et al. Genome-wide analysis of a collective grave from Mentesh Tepe provides insight into the population structure of early neolithic population in the South Caucasus. Commun. Biol. 6, 309 (2023).
Korenevskiy, S. N. Rozhdenie Kurgana: (pogrebalnye Pamyatniki Eneoliticheskogo Vremeni Predkavkazya I Volgo-Donskogo Mezhdurechia) (TAUS, 2012).
Gorelik, A., Tsybryi, A. V. & Tsybryi, V. V. âNeolithisationâ in the NE Sea of Azov region: one step forward, two steps back? Doc. Praehist. 43, 139â160 (2016).
Hollund, H. I., Higham, T., Belinskij, A. & Korenevskij, S. Investigation of palaeodiet in the North Caucasus (South Russia) Bronze Age using stable isotope analysis and AMS dating of human and animal bones. J. Archaeol. Sci. 37, 2971â2983 (2010).
Shishlina, N., Zazovskaya, E., van der Plicht, J. & Sevastyanov, V. Isotopes, plants, and reservoir effects: case study from the Caspian steppe Bronze Age. Radiocarbon 54, 749â760 (2012).
Reinhold, S. Der Kaukasus und die Eurasische Steppe â Konjunkturen einer kulturellen Kontaktzone während der Bronze- und frühen Eisenzeit. in You Only See What You Know. You Only Know What You See. Global Historic Perspectives on Intercultural Phenomena of Mobility. Festschrift Für Hermann Parzinger zum 65. Geburtstag (eds Schneeweiss, J. et al.) 437â460 (Marie Leidorf, 2024).
Hermes, T. R. et al. Early integration of pastoralism and millet cultivation in Bronze Age Eurasia. Proc. Biol. Sci. 286, 20191273 (2019).
Honeychurch, W. et al. The earliest herders of East Asia: examining Afanasievo entry to Central Mongolia. Archaeol. Res. Asia 26, 100264 (2021).
Korenevskiy, S. N. & Berezin, Y. B. Eneolithic Kurgan Burials from the Cemetery of Konstantinovsky-6. Strat. Plus 2/17, 385â396 (2017).
Knipper, C. et al. Diet and subsistence in Bronze Age pastoral communities from the southern Russian steppes and the North Caucasus. PLoS ONE 15, e0239861 (2020).
Reinhold, S. et al. in Appropriating Innovations (eds Stockhammer, P. W. & Maran, J.) 78â97 (Oxbow, 2017).
Librado, P. et al. The origins and spread of domestic horses from the Western Eurasian steppes. Nature 598, 634â640 (2021).
de Barros Damgaard, P. et al. The first horse herders and the impact of early Bronze Age steppe expansions into Asia. Science 360, eaar7711 (2018).
Burmeister, S. in Appropriating Innovations (eds Stockhammer, P. W. & Maran, J.) 69â77 (Oxbow, 2017).
Shishlina, N., Ankusheva, P., Orfinskaya, O. & Kiseleva, D. in The Indo-European Puzzle Revisited. Integrating Archaeology, Genetics, and Linguistics (eds Kristian, K., Kroonen, G. & Willerslev, E.) 275â281 (Cambridge Univ. Press, 2023).
Hansen, S. in The Caucasus (eds Giemsch, L. & Hansen, S.) 31â86 (Schnell & Steiner, 2021).
Jeong, C. et al. Bronze Age population dynamics and the rise of dairy pastoralism on the eastern Eurasian steppe. Proc. Natl Acad. Sci. USA 115, E11248âE11255 (2018).
Wilkin, S. et al. Dairying enabled Early Bronze Age Yamnaya steppe expansions. Nature 598, 629â633 (2021).
Kaiser, E. Das dritte Jahrtausend im osteuropäischen Steppenraum: Kulturhistorische Studien zu prähistorischer Subsistenzwirtschaft und Interaktion mit benachbarten Räumen (Freie Universität Berlin, 2019).
Batiuk, S. D. The fruits of migration: understanding the âlongue duréeâ and the socio-economic relations of the Early Transcaucasian Culture. J. Anthropol. Archaeol. 32, 449â477 (2013).
PerÅoiu, A., Ionita, M. & Weiss, H. Atmospheric blocking induced by the strengthened Siberian High led to drying in west Asia during the 4.2 ka BP event â a hypothesis. Clim. Past 15, 781â793 (2019).
Reinhold, S. et al. At the onset of settled pastoralism - implications of archaeozoological and isotope analyses from Bronze Age sites in the North Caucasus. Quat. Int. https://doi.org/10.1016/j.quaint.2023.05.008 (2023).
Reinhold, S., Korobov, D. S. & Belinskiy, A. B. Landschaftsarchäologie im Nordkaukasus: Studien zu einer neu entdeckten bronzezeitlichen Kulturlandschaft im Hochgebirge des Nordkaukasus Vol. 38 (Habelt, 2017).
Dabney, J. et al. Complete mitochondrial genome sequence of a Middle Pleistocene cave bear reconstructed from ultrashort DNA fragments. Proc. Natl Acad. Sci. USA 110, 15758â15763 (2013).
Meyer, M. & Kircher, M. Illumina sequencing library preparation for highly multiplexed target capture and sequencing. Cold Spring Harb. Protoc. 2010, db.prot5448 (2010).
Rohland, N., Harney, E., Mallick, S., Nordenfelt, S. & Reich, D. Partial uracil-DNA-glycosylase treatment for screening of ancient DNA. Philos. Trans. R. Soc. B 370, 20130624 (2015).
Kircher, M., Sawyer, S. & Meyer, M. Double indexing overcomes inaccuracies in multiplex sequencing on the Illumina platform. Nucleic Acids Res. 40, e3 (2012).
Gansauge, M.-T. & Meyer, M. Single-stranded DNA library preparation for the sequencing of ancient or damaged DNA. Nat. Protoc. 8, 737â748 (2013).
Gansauge, M.-T., Aximu-Petri, A., Nagel, S. & Meyer, M. Manual and automated preparation of single-stranded DNA libraries for the sequencing of DNA from ancient biological remains and other sources of highly degraded DNA. Nat. Protoc. 15, 2279â2300 (2020).
Kapp, J. D., Green, R. E. & Shapiro, B. A fast and efficient single-stranded genomic library preparation method optimized for ancient DNA. J. Hered. 112, 241â249 (2021).
Peltzer, A. et al. EAGER: efficient ancient genome reconstruction. Genome Biol. 17, 60 (2016).
Maricic, T., Whitten, M. & Pääbo, S. Multiplexed DNA sequence capture of mitochondrial genomes using PCR products. PLoS ONE 5, e14004 (2010).
Rohrlach, A. B. et al. Using Y-chromosome capture enrichment to resolve haplogroup H2 shows new evidence for a two-path Neolithic expansion to Western Europe. Sci. Rep. 11, 15005 (2021).
Jun, G., Wing, M. K., Abecasis, G. R. & Kang, H. M. An efficient and scalable analysis framework for variant extraction and refinement from population-scale DNA sequence data. Genome Res. 25, 918â925 (2015).
Korneliussen, T. S., Albrechtsen, A. & Nielsen, R. ANGSD: Analysis of Next Generation Sequencing Data. BMC Bioinform. 15, 356 (2014).
Renaud, G., Slon, V., Duggan, A. T. & Kelso, J. Schmutzi: estimation of contamination and endogenous mitochondrial consensus calling for ancient DNA. Genome Biol. 16, 224 (2015).
Fu, Q. et al. A revised timescale for human evolution based on ancient mitochondrial genomes. Curr. Biol. 23, 553â559 (2013).
Furtwängler, A. et al. Ratio of mitochondrial to nuclear DNA affects contamination estimates in ancient DNA analysis. Sci. Rep. 8, 1â8 (2018).
Weissensteiner, H. et al. HaploGrep 2: mitochondrial haplogroup classification in the era of high-throughput sequencing. Nucleic Acids Res. 44, W58âW63 (2016).
Rohrlach, A. B., Tuke, J., Popli, D. & Haak, W. BREADR: An R package for the Bayesian estimation of genetic relatedness from low-coverage genotype data. Preprint at bioRxiv https://doi.org/10.1101/2023.04.17.537144 (2023).
Alaçamlı, E. et al. READv2: advanced and user-friendly detection of biological relatedness in archaeogenomics. Genome Biol. 25, 216 (2024).
Chao, F., Gerland, P., Cook, A. R. & Alkema, L. Systematic assessment of the sex ratio at birth for all countries and estimation of national imbalances and regional reference levels. Proc. Natl Acad. Sci. USA 116, 9303â9311 (2019).
Sedgwick, P. Multiple significance tests: the Bonferroni correction. BMJ 344, e509 (2012).
Champely, S. Pwr: Basic Functions for Power Analysis. https://CRAN.R-project.org/package=pwr (2020).
Cohen, J. Statistical Power Analysis for the Behavioral Sciences (Routledge, 2013).
Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for present-day Europeans. Nature 513, 409â413 (2014).
Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065â1093 (2012).
Rasmussen, M. et al. Ancient human genome sequence of an extinct Palaeo-Eskimo. Nature 463, 757â762 (2010).
Rasmussen, M. et al. The genome of a Late Pleistocene human from a Clovis burial site in western Montana. Nature 506, 225â229 (2014).
Seguin-Orlando, A. et al. Genomic structure in Europeans dating back at least 36,200 years. Science 346, 1113â1118 (2014).
Raghavan, M. et al. Upper Palaeolithic Siberian genome reveals dual ancestry of Native Americans. Nature 505, 87â91 (2014).
Rasmussen, M. et al. The ancestry and affiliations of Kennewick Man. Nature 523, 455â458 (2015).
Kılınç, G. M. et al. The Demographic development of the first farmers in Anatolia. Curr. Biol. 26, 2659â2666 (2016).
Broushaki, F. et al. Early Neolithic genomes from the eastern Fertile Crescent. Science 353, 499â503 (2016).
Fu, Q. et al. The genetic history of Ice Age Europe. Nature 534, 200â205 (2016).
van den Brink, E. C. M. et al. A Late Bronze Age II clay coffin from Tel Shaddud in the Central Jezreel Valley, Israel: context and historical implications. Levantina 49, 105â135 (2017).
Mittnik, A. et al. The genetic prehistory of the Baltic Sea region. Nat. Commun. 9, 442 (2018).
Harney, Ã. et al. Ancient DNA from Chalcolithic Israel reveals the role of population mixture in cultural transformation. Nat. Commun. 9, 3336 (2018).
Mathieson, I. et al. The genomic history of southeastern Europe. Nature 555, 197â203 (2018).
KrzewiÅska, M. et al. Ancient genomes suggest the eastern Pontic-Caspian steppe as the source of western Iron Age nomads. Sci. Adv. 4, eaat4457 (2018).
de Damgaard, P. B. et al. 137 ancient human genomes from across the Eurasian steppes. Nature 557, 369â374 (2018).
Feldman, M. et al. Late Pleistocene human genome suggests a local origin for the first farmers of central Anatolia. Nat. Commun. 10, 1218 (2019).
Sikora, M. et al. The population history of northeastern Siberia since the Pleistocene. Nature 570, 182â188 (2019).
Flegontov, P. et al. Palaeo-Eskimo genetic ancestry and the peopling of Chukotka and North America. Nature 570, 236â240 (2019).
Jeong, C. et al. The genetic history of admixture across inner Eurasia. Nat. Ecol. Evol. 3, 966â976 (2019).
Rivollat, M. et al. Ancient genome-wide DNA from France highlights the complexity of interactions between Mesolithic hunter-gatherers and Neolithic farmers. Sci. Adv. 6, eaaz5344 (2020).
Yu, H. et al. Paleolithic to Bronze Age Siberians reveal connections with First Americans and across Eurasia. Cell 181, 1232â1245 (2020).
Agranat-Tamir, L. et al. The genomic history of the Bronze Age Southern Levant. Cell 181, 1146â1157 (2020).
Jeong, C. et al. A dynamic 6,000-year genetic history of Eurasiaâs Eastern Steppe. Cell 183, 890â904 (2020).
Kılınç, G. M. et al. Human population dynamics and Yersinia pestis in ancient northeast Asia. Sci. Adv. 7, eabc4587 (2021).
Saag, L. et al. Genetic ancestry changes in Stone to Bronze Age transition in the East European plain. Sci. Adv. 7, eabd6535 (2021).
Zhang, F. et al. The genomic origins of the Bronze Age Tarim Basin mummies. Nature 599, 256â261 (2021).
AltınıÅık, N. E. et al. A genomic snapshot of demographic and cultural dynamism in Upper Mesopotamia during the Neolithic Transition. Sci. Adv. 8, eabo3609 (2022).
Koptekin, D. et al. Spatial and temporal heterogeneity in human mobility patterns in Holocene Southwest Asia and the East Mediterranean. Curr. Biol. 33, 41â57 (2023).
Wang, X. et al. Isotopic and DNA analyses reveal multiscale PPNB mobility and migration across Southeastern Anatolia and the Southern Levant. Proc. Natl Acad. Sci. USA 120, e2210611120 (2023).
Mallick, S. et al. The Allen Ancient DNA Resource (AADR): a curated compendium of ancient human genomes. Sci. Data 11:182 (2024).
Schmid, C. et al. Poseidon â a framework for archaeogenetic human genotype data management. eLife https://doi.org/10.7554/elife.98317.1 (2024).
Patterson, N., Price, A. L. & Reich, D. Population structure and eigenanalysis. PLoS Genet. 2, e190 (2006).
Alexander, D. H., Novembre, J. & Lange, K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 19, 1655â1664 (2009).
Petr, M., Vernot, B. & Kelso, J. admixrâR package for reproducible analyses using ADMIXTOOLS. Bioinformatics 35, 3194â3195 (2019).
Harney, Ã., Patterson, N., Reich, D. & Wakeley, J. Assessing the performance of qpAdm: a statistical tool for studying population admixture. Genetics 217, iyaa045 (2021).
Delaneau, O. GLIMPSE: Low Coverage Calling of Genotypes https://github.com/odelaneau/GLIMPSE (2020).
Rubinacci, S., Ribeiro, D. M., Hofmeister, R. J. & Delaneau, O. Efficient phasing and imputation of low-coverage sequencing data using large reference panels. Nat. Genet. 53, 120â126 (2021).
Rohrlach, B. BenRohrlach/CaucasusIIAncientDNAStudy: v1.0.0. Zenodo https://doi.org/10.5281/zenodo.13709775 (2024).
Acknowledgements
We thank E.âKukral, A.âWissgott, G.âBrandt, S.âNagel and E.âEssel for support in the ancient DNA analyses, and K.âPrüfer for the processing of the raw sequence data; the teams at (former) MPI-SHH-Archaeogenetics and MPI-EVA-Archaeogenetics, the Population Genetics and PALEoRIDER groups, especially E.âSkourtanioti, M.âRivollat and V.âVillalba-Mouco, for continued support and discussion, and A. Evteev for sampling support; all excavation teams that produced and stored the archaeological material and the radiocarbon laboratory team of Curt-Engelhorn-Centre for Archaeometry for their support. This research was financed by the Max Planck Society, the European Research Council under the European Unionâs Horizon 2020 research and innovation programme (grant agreement numbers 771234-PALEoRIDER, W.H.; 834616-ARCHCAUCASUS, S.H.) and the ERA.Net RUS Plus initiative (S&T-277-BIOARCCAUCASUS, S.H. and A.P.B.).
Funding
Open access funding provided by Max Planck Society.
Author information
Authors and Affiliations
Contributions
W.H., S.H., S.R. and A.G. designed the study. S.R., A.A.K., K.B.-L., A.B.B., N.Y.B., Y.B.B., A.P.B., V.R.E., L.F.-S., I.G., A.R.K., K.B.K., D.L., R.G.M., K.M.-C., D.M., V.E.M., L.M., A.N., H.F.N., M.O., Y.Y.P., M.S., A.G.S., T.S., J.T., B.V., C.P. and J.K. provided materials and resources. F.A., L.S. and N.B. carried out laboratory experiments. A.G. analysed the data with the help of A.B.R. and H.Y., and A.C., W.H., S.R., C.W., C.P. and S.H. assisted with data interpretation. A.G., S.R., A.A.K., S.H., A.B.R., C.W. and W.H. wrote the manuscript with contributions from all co-authors.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature thanks Michael Frachetti, Andrew Millard, Ludovic Orlando and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisherâs note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Principal component analysis (PCA) of prehistoric individuals from the Caucasus region.
a, extended West Eurasian PCA with 1522 individuals from 102 populations. b, West Eurasian PCA with 1243 individuals from 82 populations. The new individuals are shown with black outline.
Extended Data Fig. 2 Results of ADMIXTURE analysis (k = 10) of the individuals from the Caucasus.
a, Individuals belonging to the Steppe cluster, b, Individuals belonging to the Caucasus cluster, and c, a representative selection of published reference individuals. Newly genotyped individuals (black outline) and published individuals are sorted by archaeological or genetic groups in chronological order from left to right.
Extended Data Fig. 3 Outgroup f3-statistics comparing genetic affinities of Mesolithic and Neolithic groups.
a, Heatmap showing shared genetic drift estimated by outgroup f3-statistics of the form f3(Mbuti; X, Y) between all EHG-related groups published to date. Lighter colours indicate higher f3 values which corresponds to higher shared genetic drift. b, Scatterplot of outgroup f3-statistics measuring shared drift between Neolithic and Chalcolithic groups to Anatolia and Levant PPN. Groups from the Caucasus cluster are shown in red and brown symbols. The trend line is given by the red dashed line. Error bars indicate ± 1 s.e. and were calculated using a weighted block jackknife across all autosomes on the 1,240,000 panel (nSNPs = 1,150,639) and a block size of 5 Mb.
Extended Data Fig. 4 Formal test for temporal genetic changes in the Caucasus region.
a, f4-statistic tests where X denotes Anatolia N, EHG, or Iran_N and b, where X denotes CHG, EHG, and Iran_N, and Test denotes various genetic groups from the Caucasus and Steppe clusters on the y axis. Significant Z-scores (|Z| > 3) are highlighted in orange and error bars indicate ± 3 s.e. and were calculated using a weighted block jackknife across all autosomes on the 1,240,000 panel (nSNPs = 1,150,639) and a block size of 5 Mb.
Extended Data Fig. 5 Admixture modelling of Neolithic South Caucasus groups and Admixture dating.
Ternary plot of distal three-way qpAdm models for Neolithic and Chalcolithic groups from South Caucasus, Anatolia and Mesopotamia. a, Iran_N and b, CHG together with Anatolia_PPN, Levant_PPN are used as sources. c, Admixture date estimates calculated using the method DATES and two different pools of sources (Supplementary Table 11). Radiocarbon date intervals for each genetic group are shown as grey bars, circles and diamonds are the estimated admixture dates converted to calendar dates (assuming a generation time of 28 years) based on the mean and the oldest 14C dates, respectively. Error bars indicate ± 1 s.e.
Extended Data Fig. 6 Key f4-statistics and qpAdm results of Eneolithic and Yamnaya groups.
a, Variation in genetic affinities of Eneolithic groups to EHG or CHG. Significant Z-scores (|Z| > 3) are highlighted in orange and error bars indicate ±1 (black) and ±3 (light gray) s.e. and were calculated using a weighted block jackknife across all autosomes on the 1,240,000 panel (nSNPs = 1,150,639) and a block size of 5 Mb. b, Ternary plot of qpAdm modelling for Steppe Eneolithic and Yamnaya groups (Supplementary Tables 10 and 15).
Extended Data Fig. 7 Results of genetic sex analysis and uniparentally inherited markers.
a, Total count of XX and XY individuals per genetic grouping from this study and Wang et al.5. The chrono-cultural and genetic groups are chronologically ordered within the Steppe and Caucasus clusters. b, Y-chromosome haplogroups and c, mitochondrial haplogroups compared across the three main genetic clusters of Caucasus, Intermediate and Steppe. The x axis shows the total number of individuals per genetic cluster.
Extended Data Fig. 8 Kurgans as demographic snapshots and micro-transects through time.
a, and b, show the general layout of the Marinskaya 5 and Marfa kurgans (Bronze Age only). Analysed individuals are highlighted in different colours. c, and d, PCA plots highlighting the genetic heterogeneity of individuals buried in these kurgans. e, and f, Estimates of genetic relatedness calculated as mean pairwise mismatch rates (PMR) for individuals from Marinskaya 5 and Marfa, with error bars indicating ± 2 s.e. Dashed lines represent the expected PMR for each relatedness degree.
Extended Data Fig. 9 IBD analysis results per pair of individuals with ancIBD27.
a, Visualising the sum (x axis) and the number (y axis) of all IBD tracts with a length of at least 12cM. Only pairs with at least two shared chunks of 20cM length are shown. b, Map showing IBD networks of closely related individuals within and between sites of the piedmont and steppe zone. The map was generated using Base Relief: Mapzen, OpenStreetMap, and rivers, lakes and borders were added using free vector and raster map data from Natural Earth (https://www.naturalearthdata.com). OpenStreetMap is open data, licensed under the Open Data Commons Open Database Licence by the OpenStreetMap Foundation.
Extended Data Fig. 10 Assessment of runs of homozygosity and inbreeding.
a, b, and c, Inferred run of homozygosity ROH per individual in Steppe and Caucasus clusters, respectively. Results are plotted by genetic and cultural groups in relative chronological order from left to right. b, Full scale of inferred ROH for individuals in the Maykop group. The legend illustrates the expected ROH for offspring of small populations or closely related individuals. d, Karyogram of VS5001. The positions of ROH longer than 4 cM are marked on the 22 autosomes (maroon).
Supplementary information
Supplementary Information
Supplementary Notes Sections 1â5 (see contents for details).
Supplementary Information
A guide to Supplementary Tables 1â20 (tables supplied separately).
Supplementary Tables
Supplementary Tables 1â20.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the articleâs Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the articleâs Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Ghalichi, A., Reinhold, S., Rohrlach, A.B. et al. The rise and transformation of Bronze Age pastoralists in the Caucasus. Nature 635, 917â925 (2024). https://doi.org/10.1038/s41586-024-08113-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41586-024-08113-5
