Was the Maikop culture the PIE homeland?

Figure S4. Eurasian PCA with Neolithic to Bronze Age Individuals from Iran and Turan and the Genetic Outlier Individual from Alalakh (ALA019), Related to Figure 7

Scatter plot of PC1 and PC2 from PCA computed on modern-day Eurasian populations (gray points) shows that the Alalakh_MLBA_outlier (ALA019) is genetically closer to individuals from Chalcolithic and Bronze Age Iran/Turan. Colored labels and points refer to ancient populations and black labels to modern-day populations.

Discussion
Genetic Homogenization across Anatolia and the Southern Caucasus prior to the Bronze Age
Our study covers time transects of ∼4,000 years of Syro-Anatolian and ∼2,000 years of Southern Caucasian history, both starting from the 6th millennium BCE. In addition, our admixture date estimates allow us to infer a millennium further back in time to the Neolithic period. We describe a Late Neolithic/Early Chalcolithic (6th millennium BCE) genetic cline stretching from Western Anatolia (i.e., area around the Sea of Marmara) to the lowlands of the Southern Caucasus that was formed by an admixture process that started at the beginning of Late Neolithic (∼6500 years BCE). The eastern end of this cline extends beyond the Zagros mountains with minute proportions of Anatolian (i.e., Western Anatolian-like) ancestry reaching as far as Chalcolithic and Bronze Age Central Asia (Narasimhan et al., 2019). To the south, Anatolian ancestry is present in the Southern Levantine Neolithic populations (Lazaridis et al., 2016), and to the north, in the Chalcolithic and Bronze Age populations from the Caucasus (mainly mountainous area) (Allentoft et al., 2015, Lazaridis et al., 2016, Wang et al., 2019), most likely as a result of the Late Neolithic admixture.

Evidence for genetic homogenization across larger geographic distances also comes from the uniparentally inherited Y chromosome lineages (Table S9). We observe the most common male lineages J1a, J2a, J2b, and G2a in all spatiotemporal groups of the region. Alongside the less frequent lineages H2 and T1a, these all form part of the genetic legacy that dates to the Neolithic or was already present in the region during the Upper Paleolithic (Wang et al., 2019, Lazaridis et al., 2016, Jones et al., 2015, Feldman et al., 2019, Broushaki et al., 2016). A few notable exceptions provide rather anecdotal but nonetheless important evidence for long distance mobility and extended Y-haplogroup diversity. For example, individual ART038 carries Y-haplotype R1b-V1636 (R1b1a2), which is a rare clade related to other early R1b-lineages, such as R1b-V88 that was found in low frequency in Neolithic Europe (e.g., Haak et al., 2015) and R1b-Z2103—the main Y-lineage that is associated with the spread of “steppe ancestry” across West Eurasia during the early Bronze Age. However, R1b-V1636 and R1b-Z2103 lineages split long before (∼17 kya) and therefore there is no direct evidence for an early incursion from the Pontic steppe during the main era of Arslantepe. Lineage L2-L595 found in ALA084 (Alalakh) has previously been reported in one individual from Chalcolithic Northern Iran (Narasimhan et al., 2019) and in three males from the Late Maykop phase in the North Caucasus (Wang et al., 2019). These three share ancestry from the common Anatolian/Iranian ancestry cline described here, which indicates a widespread distribution that also reached the southern margins of the steppe zone north of the Caucasus mountain range.

Dating the formation of the west-to-east cline during the 7th millennium BCE enables the contextualization of these genetic signals observed on both autosomal and uniparental markers with archaeological evidence of human mobility and changes in socio-cultural practices. The time around 6500–6400 years BCE marks a significant junction in the Anatolian Neolithic because it saw a sudden and massive expansion of sedentary communities into areas that had previously supported no or very few food-producing communities before (Brami, 2015, Düring, 2013). Subsequently, in the Southern Caucasus, the abrupt appearance of a Neolithic lifestyle and the introduction of exogenous domesticated animal and plant species ca. 6000 BCE suggests some type of interaction with, and eventually intrusion of Neolithic populations from the neighboring regions, among which Southeastern Anatolia—along with Zagros and the Caspian belt—could be one of the most suitable candidates (Chataigner et al., 2014). Related to these events, the genetic structure of domesticated caprine populations within the Near East began to break down, and by the Chalcolithic period, goat herds across the region were found to harbor ancestries both from eastern and western Neolithic populations (Daly et al., 2018, Kadowaki et al., 2017). Although the exact timing of this admixture is not known, the parallel between human and livestock genetic histories suggests that livestock moved not only through trade networks but also together with people, as well as their material culture, ideas, and practices. This is indicated, for instance, by the circular Neolithic architecture of the Southern Caucasus (Baudouin, 2019, Lyonnet et al., 2016), which is reminiscent of the Halaf traditions, that were developing during the early 6th millennium in North Mesopotamia and along the Anatolian stretches of the Tigris and Euphrates river valleys.

In the subsequent millennia and until the Late Bronze Age, genetic continuity persisted in North-Central and Eastern Anatolia, which is supported by the genetic similarity of these later populations and the absence of new ancestry sources after the Neolithic period. This contradicts prior hypotheses for population change based on archaeological evidence of intense cultural interactions during this period. For example, the site of İkiztepe on the Turkish Black Sea Coast contains a material culture with strong Balkan affinities, and this has been argued to signify direct contact with populations across the Black Sea (e.g., Thissen, 1993), but these contacts do not seem to be accompanied by gene flow.

The site of Arslantepe provides another representative example. At the beginning of EBA, archaeological evidence at the site strongly suggests the presence of a disruptive socio-political conflict that led to the occupation of Arslantepe by pastoral populations with a connection to the Caucasus. In PCA and f4-statistics, two individuals dating to this period show excess affinity with populations from the Caucasus and the Pontic steppe compared to their peers, while later Arslantepe_EBA individuals do not share this Caucasian affinity (Table S10). This implies that the postulated demic interaction must have been transient and of small scale, although the small sample size from Arslantepe_EBA (n = 4) may not be sufficient to detect it. Subtle gene flow is consistent with recent findings from the site, suggesting that the EBA pastoralists occupying the site were more likely well-established local groups moving around the mountains rather than intrusive groups from the Caucasus (Frangipane, 2014).

The genetic landscape of Arslantepe also has important implications with respect to the interactions with the Mesopotamian world. Archaeological evidence indicates that in the 4th millennium BCE Mesopotamian populations established colonies in Southeastern Anatolia and Northern Syria (e.g., Habuba Kabira, Jebel Aruda, Hacinebi) during a period referred to as the Uruk Expansion (Algaze, 2005). However, the Uruk expansion was also a very complex and deep process of socio-cultural transformation that reoriented the economic, political, and cultural interests of indigenous elites toward Southern Mesopotamia. Artifacts in Arslantepe reflect this complexity, and the genetic continuity shown here supports the notion that indigenous populations adopted these broader Uruk features and ideas without the transmission of genes.

Population and Territorial State Dynamics in the Northern Levant
In contrast to the rest of Anatolia, the Northern Levant stands out as a region of the Near East with traceable post-Neolithic changes in the genetic structure. We found that the gene pools at Ebla and Alalakh can only be explained by more complex models that require additional contributions both from the Caucasus and Southern Levant. The inclusion of a source related to the Caucasus in our proposed model raises the question whether this turnover could be linked to the expansion of the Transcaucasian Kura-Araxes material culture to the Levant. This expansion is recorded in the region of the Northern Levant ca. 2800 BCE and could be associated with the movement/migration of people from Eastern Anatolia and the Southern Caucasian highlands (Greenberg and Palumbi, 2015, Greenberg et al., 2014). However, our results do not support this scenario for a number of reasons: (1) we do not find any substantial increase of Caucasus-related ancestry in the populations of the primary expansion area of Kura-Araxes (e.g., Eastern Anatolia), (2) populations from the highlands of the Southern Caucasus—including individuals from a Kura-Araxes context (“L.Caucasus_EBA”)—as secondary source populations also fail, and (3) so do models with Arslantepe from Eastern Anatolia, a population located mid-way along the proposed expansion route from the Southern Caucasus to the Northern Levant.

Consequently, these interpretative caveats call for consideration of alternative historical scenarios, including scenarios of multiple gene-flow events that could have taken place in the intervening two millennia between the Tell Kurdu population and those of Bronze Age Ebla and Alalakh. However, written sources, archaeological, and paleoclimatic evidence suggest that a narrower time period—the end of the EBA—had been very critical with respect to political tensions and population mobility. It was during this period, for example, that Ebla was destroyed twice and re-established at the beginning of MBA. There are extensive textual references from the end of the EBA through the LBA referring to groups of people arriving into the area of the Amuq Valley. Although these groups were named, likely based on designations (e.g., Amorites, Hurrians), the formative context of their (cultural) identity and their geographic origins remain debated. One recent hypothesis (Weiss, 2014, Weiss, 2017, Akar and Kara, 2020) associates the arrival of these groups with climate-forced population movement during the “4.2k BP event,” a Mega Drought that led to the abandonment of the entire Khabur river valley in Northern Mesopotamia and the search of nearby habitable areas.

Taking the above into consideration, we suggest that the ancestries we inferred for Alalakh and Ebla might best describe the genetic make-up of the EBA populations of unsampled Northern Mesopotamia. During the following MBA and LBA, we find no evidence of genetic disruption, even though shifts in territorial control dynamics between kingdoms/empires affected Ebla’s and Alalakh’s socio-cultural development (see STAR Methods). Nevertheless, the case of the Alalakh individual with likely Central Asian origin is a finding that can be interpreted within the context of nascent internationalism of the Middle and Late Bronze Age Eastern Mediterranean societies. It calls for future research on the various societal features of this phenomenon and how these are reflected on the individual life histories.

Conclusions
Overall, our large-scale genomic analysis reveals two major genetic events. First, during the Late Neolithic, gene pools across Anatolia and the Southern Caucasus mixed, resulting in an admixture cline. Second, during the Early Bronze Age, Northern Levant populations experienced gene flow in a process that likely involved yet to-be-sampled neighboring populations from Mesopotamia. Even though we could detect subtle and transient gene flow in Arslantepe, we acknowledge that disentangling questions related to local-scale population dynamics within the homogeneous Chalcolithic and Bronze Age Anatolian gene pool might be beyond the resolution of current analytical tools. Furthermore, while our sampling expands in number and geographic range on previous studies, the critical area of Mesopotamia remains unsampled; thus, although our picture of the genetic landscape of the Near East is highly suggestive, it remains incomplete. Nevertheless, the cumulative genetic dataset of Anatolia, the Southern Caucasus, and the Northern Levant between the Early and Late Bronze Age indicates that, following the genetic events of the Late Neolithic and Early Bronze Age, there was no intrusion of genetically distinct populations in this region. This conclusion is of great importance with respect to our understanding of the formation of complex Bronze Age socio-political entities.