Ancient DNA | Neolithic Anatolia | Gobekli Tepe

Who built Gobekli Tepe?

More than 11,000 years ago, hunter-gatherers in southeastern Turkey carved giant stone pillars and raised them in circles, thousands of years before Stonehenge or the pyramids. No DNA has yet been published from the site itself, but the genomes of the people who lived around it tell us who its builders most likely were. We tested them in Global25.

9600 BCapproximate date of the oldest monumental enclosures, built before farming, pottery or writing
4 ancestriesfrom Anatolia, the Levant, the Zagros and the Caucasus met in the people of the region, in a gradient from west to east
0 genomesfrom Gobekli Tepe itself so far; the first intact graves were only published in September 2026

Hero image: AI generated illustration of hunter-gatherers raising a T-shaped limestone pillar with wooden poles and ropes at Gobekli Tepe more than 11,000 years ago, among pillars carved with a fox, a boar, a snake and a bird of prey. Clothing, carvings and the scene are approximations, and it is not a reconstruction of a documented event or of the individuals whose genomes are discussed here.

Gobekli Tepe, the "pot-bellied hill" near the city of Sanliurfa, is one of the most famous archaeological sites in the world. Its stone enclosures, ringed by T-shaped pillars up to 5.5 metres tall and carved with foxes, boars, snakes and vultures, were built by people who still lived by hunting and gathering, some 11,600 years ago. It has been called the world's first temple, and it has inspired theories ranging from a lost civilisation to visitors from the stars.

The real question is simpler and more interesting: who were the people who built it, where did their ancestors come from, and what became of their descendants? No genome from Gobekli Tepe has been published yet. But since 2019, ancient DNA has been recovered from hunter-gatherers and early farmers all around it, from central Anatolia to northern Iraq. We read the studies and modelled the genomes ourselves.

Gobekli Tepe and its neighboursKey dates, from the last hunters of Anatolia to the first genomes of Upper Mesopotamiaabout 13,600 to 13,100 BCThe hunter of PinarbasiThe oldest Anatolian genome: a hunter-gatherer whosepeople became the first farmers of central Anatolia.about 9600 to 8000 BCGobekli TepeHunter-gatherers raise T-shaped pillars up to 5.5 metrestall in great stone enclosures near Sanliurfa.about 8750 BCBoncuklu Tarla and NemrikThe oldest genomes from Upper Mesopotamia, nearMardin in Turkey and in northern Iraq.8500 to 7500 BCCayonuThirteen genomes from a village on the upper Tigriswith monumental buildings of its own.about 7100 BCCatalhoyukThe great farming town of central Anatolia, whose peoplecarry new ancestry from the south and east.1995 to 2026Excavation and the deadKlaus Schmidt digs from 1995; carved skulls in 2017;the first intact graves published in 2026.Dates are approximate, from radiocarbon dating of each site.
Key dates around Gobekli Tepe, from the oldest Anatolian genome to the first burials published in 2026. Dates are approximate, from radiocarbon dating of each site.

A temple before farming?

The site was first noticed in a survey in 1963, when the broken limestone slabs on the hill were taken for medieval gravestones. In 1994 the German archaeologist Klaus Schmidt recognised what they were, and from 1995 until his death in 2014 he excavated the great enclosures. His interpretation made headlines: Gobekli Tepe was a sanctuary where scattered bands of hunter-gatherers gathered for feasts and rituals, and the effort of building it may have pushed people towards farming. First came the temple, then the city, as his famous summary put it.

That picture has since been refined. Under Lee Clare of the German Archaeological Institute and Turkish colleagues, excavations have found rainwater cisterns, domestic buildings and thousands of grinding stones, the signs of a place where people lived, not only met. Gobekli Tepe is also no longer unique: Karahan Tepe and a dozen other sites of the Sanliurfa region, the "Tas Tepeler" or stone hills, share its pillars and imagery. Gobekli Tepe became a UNESCO World Heritage Site in 2018. And genetic studies of wheat have shown that einkorn, one of the first domesticated cereals, was probably domesticated in the nearby Karacadag mountains. Whatever its builders believed, they lived at the very place and time where farming began.

The dead of Gobekli Tepe

For three decades, the only human remains from the site were hundreds of fragments of broken bone mixed into the fill of the great buildings. In 2017, Julia Gresky and colleagues showed that three of the skull fragments had been carved with deep grooves, one also drilled, and suggested a skull cult. In September 2026, Gresky and Clare published the first two intact burials, dated between about 8700 and 8000 BC: a pit beneath the plaster floor of a building holding a woman, a young man and an adolescent girl, and the single grave of a woman laid on her side. They now count incised skull fragments from five people, and they suggest that many of the 928 loose bone fragments came from ordinary burials higher on the mound, carried down into the buildings by erosion.

These graves are the first real hope of DNA from Gobekli Tepe itself. Until it is published, the answer to our question must come from the neighbours.

The neighbours have spoken

Three studies matter most. In 2019, Michal Feldman and colleagues sequenced a hunter-gatherer from Pinarbasi in central Anatolia, who lived about 15,000 years ago, and showed that the first farmers of central Anatolia descended for 80 to 90% of their ancestry from local hunters like him: farming there was adopted by locals, not brought by migrants. In 2022, N. Ezgi Altinisik and colleagues published 13 genomes from Cayonu, a Neolithic village on the upper Tigris, about 150 km northeast of Gobekli Tepe, occupied between 8500 and 7500 BC and famous for its own monumental buildings, including a "skull building" that held the remains of hundreds of people. The same year, Iosif Lazaridis, David Reich and colleagues published the oldest genomes of Upper Mesopotamia, from Boncuklu Tarla near Mardin, about 250 km east of Gobekli Tepe, and from Nemrik 9 in northern Iraq, both from around 8750 BC.

The common finding is that the people of Upper Mesopotamia were genetically in between. They carried ancestry related to the hunters and first farmers of Anatolia, to the Natufians and early farmers of the Levant, to the early farmers of the Zagros mountains of Iran, and to the hunter-gatherers of the Caucasus. Lazaridis and colleagues also found that ancestry from Upper Mesopotamia spread west into Anatolia, in distinct movements during the Pre-Pottery and the Pottery Neolithic.

In the middle of everything

We placed the hunter-gatherers and early farmers of West Asia on a local principal component analysis built from Global25 group averages, then projected the six Upper Mesopotamian genomes onto it. The first axis, which captures 77% of the variation, runs from west (Anatolia and the Levant) to east (Zagros and Caucasus).

In the middle of everythingLocal PCA of early Neolithic and late hunter-gatherer genomes of West Asia, Global25 (our analysis)Pinarbasi, 13,000 BCBoncuklu (C. Anatolia)Asikli (C. Anatolia)CatalhoyukBarcin (NW Anatolia)Natufians (Levant)Levant PPNB (Israel)Levant PPNB (Jordan)Ganj Dareh (Zagros)Abdul Hosein (Zagros)Wezmeh (Zagros)Kotias (Caucasus HG)Satsurblia (Caucasus HG)Nemrik 9, IraqBoncuklu Tarla, MardinCayonu, migrant childCayonu (3 genomes)PC1: west (Anatolia, Levant) to east (Zagros, Caucasus), 77% of the variationAnatoliaLevantZagrosCaucasusUpper MesopotamiaUpper Mesopotamian genomes are projected onto the axes built from the other groups. Coverage 7 to 39%: positions are approximate.
Local principal component analysis of Global25 group averages of early Neolithic farmers and late hunter-gatherers of West Asia, with the six Upper Mesopotamian genomes projected onto it (our analysis). Their coverage ranges from 7 to 39%, so their positions are approximate.

The Upper Mesopotamians fall exactly where their geography predicts: in the middle. The genomes from Cayonu sit closer to Anatolia and the Levant, those of Boncuklu Tarla and Nemrik further east, towards the Zagros. These genomes are of low coverage, from 7 to 39% in our data, so each position is approximate, but the pattern is clear and agrees with the published studies.

We then modelled each genome as a mixture of four sources: the early farmers of Boncuklu in central Anatolia, those of the Levant (Jordan, Pre-Pottery Neolithic B), those of Ganj Dareh in the Zagros, and the hunter-gatherers of Kotias cave in Georgia.

Four ancestries in one villageUpper Mesopotamian genomes modelled with Central Anatolian, Levantine, Zagros and Caucasus sources (our analysis)0%25%50%75%100%fitCayonu, cay007 (8400 BC)31491439Cayonu, cay013 (7740 BC)28531875Cayonu, cay004 (7590 BC)17602258Cayonu, migrant child cay0081543113256Boncuklu Tarla, Mardin (8750 BC)55252051Nemrik 9, Iraq (8750 BC)46332139Upper Mesopotamia average1155152046Central Anatolia (Boncuklu)Levant (PPNB)Zagros (Ganj Dareh)Caucasus (Kotias)Single genomes of low coverage (7 to 39%): read the west to east gradient, not the exact shares.
Four-source models of the Upper Mesopotamian genomes, with central Anatolian (Boncuklu), Levantine (Jordan PPNB), Zagros (Ganj Dareh) and Caucasus (Kotias) sources; fit distances x 1000 on the right (our analysis). Exact shares move by several points with other sources; the west to east gradient does not.

Every one of them needs all four directions, but in different doses. In the three Cayonu genomes, the eastern share (Zagros plus Caucasus) is about 20%. At Boncuklu Tarla it rises to 45%, and at Nemrik 9, in Iraq, to 54%. One child from Cayonu, cay008, stands apart from the rest of the village, with 43% eastern ancestry: someone whose family had come from further east, a reminder that these communities were connected by movement. The large Levantine share in all of them should not be read as migration from Jordan; it stands for a southern ancestry that was widespread across the Fertile Crescent. With single genomes of this coverage, the exact shares move by several points with the choice of sources; the west to east gradient does not.

We also checked the spread of this ancestry into Anatolia. The Pottery Neolithic farmers of Tepecik-Ciftlik, in Cappadocia, fit better with an Upper Mesopotamian source added to central Anatolian ancestry (fit distance 32.8) than with a Levantine one (40.2), while those of Catalhoyuk and of Barcin in the northwest show no such gain. This agrees with the published finding that people from the east reached central Anatolia during the Neolithic.

So who built Gobekli Tepe?

Gobekli Tepe lies at the western end of this gradient, about 150 km southwest of Cayonu. If the pattern holds, its builders were hunter-gatherers whose ancestry was mostly the western, Anatolian and Levantine kind, with a minority of ancestry from the Zagros and the Caucasus, probably close to the people of Cayonu or a little more western. They were not a lost race, not strangers from afar and not a single tribe, but local people at a crossroads where several ancestral streams had been meeting for thousands of years.

That position may help explain the monuments themselves. Gobekli Tepe sits in a region where people from different directions met, exchanged obsidian from central and eastern Anatolia, and shared a striking repertoire of symbols over hundreds of kilometres. Building enclosures of this size required many hands and long-distance networks; the genomes show that those networks moved people as well as ideas. This is our interpretation, and genomes alone cannot prove it, but it fits what archaeologists see on the ground.

The descendants of these people did not disappear. Upper Mesopotamian ancestry spread into Anatolia, and from there, with the Anatolian farmers, part of it reached Europe; it also contributed to the later populations of the Near East. In a sense, many people in Europe and West Asia today carry a small inheritance from the world of Gobekli Tepe.

Who looks most like them today?

We compared the average of the four best-covered Upper Mesopotamian genomes, from Cayonu, Boncuklu Tarla and Nemrik 9, with present-day populations in Global25.

Who looks most like the people of the Gobekli Tepe era?Global25 distance from the Upper Mesopotamian Neolithic average to present-day populations (our analysis)020406080100Lebanese Sunni (Dinniyeh)36Lebanese Maronite (Zgharta)39Lebanese Orthodox (Koura)39Samaritans43Iraqi Jews44Druze45Armenians of Urfa50Assyrians of Mardin55Cypriots57Kurds of Turkey (Kurmanji)94Turks of Kayseri94Distances x 1000. Gold: communities from the region of Gobekli Tepe itself. No living group is truly close.
Global25 distance from the average of the four best-covered Upper Mesopotamian genomes (Cayonu cay007 and cay013, Boncuklu Tarla, Nemrik 9) to present-day populations (our analysis).

The closest living populations are in the Levant: Lebanese communities, Samaritans, Iraqi Jews and Druze, at 36 to 45 units. Then come the Armenians whose families came from Urfa, the city that is today Sanliurfa, at 50, and the Assyrians of Mardin at 55. Kurds and Turks of Anatolia are further, at 94. No living group is truly close: 36 units is a large distance, and 11,000 years of history, including Bronze Age movements from the Caucasus and the steppe and, for Anatolian Turks, medieval migrations from Central Asia, have reshaped every population of the region. Levantine populations simply kept a larger share of the old Neolithic mixture.

These distances describe similarity, not ownership. Everyone who lives in southeastern Turkey today inherits the history of the land in one way or another, and no modern group can claim to be the sole heir of Gobekli Tepe.

Some questions remain open. Without DNA from Gobekli Tepe, we cannot say whether its builders were exactly like the people of Cayonu. The Upper Mesopotamian genomes are few and of low coverage. And DNA cannot tell us what the carved animals meant, or what language the builders spoke. The burials published in 2026 may soon change part of this story.

Key points

  • Gobekli Tepe was built by hunter-gatherers from about 9600 BC; no DNA from the site has been published, but the first intact burials were reported in 2026.
  • Genomes from nearby Cayonu, Boncuklu Tarla and Nemrik 9 show that the people of Upper Mesopotamia mixed Anatolian, Levantine, Zagros and Caucasus ancestry.
  • In our Global25 analysis, the eastern share rises from about 20% at Cayonu to 45% at Boncuklu Tarla and 54% at Nemrik 9: a gradient from west to east.
  • Gobekli Tepe lies at the western end of that gradient, so its builders probably carried mostly Anatolian and Levantine-related ancestry with an eastern minority.
  • Upper Mesopotamian ancestry spread west into Anatolia; in our tests, Pottery Neolithic Tepecik-Ciftlik fits better with it than with a Levantine source.
  • No living population is close; the nearest today are Levantine communities, followed by Armenians of Urfa and Assyrians of Mardin.

Global25 coordinates used in this article

The Upper Mesopotamian genomes and their average, the four model sources and a few present-day populations. They can be pasted directly into the Vahaduo tool.

Upper_Mesopotamia_PPN_avg6,0.080814,0.153514,-0.058202,-0.095931,-0.016978,-0.034675,-0.003173,-0.007961,-0.013839,0.012483,0.012666,-0.000874,0.001412,0.007798,-0.016626,-0.004331,-0.000804,0.006440,0.005950,0.003752,0.004638,0.013994,-0.006429,-0.010102,0.004071
Upper_Mesopotamia_PPN_avg4_bestcovered,0.080530,0.151568,-0.062130,-0.096900,-0.018542,-0.033536,-0.000059,-0.009404,-0.012425,0.011344,0.012504,-0.002548,0.003865,0.008636,-0.015031,-0.002784,-0.002477,0.005764,0.005719,-0.000906,0.005241,0.014344,-0.007795,-0.010905,0.005359
Nemrik9_Iraq_I6445_8750BC,0.064879,0.135065,-0.084852,-0.068476,-0.049548,-0.016733,0.002585,-0.012461,-0.034155,-0.017312,0.006658,-0.002847,0.012636,0.003303,-0.007872,0.010740,0.004955,0.011149,0.003520,-0.004127,0.014974,0.007172,-0.008134,-0.015424,0.007185
BoncukluTarla_Mardin_I8432_8750BC,0.072847,0.136081,-0.084852,-0.096900,-0.046778,-0.034582,-0.004700,-0.011769,-0.019839,-0.010205,0.021273,-0.004946,0.010109,0.009771,-0.002850,-0.000133,-0.004172,0.001774,-0.003017,0.001126,0.005366,0.012489,-0.000616,-0.021208,0.013292
Cayonu_cay007_8400BC,0.095611,0.159438,-0.040352,-0.103360,0.009848,-0.042112,-0.003760,-0.010846,0.007158,0.036447,0.010555,0.001199,-0.004014,0.007569,-0.026194,-0.003713,0.003520,0.003294,0.005405,0.001376,0.000624,0.008408,-0.005916,-0.000602,-0.002874
Cayonu_cay013_7740BC,0.088782,0.175687,-0.038466,-0.118865,0.012310,-0.040718,0.005640,-0.002538,-0.002863,0.036447,0.011530,-0.003597,-0.003271,0.013900,-0.023208,-0.018032,-0.014212,0.006841,0.016969,-0.002001,0.000000,0.029306,-0.016515,-0.006386,0.003832
Boncuklu_CentralAnatolia_N,0.116327,0.169187,0.013576,-0.070737,0.055210,-0.036423,-0.003337,-0.003138,0.034237,0.078835,0.009483,0.014897,-0.030148,0.002615,-0.038762,-0.005595,0.029441,0.005143,0.010886,-0.009054,-0.007736,0.007666,-0.007296,-0.010748,-0.005868
Levant_PPNB_Jordan,0.075692,0.161977,-0.029887,-0.134611,0.031544,-0.061356,-0.007520,-0.007788,0.076339,0.038543,0.020380,-0.015062,0.027205,0.001273,-0.027280,0.002718,0.015548,0.001045,-0.007007,0.013038,-0.003837,0.005935,0.003821,-0.004820,-0.003323
GanjDareh_Zagros_N,0.043025,0.067431,-0.153488,0.005556,-0.123962,0.024375,0.015464,0.000277,-0.081605,-0.054270,-0.003248,-0.001619,0.005382,-0.007845,0.031949,0.056775,-0.005815,0.007576,0.014405,-0.032741,0.007661,-0.030048,-0.010920,-0.038777,0.022944
Kotias_Caucasus_HG,0.089920,0.103584,-0.088246,-0.004845,-0.089555,0.020359,0.023266,-0.001154,-0.130486,-0.080002,-0.007632,0.024878,-0.052626,0.007707,0.026465,-0.029700,0.027772,-0.010769,-0.023757,0.037518,0.030945,-0.006677,0.005300,-0.020846,0.001557
Pinarbasi_Anatolia_HG,0.114961,0.167562,0.030170,-0.080750,0.072629,-0.039881,-0.000470,0.000231,0.046836,0.083282,0.004222,0.007793,-0.022150,-0.011836,-0.041123,-0.003845,0.043027,0.001140,0.008799,-0.004252,-0.011729,0.006801,-0.006286,-0.010122,-0.007784
Lebanese_Sunni_Dinniyeh_today,0.090020,0.144722,-0.052519,-0.079458,-0.014972,-0.028373,-0.001864,-0.006587,0.000524,0.007670,0.007222,-0.004914,0.010156,0.004088,-0.009127,0.002419,-0.004300,0.002327,0.002212,-0.002481,0.000105,0.003180,-0.002048,0.001831,-0.000326
Lebanese_Maronite_Zgharta_today,0.090072,0.143446,-0.052730,-0.078836,-0.012616,-0.027342,-0.001917,-0.006092,0.002606,0.008529,0.006113,-0.004902,0.011294,0.003043,-0.008313,0.004614,-0.002487,0.001532,0.002984,-0.002396,-0.000208,0.002539,-0.001719,0.001531,0.001252
Armenian_Urfa_today,0.101140,0.139273,-0.059585,-0.070507,-0.026730,-0.019682,0.003760,-0.005143,-0.026121,0.002421,0.003596,0.002034,0.000425,0.004247,-0.007775,0.000890,0.001639,0.000163,0.002083,-0.002590,0.000267,0.000707,-0.001409,-0.001119,0.001061
Assyrian_Mardin_today,0.097319,0.137604,-0.058642,-0.069122,-0.029390,-0.017849,-0.001058,-0.007846,-0.015442,-0.005832,0.005440,-0.001499,0.006913,0.001927,-0.006447,0.010939,0.009583,0.000190,-0.001320,-0.002751,0.002246,0.001546,-0.001541,0.005965,-0.001617
Kurd_Kurmanji_Turkey_today,0.097649,0.111926,-0.055491,-0.036891,-0.037255,-0.005001,0.004015,-0.004198,-0.028938,-0.011134,0.000509,-0.000983,0.001305,-0.003528,0.004460,0.008611,-0.003112,0.003259,0.004906,-0.009111,-0.001182,-0.001721,-0.000945,-0.002347,0.002834
Turkish_Kayseri_today,0.099514,0.093864,-0.034129,-0.043513,-0.024818,-0.012809,0.004482,-0.000676,-0.022746,0.001341,-0.001113,0.001220,-0.003780,0.001740,-0.004518,-0.000322,0.000084,0.000778,0.001401,-0.002957,-0.000811,0.001599,-0.002007,-0.002264,-0.000402

References

  1. Ancient DNA Altinisik NE, Kazanci DD, Aydogan A, Gemici HC, Erdal OD, et al., Somel M (2022). A genomic snapshot of demographic and cultural dynamism in Upper Mesopotamia during the Neolithic Transition. Science Advances 8(44): eabo3609. doi.org/10.1126/sciadv.abo3609
  2. Ancient DNA Lazaridis I, Alpaslan-Roodenberg S, Acar A, Acikkol A, Agelarakis A, et al., Reich D (2022). Ancient DNA from Mesopotamia suggests distinct Pre-Pottery and Pottery Neolithic migrations into Anatolia. Science 377(6609): 982 to 987. doi.org/10.1126/science.abq0762
  3. Ancient DNA Feldman M, Fernandez-Dominguez E, Reynolds L, Baird D, Pearson J, et al., Krause J (2019). Late Pleistocene human genome suggests a local origin for the first farmers of central Anatolia. Nature Communications 10: 1218. doi.org/10.1038/s41467-019-09209-7
  4. Bioarchaeology Gresky J, Clare L (2026). First burials give insights into mortuary rites at Gobeklitepe, Early Neolithic Turkiye. PLOS One 21(9): e0358979. doi.org/10.1371/journal.pone.0358979
  5. Bioarchaeology Gresky J, Haelm J, Clare L (2017). Modified human crania from Gobekli Tepe provide evidence for a new form of Neolithic skull cult. Science Advances 3(6): e1700564. doi.org/10.1126/sciadv.1700564
  6. Archaeobotany Heun M, Schafer-Pregl R, Klawan D, Castagna R, Accerbi M, Borghi B, Salamini F (1997). Site of einkorn wheat domestication identified by DNA fingerprinting. Science 278(5341): 1312 to 1314. doi.org/10.1126/science.278.5341.1312
  7. Archaeology Schmidt K (2012). Gobekli Tepe: A Stone Age Sanctuary in South-Eastern Anatolia. Berlin: ex oriente.
  8. Data Global25 coordinates by Davidski (Eurogenes), and the Moriopoulos Collection 2026 of averages. Admixture models and figures: ExploreYourDNA.