ArticlePDF Available

The genetic origin of the Indo-Europeans

February 2025
Nature 639(8053):132-142

DOI:10.1038/s41586-024-08531-5

Authors:

Iosif Lazaridis

Harvard Medical School

David Anthony

Hartwick College

Leonid Vyazov

University of Ostrava

Show all 94 authorsHide

The Yamnaya archaeological complex appeared around 3300 bc across the steppes north of the Black and Caspian Seas, and by 3000 bc it reached its maximal extent, ranging from Hungary in the west to Kazakhstan in the east. To localize Yamnaya origins among the preceding Eneolithic people, we assembled ancient DNA from 435 individuals, demonstrating three genetic clines. A Caucasus–lower Volga (CLV) cline suffused with Caucasus hunter-gatherer¹ ancestry extended between a Caucasus Neolithic southern end and a northern end at Berezhnovka along the lower Volga river. Bidirectional gene flow created intermediate populations, such as the north Caucasus Maikop people, and those at Remontnoye on the steppe. The Volga cline was formed as CLV people mixed with upriver populations of Eastern hunter-gatherer² ancestry, creating hypervariable groups, including one at Khvalynsk. The Dnipro cline was formed when CLV people moved west, mixing with people with Ukraine Neolithic hunter-gatherer ancestry³ along the Dnipro and Don rivers to establish Serednii Stih groups, from whom Yamnaya ancestors formed around 4000 bc and grew rapidly after 3750–3350 bc. The CLV people contributed around four-fifths of the ancestry of the Yamnaya and, entering Anatolia, probably from the east, at least one-tenth of the ancestry of Bronze Age central Anatolians, who spoke Hittite4,5. We therefore propose that the final unity of the speakers of ‘proto-Indo-Anatolian’, the language ancestral to both Anatolian and Indo-European people, occurred in CLV people some time between 4400 bc and 4000 bc.

Three Eneolithic clines and their neighbours in space and time a, Map with analysed sites. b, PCA using axes formed by a set of ancient West European hunter-gatherer and Siberian, West Asian and European farmer populations. Selected individuals relevant to this study are projected (Methods). c, qpAdm models fitted on individuals of the populations of the clines. The Volga cline is generated by admixture between lower Volga (BPgroup) people with upriver EHG populations. People of the Dnipro cline have UNHG or UNHG + EHG admixture relative to the Core Yamnaya (the hunter-gatherer source along this cline is significantly variable). The Caucasus–lower Volga cline is generated by admixture of lower Volga people with those from the Neolithic Caucasus (Aknashen related). The map was drawn using public-domain Natural Earth data with the rnaturalearth package in R⁵⁴.

…

The three clines in the context of Eneolithic and Bronze Age admixture Six three-source qpAdm models elucidate a complex history of admixture. a, Caucasus and European hunter-gatherer admixtures in the ‘old steppe’: Krivyansky on the lower Don received much more CHG-related admixture than did upriver people of the middle Don at Golubaya Krinitsa. In the middle and upper Volga and the Kama River, populations had negligible CHG-related influence. b, The Don–Volga difference. On the lower Volga and North Caucasus piedmont, the BPgroup received CHG-related ancestry, similar to its western lower Don counterpart at Krivyansky. But it also received ancestry from Central Asia, and this eastern influence was greater still in the Bronze Age steppe Maikop. c, The Volga basin Eneolithic populations with regard to the Don. Populations at Khvalynsk, Klopkov Bugor and Ekaterinovka form a Volga cline between the Berezhnvoka cluster on the lower Volga and the upriver EHG-like populations of the middle Volga (Labazy and Lebyazhinka). d, The Volga basin Eneolithic populations with regard to Central Asia. There is a slight excess of Central Asian ancestry in the Khi subset of Khvalynsk. e, The Dnipro cline. The Core Yamnaya are on one end of a cline that also includes the Don Yamnaya and Serednii Stih populations, formed by admixture of North Pontic hunter-gatherers with those from the CLV cline of differential admixture of Neolithic Caucasus and BPgroup people. The CLV cline includes diverse people buried in kurgans at Berezhnovka, Progress-2, Remontnoye and the Maikop sites of Klady and Dlinnaya-Polyana, dating from around 5000–3000 bc. f, West Asia. CLV ancestry first appears in the Chalcolithic population at Areni-1 in Armenia and is also present in the Bronze Age at Maikop. Most of the ancestry is from West Asian sources from the Mesopotamia–Caucasus (or Çayönü–Masis Blur–Aknashen) cline. Chalcolithic and Bronze Age Anatolians lack CLV ancestry, but traces of it can be found in Bronze Age central Anatolians.

…

Patrilineal succession Temporal distribution of key Y-chromosome haplogroups from Kazakhstan, Kyrgyzstan, Mongolia, Russia, Turkmenistan, Ukraine, Uzbekistan and comparative regions of Europe and West Asia, 6000–1000 bc. The Early and Middle Bronze Age group includes the Yamnaya, Afanasievo, Poltavka, Catacomb, Chemurchek and North Caucasus cultures; the Middle and Late Bronze Age group includes individuals of diverse cultures down to 1000 bc, including those of the Sintashta, Andronovo, Potapovka and Srubnaya cultures. Information on which individuals are plotted can be found in Supplementary Information (Supplementary Table 6).

…

Trajectory of the Yamnaya expansion We used HapNe-LD to estimate the changes in effective population size (Ne) over time of Yamnaya ancestors. We carried out the computation separately for the individuals from the earlier 300 years of our sampling and the later 300 years; shading shows confidence intervals (dark, 50%; light, 95%). Jointly displaying these two trajectories reveals an extraordinary population expansion at 3642–3374 bc (intersection of 95% confidence intervals for the two analyses for the minimum), from when the effective size is a few thousand to an order of magnitude larger. The offset on the x axis is due to the difference in sampling time between the two groups.

…

IBD analysis of the Yamnaya and their predecessors a,b, Pairs of individuals linked by at least one IBD segment at least 20 cM in length reveal a sparse but highly connected network in the pre-Yamnaya (Methods) (a) and Yamnaya (b) groups. No detectable IBD is found in the pre-Yamnaya period beyond the scale of 1,000 km. c,d, Yamnaya share more IBD with each other at short distance scales compared with the pre-Yamnaya people (c), but IBD sharing extends all the way to the roughly 6,000 km scale of their geographical distribution (d). e,f, However, closely related individuals occur only at short distance scales in both pre-Yamnaya (e) and Yamnaya (f) groups, indicating that IBD sharing in the Yamnaya was a legacy of their common origin. In c–f, two-sided 95% confidence intervals are shown as a vertical interval (at distance = 0) or a rectangle (at distance ranges greater than 0); the fraction of number of pairs of individuals sharing IBD (I)/total number of pairs of individuals (T) is shown in red. g, In a set of 9 Yamnaya cemeteries and a total of 25 kurgans, closely or distantly related individuals are almost absent in inter-cemetery comparisons, more are found in inter-kurgan and within-cemetery comparisons, and even more are found in intra-kurgan comparisons; nonetheless, most Yamnaya individuals in all comparisons were unrelated. Kurgan burial of close kin was less common than in the case of a local patrilineal dynasty, as at a Neolithic long cairn at Neolithic Hazleton North⁴⁵, but was more common than in Neolithic monuments in Ireland⁵⁵. Two-sided 95% confidence intervals are shown. The map was drawn using public-domain Natural Earth data with the rnaturalearth package in R⁵⁴.

…

Figures - available from: Nature

This content is subject to copyright. Terms and conditions apply.

Content uploaded by Harald Ringbauer

Content may be subject to copyright.

Nature | www.nature.com | 1

Article

The genetic origin of theIndo-Europeans

Iosif Lazaridis1,2,67 ✉, Nick Patterson1,3,67 ✉, David Anthony1,4,67 ✉, Leonid Vyazov1,5,67 ✉,

Romain Fournier6, Harald Ringbauer1,7, Iñigo Olalde1,8,9, Alexander A. Khokhlov10,

Egor P. Kitov11, Natalia I. Shishlina12, Sorin C. Ailincăi13, Danila S. Agapov14, Sergey A. Agapov14,

Elena Batieva15, Baitanayev Bauyrzhan16, Zsolt Bereczki17, Alexandra Buzhilova18,

Piya Changmai5, Andrey A. Chizhevsky19, Ion Ciobanu20, Mihai Constantinescu21,

Marietta Csányi22, János Dani23,24, Peter K. Dashkovskiy25, Sándor Évinger26, Anatoly Faifert27,

Pavel Flegontov1,5,28, Alin Frînculeasa29, Mădălina N. Frînculeasa30, Tamás Hajdu31,

Tom Higham32,33, Paweł Jarosz34, Pavol Jelínek35, Valeri I. Khartanovich36,

Eduard N. Kirginekov37, Viktória Kiss38, Alexandera Kitova39, Alexeiy V. Kiyashko40,

Jovan Koledin41, Arkady Korolev10, Pavel Kosintsev42,43, Gabriella Kulcsár38, Pavel Kuznetsov10,

Rabadan Magomedov44, Aslan M. Mamedov16, Eszter Melis38, Vyacheslav Moiseyev36,

Erika Molnár17, Janet Monge45, Octav Negrea29, Nadezhda A. Nikolaeva46, Mario Novak47,48,

Maria Ochir-Goryaeva49, György Pálfi17, Sergiu Popovici50, Marina P. Rykun51,

Tatyana M. Savenkova52, Vladimir P. Semibratov53, Nikolai N. Seregin54, Alena Šefčáková55,

Raikhan S. Mussayeva16, Irina Shingiray56, Vladimir N. Shirokov57, Angela Simalcsik20,58,

Kendra Sirak1,2, Konstantin N. Solodovnikov59, Judit Tárnoki22, Alexey A. Tishkin53,

Viktor Trifonov60, Sergey Vasilyev61, Ali Akbari1,2, Esther S. Brielle1, Kim Callan2,62,

Francesca Candilio63, Olivia Cheronet32,33, Elizabeth Curtis2,62, Olga Flegontova5,28,

Lora Iliev2,62, Aisling Kearns2, Denise Keating64, Ann Marie Lawson2,62, Matthew Mah2,3,62,

Adam Micco2, Megan Michel1,2,7,62, Jonas Oppenheimer2,62, Lijun Qiu2,62, J. Noah Workman2,62,

Fatma Zalzala2,62, Anna Szécsényi-Nagy65, Pier Francesco Palamara6,66, Swapan Mallick2,3,62,

Nadin Rohland2, Ron Pinhasi32,33 ✉ & David Reich1,2,3,62 ✉

The Yamnaya archaeological complex appeared around 3300 BC across the steppes

north of the Black and Caspian Seas, and by 3000 BC it reached its maximal extent,

ranging from Hungary in the west to Kazakhstan in the east. To localize Yamnaya

origins among the preceding Eneolithic people, we assembled ancient DNA from 435

individuals, demonstrating three genetic clines. A Caucasus–lower Volga (CLV) cline

suused with Caucasus hunter-gatherer1 ancestry extended between a Caucasus

Neolithic southern end and a northern end at Berezhnovka along the lower Volga river.

Bidirectional gene ow created intermediate populations, such as the north Caucasus

Maikop people, and those at Remontnoye on the steppe. The Volga cline was formed

as CLV people mixed with upriver populations of Eastern hunter-gatherer2 ancestry,

creating hypervariable groups, including one at Khvalynsk. The Dnipro cline was

formed when CLV people moved west, mixing with people with Ukraine Neolithic

hunter-gatherer ancestry3 along the Dnipro and Don rivers to establish Serednii Stih

groups, from whom Yamnaya ancestors formed around 4000 BC and grew rapidly

after 3750–3350 BC. The CLV people contributed around four-fths of the ancestry of

the Yamnaya and, entering Anatolia, probably from the east, at least one-tenth of the

ancestry of Bronze Age central Anatolians, who spoke Hittite4,5. We therefore propose

that the nal unity of the speakers of ‘proto-Indo-Anatolian’, the language ancestral

to both Anatolian and Indo-European people, occurred in CLV people some time

between 4400 BC and 400 0 BC.

Between 3300 BC and 1500 BC, people of the Yamnaya archaeologi-

cal complex and their descendants spread Indo-European languages

from the steppe

2,6–12

and transformed Europe, Central and South Asia,

Siberia and the Caucasus. Sparse sampling of Yamnaya people and their

Eneolithic precursors creates a problem for understanding the origins

of this Bronze Age culture. It is broadly accepted that the Yamnaya

had two ancestries: northern, eastern hunter-gatherer (EHG) ancestry

from far-eastern Europe, and southern, West Asian ancestry2 from

Caucasus hunter-gatherers (CHG) in Georgia1 and Neolithic people

from Zagros

andthe south Caucasus

10,14,15

. These two groups inter-

acted across West Asia and eastern Europe13, but it has not been clear

where or how the Eneolithic ancestors of the Yamnaya first appeared.

Potential northern ancestors include the EHG, and EHG mixed

with western hunter-gatherers16 (WHG), for example in the Dnipro

https://doi.org/10.1038/s41586-024-08531-5

Received: 16 April 2024

Accepted: 17 December 2024

Published online: xx xx xxxx

Check for updates

A list of affiliations appears at the end of the paper.

2 | Nature | www.nature.com

Article

valley3, where they formed the Ukraine Neolithic hunter-gatherers

(UNHG). But the Yamnaya also received Anatolian Neolithic ances-

try9, mediated by Caucasus Neolithic populations, such as those

sampled at Aknashen and Masis Blur in Armenia10, and even possi-

bly Siberian ancestry that reached the European steppe before their

emergence9.

We present a genetic analysis of 367 newly reported individuals

(6400–2000 BC) and increased data quality for 68 individuals6

(a total of 435individuals). The present study is the formal report for

291 and 63 of these, respectively; more than 80% are from Russia, and

the rest are largely from the western expansion into the Danube valley

(Supplementary Information section1 and Supplementary Table1).

Details of 803 ancient DNA libraries (195 that failed screening) are in

Supplementary Information section 1 and Supplementary Table2, and

198 new radiocarbon dates are in Supplementary Table3. A parallel

study17 of the North Pontic Region (Ukraine and Moldova) is the for-

mal report for the remaining individuals. We labelled individuals on

the basis of geographical and temporal information, archaeological

context and genetic clustering (Supplementary Information section 1

and Supplementary Table4). The combined dataset adds 79 Eneolithic

people from the European steppe and its environs to 82 published. It

also adds 211 Yamnaya (and related Afanasievo) individuals to the 75

previously published (Methods).

Three pre-Bronze Age genetic clines

Principal component analysis (PCA) of ancient individuals from the

Ponti c–Casp ian step pe and adj acent are as reveal s that Ene olithi c peo-

ple and the Bronze Age Yamnaya fall on non-overlapping gradients

(Fig.1 and Supplementary Table5). PC1 correlates (right to left) to

differentiation between inland West Asian (Caucasus and Iran) and

East Mediterranean populations (Anatolian–European)

, but interpre-

tation is not clear because this axis also correlates to differentiation

between Siberian and European hunter-gatherers. PC2 differentiates

between northern Eurasians (top, including Europe and Siberia) and

Wes t As ia ns ( bo tto m, Ana to lia –M eso po tam ia –Ca uc asu s– Ira n). En eo-

lithic and Bronze Age people occupy the middle, indicating that they

formed by mixture.

To distinguish alternative mixture scenarios that could explain these

patterns, we implemen ted a competition frame work around qpWave/

qpAdm

2,18

(Methods and Supplementary Information section2). The

idea is that model X (a s et o f ad mix in g so urce s) desc ri bes a ta rget pop u-

lation T if: i t re con str uct s th e sh are d ge net ic d rif t of T with both distant

outgroup populations and the sources of alternative models; and also

renders these models infeasible if they cannot model shared drift with

the sources of X. Models are thus first filtered against a set of distant

outgroups; having survived this step, they are compared all-against-all

to produce a set of promising models.

Three PCA clines (denoted geographically as Volga, Dnipro and

Caucasus–lower Volga) diverge from the area enclosed by the Lower

Don (at Krivyansky), lower Volga (at Berezhnovka-2) and north Cau-

casus (at Progress-2, Vonyuchka-1 and Sharakhalsun

). They extend

from there towards: EHG and UNHG, representing the pre-Eneolithic

people of the Volga–Don–Dnipro area of eastern Europe; and CHG

and Caucasus Neolithic, representing the pre-Eneolithic people of the

Caucasus and West Asia.

The Volga cline

Distinct upriver and downriver gradients formed by Eneolithic

individuals who lived on waterways that drain into the Caspian Sea

delineate zones of ongoing human contact. PCA positions correlate

well to positions along the Volga: the Volosovo-attributed Sakhtysh

(in the upper Volga) and Murzikha (near the Kama–Volga confluence)

constitute the upriver European hunter-gatherer cline, between EHG

and UNHG. A ‘bend’ separates the two clines and is occupied by EHG

groups, including middle Volga ones and those from northwest Russia

in Karelia2,20, which is a very wide geographic distribution indicating

that EHG was the earlier established population. Downriver and past

the bend, we find the Volga cline: hunter-gatherer affinity decreases

at the middle Volga at Labazy, Lebyazhinka, Ekaterinovka, Syezzheye

then Khvalynsk (4500–4350 BC) and Khlopkov Bugor, before reach-

ing the lower Volga at Berezhnovka-2 (4450–3960 BC) (Fig.1a,b). This

decrease is counterbalanced by increased affinity to the Caucasus,

driven by an unsampled CHG-related source, somewhere between

Georgia (the sampling location of CHG1) and the lower Volga, inter-

acting with EHG people. Archaeological correlates for such interac-

tions begin with the expansion of the Seroglazovo forager culture

around the lower Volga estuary in around 6200 BC, which parallels

cultures of the Caucasus in ceramics and lithics, and continue to the

north Caucasus Neolithic cemetery near Nalchik, dated to around

4800 BC21,22.

At the end of the Volga cline, four lower-Volga individuals from

Berezhnovka-2 can be grouped with the north Caucasus PG2004 individ-

ual from Progress-2 (ref. 9), dated to 4240–4047 BC, into a Berezhnovka 2–

Progress -2 c luster labelled the BPgroup. The s econd P rogress-2

individual (PG2001; 4994–4802 cal BC) groups with another north

Caucasus individual from Vonyuchka-1 (ref. 9; VJ1001; 4337–4177 BC)

into a Progress-2–Vonyuchka 1 cluster (the PVgroup). The BPgroup

and PVgroup are distinct (P = 0.0006) but little differentiated (fixation

indexF

= −0.002 ± 0.002; Extended Data Table1), indicating move-

ment between the north Caucasus piedmont and the lower Volga. These

two locations also shared a distinctive burial pose, on the back with

raised knees, which was later typical of the Yamnaya and dated earliest

in four individuals from Ekaterinovka (4800–4500 BC), contrasting

with 95% of the graves, which had individuals posed supine with legs

extended straight, and also a female (individual 2) from Lebyazhinka-5,

grave 12 (4838–4612 BC). BPgroup is shifted relative to PVgroup (Fig.1b)

towards Afontova Gora-3 from Upper Palaeolithic Siberia

, West Sibe-

rian hunter-gatherers8 and a Neolithic individual dated at 7,500 years

ago from Tutkaul (TTK) from Central Asia20.

A natural interpretation is that upriver, EHG-related, and downriver,

Berezhnovka-related, ancestors came together along the Volga, form-

ing the genetic gradient. The upriver ancestry has long-established east-

ern European antecedents20, unlike the downriver ancestry, because:

first, there are no earlier sequenced individuals from the lower Volga;

second, the Berezhnovka people are distinct from preceding groups;

and third, BPgroup cannot be modelled as a clade with contempo-

rary or earlier groups (P < 0.001). Whatever BPgroup’s origins are,

we can use it as one proximate source for the Volga cline together

with an EHG source from Karelia2,20, which is well outside the Volga

area and is thus unlikely to be part of the riverine mating network.

Seven Volga cline populations fit this model (P-values of 0.04 for

Ekaterinovka and 0.12–0.72 for the others) with consistently poor fits

only for upper Volga, Murzikha, Maximovka and Klo (the Khvalynsk

individuals with low Berezhnovka relatedness) (P-values from 1 × 10

−66

to 0.006). Three of these (other than Klo) are arrayed in the upriver EHG

cline (Fig.1c).

People buried at Ekaterinovka (5050–4450 BC, based on three

herbivore bone radiocarbon dates unaffected by marine reservoir

effects; Supplementary Table1) were already mixing with lower Volga

Berezhnovka-related people (24.3 ± 1.3%). This contrasts with the earlier

hunter-gatherers from Lebyazhinka (7.9 ± 3.6%; consistent with zero,

P = 0.21). A century or two later at Khvalynsk

, around 120 km from

Ekaterinovka (4500–4350 BC, based on two herbivore bones), there is an

admixture gradient, divided for convenience into: Khvalynsk high (Khi;

76.8 ± 1.9% BPgroup), Khvalynsk medium (Kmed; 57.3 ± 1.7% BPgroup)

and Khvalynsk low (Klo; 41.2 ± 1.6% BPgroup). Volga cline individuals

had around 14–89% Berezhnovka ancestry (Fig.1c), dominated by nei-

ther the old native EHG group nor the lower Volga newcomers. Genetic

differentiation between lower Volga (BPgroup) and Ekaterinovka was

Nature | www.nature.com | 3

strong (F

= 0.030 ± 0.001; Extended Data Table1), probably reflecting

different linguistic–cultural communities.

A genetically Volga cline individual from Csongrád-Kettőshalom in

Hungary (4331–4073 BC) had 87.9 ± 3.5% BPgroup ancestry (Fig.1c),

similar to Khi individuals. This individual was from late fifth millen-

nium BC steppe-like graves in southeastern Europe that included a

cemetery at Mayaky in Ukraine17,25,26 and a cemetery at Giurgiuleşti27

in Moldova, from which one individual (I20072; 4330–4058 BC) is a

clade with BPgroup (P = 0.90). Archaeology has documented Balkan

copper on the Volga cline site of Khvalynsk

, and the Csongrád and

Giurgiuleşti individuals were plausibly part of this cultural exchange,

leapfrogging the intervening Dnipro and Don basins without picking

up ancestry from them17.

The Dnipro cline

The Dnipro cline is formed by Neolithic individuals who lived along

the Dnipro River rapids (UNHG; 6242-4542 BC) and the Serednii Stih

population, represented by 13 individuals (4996–3372 BC; uncorrected

for freshwater-reservoir effects). This cline also includes most later

Yamnaya individuals, a high-quality and genetically homogeneous

subset (n = 104) that we term Core Yamnaya (Supplementary Informa-

tion section2). Close to Core Yamnaya (Fig.1b) are some Eneolithic

individuals: the Serednii Stih individual from Krivyansky in the lower

Don (4359–4251 BC) and the PVgroup from the north Caucasus. None-

theless, the Core Yamnaya cannot be modelled as derived from them

or any other single source (P < 1 × 10

−4

). Dnipro cline people are also

BPgroup

ChekalinoIV

CHG

CoreYamnaya

Csongrád_I5124

Ekaterinovka

Ekaterinovka_o

GK1

GK2

Hungary_Yamnaya

Igren_o

Kazakhstan_Kumsay_EBA

Khi

KhlopkovBugor

Klo

Kmed

Krivyansky

Labazy

Lebyazhinka_Eneolithic

Lebyazhinka_HG

Maikop

Maikop_I4429

Maximovka

Murzikha

PVgroup

Remontnoye

Russia_Don_EBA_Yamnaya

Russia_Sidelkino.SG

Russia_Steppe_Maikop

SE_Europe_Yamnaya

Sharakhalsun_SA6010

Sharakhalsun_SA6013

SShi

SSlo

SSmed

Syezzheye

Trypillia

Ukraine_EBA_Deriivka_I4110

Ukraine_EBA_Deriivka_I5882

Ukraine_EBA_Deriivka_I5884

Ukraine_EBA_Ozera_I1917

Ukraine_N

Ukraine_N_Deriivka_I3719

Unakozovskaya

UpperVolga

Usatove

Armenia_Aknashen_N

Armenia_C

Armenia_MasisBlur_N

Azerbaijan_C

Azerbaijan_N

Iran_C_SehGabi

Iran_C_TepeHissar

Iran_GanjDareh_N

Iran_HajjiFiruz_C

TUR_Aegean_BA

TUR_BlackSea_BA

TUR_BlackSea_ChL

TUR_C_BA

TUR_C_ChL

TUR_E_BA

TUR_E_ChL

TUR_Hatay_BA

TUR_Hatay_ChL

TUR_Marmara_Barcın_N

TUR_Marmara_ChL

TUR_Med_BA

TUR_SE_BA

TUR_SE_Çayönü_PPN

TUR_SE_ChL

Csongrád

GanjDareh

TepeHissar

SehGabi

Aknashen

Areni-1

MasisBlur

Sakhtysh

Murzikha

Kumsay

Verteba

Soldaneshti

Kopachiv

Klady

Unakozovskaya

Krivyansky

Remontnoye

Sharakhalsun

Berezhnovka

GolubayaKrinitsa

Khvalynsk

Ekaterinovka Syezzheye

KhlopkovBugor

Labazy

Lebyazhinka

Igren

Oleksandria

Deriivka

Vinogradnoe

Ural Yam.

Volgograd Yam.

Stav. Krai Yam.

Don Yam.

Chelyabinsk Yam.

Samara Yam.

CHG

Progress-2

Vonyuchka

Sidelkino

Barcin Ilipinar

Yassitepe

Harmanören-Göndürle

Kalehöyük

Ovaören

Ikiztepe

Çorum

Devret Höyük

Tell Kurdu

Arslantepe

Oylum Höyük

Tatika

Mayaki

SE Europe Yam.

Hungary Yam.

Giurgiule ti

–0.01 0 0.01 0.02 0.03

–0.02

–0.01

0.01

0.02

PC1

PC2

Armenia_Aknashen_N

Armenia_C

Armenia_MasisBlur_N

Azerbaijan_C

Azerbaijan_N

BPgroup

ChekalinoIV

CHG

CoreYamnaya

Csongrád_I5124

Ekaterinovka

Ekaterinovka_o

Giurgiule ti

GK1

GK2

Igren_o

Iran_GanjDareh_N

Kazakhstan_Kumsay_EBA

Khi

KhlopkovBugor

Klo

Kmed

Krivyansky

Labazy

Lebyazhinka_Eneolithic

Lebyazhinka_HG

Maikop

Maikop_I4429

Maximovka

Murzikha

PVgroup

Remontnoye

Russia_AfontovaGora3

Russia_Don_EBA_Yamnaya

Russia_Karelia

Russia_Sidelkino.SG

Russia_Steppe_Maikop

Sharakhalsun_SA6010

Sharakhalsun_SA6013

SShi

SSlo

SSmed

Syezzheye

Trypillia

TTK

TUR_Marmara_Barcın_N

TUR_SE_Çayönü_PPN

Ukraine_EBA_Deriivka_I4110

Ukraine_EBA_Deriivka_I5882

Ukraine_EBA_Deriivka_I5884

Ukraine_EBA_Ozera_I1917

Ukraine_N

Ukraine_N_Deriivka_I3719

Unakozovskaya

UpperVolga

Usatove

WSHG

Siberia

Central Asia

Lower

Volga

UNHG EHG

Mesopotamia

South

Caucasus

North

Caucasus

Iran

CHG

Caucasus–lower Volga cline

European hunter-gatherer cline

Upper Volga

Kama

Golubaya Krinitsa

European–Anatolian farmers

Dnipro cline

Volga cline

Russia_Karelia (n = 18)

Ekaterinovka:I23649

Ekaterinovka:I20192

Ekaterinovka:I23650

Labazy:I6910

Ekaterinovka:I8287

Ekaterinovka:I8284

Ekaterinovka:I20115

Ekaterinovka:I6069

Syezzheye:I22205

Syezzheye:I22203

Ekaterinovka:I8286

Ekaterinovka:I23648

Ekaterinovka:I6101

Ekaterinovka:I23651

Ekaterinovka:I6068

Ekaterinovka:I23652

Labazy:I6916

Ekaterinovka:I6064

Ekaterinovka:I3546

Ekaterinovka:I6061

Ekaterinovka:I6062

Klo:I0433

Ekaterinovka:I20114

Klo:I6110

Ekaterinovka:I8290

Klo:I6108

Ekaterinovka:I8740

Ekaterinovka:I8283

Syezzheye:I22204

Klo:I0122

Klo:I6740

Kmed:I0122_d

Klo:I6109

Klo:I6102

Kmed:I6738

Kmed:I6736

Kmed:I0426

Kmed:I6103

Kmed:I6735

Kmed:I6406

Kmed:I6405

Kmed:I6107

KhlopkovBugor:I6301_enhanced

Khi:I11837

Khi:I6412

Khi:I6741

Khi:I6737

Khi:I6734

Khi:I0434

KhlopkovBugor:I6300_enhanced

Khi:I6739

Csongrád:I5124

Khi:I6104

BPgroup (n = 5)

Russia_Karelia (n = 18)

Ukraine_N (n = 35)

GK2:I12490

SSlo:I1424

GK1:I12491

SSmed:I27282

SSmed:I6558

DonYamnaya:I12686

GK1:I12492

SSmed:I27283

SSmed:I28319

DonYamnaya:I8951

DonYamnaya:I11029

SSmed:I5894

SSmed:I7585_enhanced

DonYamnaya:I26638

GK1:I12493

DonYamnaya:I24088

SSmed:I4118

DonYamnaya:I12687

DonYamnaya:I10627

DonYamnaya:I26783

DonYamnaya:I26785

DonYamnaya:I24089

DonYamnaya:I12685

SShi:I1924

DonYamnaya:I6882_d

DonYamnaya:I8952

SShi:I2108

DonYamnaya:I26780

DonYamnaya:I24093

DonYamnaya:I24091

DonYamnaya:I24086

SShi:I6559

SShi:I1430

CoreYamnaya (n = 104)

BPgroup: (n = 5)

PVgroup:PG2001

PVgroup:VJ1001

Sharakhalsun:SA6010

Remontnoye:I28683

Remontnoye:I28682

LateMaikop:MK5004

Maikop:I6266

Maikop:I6267

Maikop:I1720_wNonUDG

Maikop:OSS001

LateMaikop:MK5008

Maikop:I6268

Maikop:I6272

Aknashen:I3931

Giurgiule ti

CLV clineDnipro clineVolga cline

Giurgiule ti:I20072

Fig. 1 | Three Eneoli thic clines and their nei ghbours in space and tim e.

a, Map with analysed site s. b, PCA using axes formed by a set of ancient West

Europ ean hunte r-gatherer and Siber ian, West As ian and Euro pean farm er

popul ations . Select ed indivi duals rele vant to this study are pro jected

(Method s). c, qpAdm mod els fit ted on indi viduals of the popul ations of

the cli nes. The Volga cline is generate d by admixt ure betwe en lower Volg a

(BPgroup) people wi th upriver EHG populations . People of the Dnipro cline

have UNHG or UNHG + EHG admixture relative to the Core Yamnaya (the hunter-

gatherer sour ce along this cline is signific antly variable). The Caucasus–lowe r

Volga cline is generated by admix ture of lower Volga people with those from

the Neolithic Cau casus (Aknashe n related). The map was drawn using public-

domain Natural Ea rth data with the rnatural earth package in R54.

4 | Nature | www.nature.com

Article

distinct from Volga cline individuals because no inter-riverine pairs

form a clade (P < 1 × 10

−7

). This distinctiveness spans three millennia,

commencing with the UNHG, continuing with the Eneolithic Serednii

Stih, and ending with the Early Bronze Age Yamnaya. A geographically

localized Yamnaya population of the lower Don (n = 23), many (n = 17)

from the site of Krivyansky, is distinct from the Eneolithic individual

at Krivyansky (Fig.1b) and not a clade with them (P = 8 × 10−15). The

Yamnaya can thus not be traced to the north Caucasus (PVgroup), the

lower Don (Krivyansky) or the Volga (BPgroup and the rest of the Volga

cline). Their placement on the Dnipro cline indicates their formation

by a process of admixture as descendants of the Serednii Stih culture.

Serednii Stih heterogeneity contrasts with Core Yamnaya homogene-

ity (Fig.1b), which is remarkable given the 5,000-km-wide sampling of

the latter, from Hungary to southern Siberia. The Yamnaya expanded

across this vast region, hardly admixing with locals, at least initially and

for the elite individuals buried in kurgans. Individuals of the Serednii

Stih culture are arrayed along the Dnipro cline. An individual from

Vinogradnoe, grouped with two from Oleksandria and one from Igren,

fall into an SShi cluster of greatest Core Yamnaya affinity but are not

a clade with them (P = 2 × 10

−7

). A Kopachiv female (I7585)

is part of

an SSmed cluster further along the cline, which also includes three

individuals from Oleksandria and three from Deriivka. SShi and SSmed

are largely contiguous, but I1424 from Moliukhiv Bugor (SSlo) is apart

from them, close to UNHG. Variation within the Serednii Stih plausibly

included unsampled individuals in gaps along the cline, or beyond its

sampled variation. The Don Yamnaya largely overlap with the Serednii

Stih, and at stratified sites of the lower Don Konstantinovka culture,

they continued to occupy Serednii Stih settlements, a continuity unob-

served in the Volga–Ural steppes.

All Dnipro cline groups can be well modelled with either UNHG or

GK2 (individual I12490 from Gol uba ya K rin its a in the mid dle Don ; 56 10-

5390 BC) at one extreme, and Core Yamnaya on the other (P-values

0.07–0.85). However, the hunter-gatherer end of the cline is not clearly

one or the other; although the source for SSmed upriver fits just as

well as UNHG (P = 0.27) or GK2 (P = 0.43), the Don Yamnaya upriver

source can fit only as UNHG (P = 0.08), not GK2 (P = 0.0001), and the

SShi upriver source can fit only as GK2 (P = 0.08), not UNHG (P = 0.003).

We t her efo re mod el in div id ual s f rom an y po in t al on g t he e nt ire UNH G–

EHG cline (Fig.1c), not presupposing either UNHG or GK2 as the source,

finding that UNHG ancestry predominates but more EHG ancestry is

also present (as at GK2). The hunter-gatherer source was thus from

the Dnipro–Don (UNHG–GK2), not the Volga (EHG). GK2 clusters with

Mesolithic hunter-gatherers from Vasylivka in the Dnipro17 and may

stand in for unsampled survivors there of that earlier population. Core

Yamnaya as a source for earlier populations would be ahistorical; it

must stand for an unsampled Eneolithic source.

The Don, which lies between the Dnipro and the Volga, is repre-

sented by middle Don Golubaya Krinitsa individuals and the lower

Don Krivyansky. Golubaya Krinitsa contained archaeologically con-

trasting graves, one similar to those of the Dnipro Neolithic and the

other similar to Serednii Stih

. GK2 is modelled as 66.6 ± 4.7% UNHG

and 33.4 ± 4.7% EHG (P = 0.39). Using the most ancient sources (Karelia,

UNHG and CHG), Krivyansky Eneolithic and Golubaya Krinitsa indi-

viduals have variable CHG-related ancestry (Fig.2a), maximized at

Krivyansky (58.9 ± 2.4%) and less (25.3 ± 2.1%) in three Golubaya Krinitsa

individuals grouped as GK1 (Fig.1); GK2 had none or little (4.0 ± 2.2%).

Thus, the admixture history of the Don paralleled its intermediate

geography, and included southern, CHG-related ancestry (Fig.2a).

This was already present in GK1 (individual I12491; 5557–5381 BC)

indicative of an early presence, but its absence in GK2 of a similar age

shows that it was not generally present. Dates for GK1 and GK2 may be

inflated because Golubaya Krinitsa was archaeologically interpreted as

being in cultural contact with the much later Eneolithic Serednii Stih

Moreover, a Serednii Stih outlier from Igren (I27930; 4337–4063 cal BC)

is a clade with GK2; this could be evidence of long-distance migration

from the Don to the Dnipro in a Serednii Stih time frame.

C dates at

Golubaya Krinitsa could potentially be overestimated owing to the

consumption of freshwater fish, which inflate dates by up to a millen-

nium in this region30.

It has been suggested11 that the Yamnaya had roughly 35% CHG-related

and about 65% Golubaya Krinitsa ancestry, the latter already having

around 20–30% CHG-related ancestry, implying that the main Yam-

naya source may have been hunter-gatherers of the Don area. Con-

tradicting this model, Yamnaya do not fit models with CHG-related

and either GK1 or GK2 sources

(P < 10

−6

). To better understand this,

we fitted Yamnaya to a model of Karelia + UNHG + CHG (Fig.2a) and

found that it underestimates the shared drift of Core Yamnaya with

both Afontova Gora-3 from Upper Palaeolithic Siberia (Z = −5.2) and

Anatolian Neolithic (Z = −6.8). A Volga source of the Siberian-related

ancestry is indicated by the fact that applying the same model to Volga

cline groups also underestimates shared drift with Afontova Gora-3

(P = 1 × 10−8 and Z = −4.5 for BPgroup; the Siberian ancestry is also evi-

dent in the deviation of the Dnipro cline towards Siberians in Fig.1b).

This Siberian-related ancestry is also affirmed because BPgroup can

be modelled as around 76% Krivyansky and 24% Central Asian (Sibe-

rian related) Tutkaul

(P = 0.13). When we fit Krivyansky and BPgroup

with the model that includes all relevant ancestries, CHG, GK2 and

Tutkaul (Fig.2b), Krivyansky has little to no Central Asian ancestry

(5.1 ± 3.6%), fitting as a simple two-way mix of 56.7 ± 2.6% CHG related

and 43.3 ± 2.6% GK2 (P = 0.37). By contrast, BPgroup requires 29.3 ± 2.2%

Tutkaul. Even adding Siberian-related ancestry (Tutkaul) is not suf-

ficient to model the Core Yamnaya, however, because the three-way

model in Fig.2b still fails (P = 10−9) to explain the shared drift with

Anatolian Neolithic (Z = −6.1).

Central Asian or Siberian ancestry was therefore already in the north

Caucasus steppe and Volga during the Neolithic, but with no evidence

of it further west on the Don. Adding a third, western (UNHG) or eastern

(Tutkaul), source (Fig.2c,d) to the two-source BPgroup + EHG model

for Volga cline individuals, they remain well modelled with these two

alone (Fig.2c). Some have more Tutkaul ancestry (Fig.2d). However,

deviations are minor (4.4 ± 2.6% Tutkaul ancestry for Khi). Crucially,

the Core Yamnaya fail all models of Fig.2a–d (P < 10

−8

), so they were

not formed from the CHG–EHG–UNHG–Tutkaul blend of these models.

The CLV cline

The Core Yamnaya, positioned on the opposite end of the Dnipro

cline to the UNHG and GK2 (Fig.1b), had ancestry from an unknown

source of lower or even no such ancestry. The only consistently fitting

(P = 0.67) two-way model for them involved 73.7 ± 3.4% of the SShi sub-

set of Serednii Stih and 26.3 ± 3.4% from a population represented by

two Eneolithic individuals from Sukhaya Termista I (I28682) and Ulan

IV (I28683) (4152–3637BC) near the village of Remontnoye, north of

the Manych Depression between the lower Don and the Caspian Sea.

Remontnoye is on neither the Volga nor the Dnipro cline and does not

form a clade (P < 10

−10

) with any other group. It had at least two sources:

a southern, Caucasus one, comprising either descendants of people

like those who lived in Neolithic Armenia at Aknashen10, or ancestors

of people of the Bronze Agenorth Caucasus Maikop

culture; and a

northern one, from a population like BPgroup. The southern compo-

nent can be modelled as having around half its ancestry from either

Aknashen (44.6 ± 2.7%; P = 0.66) or Maikop (48.1 ± 2.9%; P = 0.44).

We estimate −0.3 ± 2.9% UNHG or −0.5 ± 3.5% GK2 ancestry when

either is added as a third source to the Aknashen + BPgroup model, so

Remontnoye had no discernible UNHG/GK2-related ancestry as antici-

pated for the unknown source for the Yamnaya. Moreover, the main

Maikop cluster, including individuals buried in kurgans in Klady and

Dlinnaya-Polyana, had 86.2 ± 2.9% (P = 0.50) Aknashen ancestry. Thus,

there is a CLV cline: Aknashen–Maikop–Remontnoye–Berezhnovka.

These four, arrayed in order of decreasing Caucasus Neolithic com-

ponent, match their south-to-north location. North Caucasus people

Nature | www.nature.com | 5

at Progress-2 and Vonyuchka-1 bucked the latitudinal trend, having,

unlike their Maikop neighbours, little Caucasus Neolithic ancestry.

These violations document long-range connectivity across the CLV

area, and provide an important example of how genetics and geog-

raphy do not always match.

We wanted to know which group mediated the southern ancestry

of the CLV cline. It is not Aknashen, which is geographically remote

and much earlier (5985–5836 BC). It is not Maikop, which was geo-

graphically closer but later (3932–2934 BC). Unsampled Meshoko

and Svobodnoe settlements (4466–3810 BC)

are plausible for the

expansion of Aknashen-like ancestry northward and Berezhnovka-like

ancestry southward, because they exchanged exotic stone, copper

and stone mace heads with Volga cline sites. They are preceded in the

north Caucasus by the Eneolithic Unakozovskaya (ref. 9, 4607–4450

BC, and this study) and succeeded by the Maikop. The Unakozovskaya

population is not a good genetic source for Remontnoye, because

the model BPgroup + Unakozovskaya fails (P < 0.001) by overesti-

mating (Z = 3.8) CHG-related drift. Unakozovskaya is well modelled

as 95.3 ± 6.3% Maikop and 4.7 ± 6.3% CHG (P = 0.46); this group is

therefore Maikop-like, but distinct genetically (P = 2 × 10

−11

) (Fig.1b).

A recently published

individual from Nalchik (around 5000–4800 cal

BC) had more steppe affinity than the sampled Unakozovskaya, and

can be modelled (Supplementary Information section2) as a mix of

Unakozovskaya and steppe populations. Thus, in the Eneolithic north

Caucasus there were: Aknashen-related ancestry, representing the

Neolithic spread; CHG-related ancestry, indicated by the Maikop–

Unakozovskaya contrast; and northern lower Volga ancestry, constitut-

ing about one-seventh of the ancestry of the sampled Maikop.

Remontnoye, Berezhnovka and Maikop all used kurgan burial, which

was common at around 5000–3000 BC in diverse CLV cline people.

By contrast, a distinctive burial feature, with individuals posed on

the back with the knees raised and the floor of the burial pit covered

with red ochre, was shared by almostall steppe groups including the

Serednii Stihand Volga cline, while Remontnoye and Maikop burials

were contracted on one side. Some funeral customs united Maikop

with the steppes, but others separated them.

Aknashen

CHG

Tutkaul

Krivyansky

Golubaya

Krinitsa

GK1

Murzikha

UpperVolga UpperVolga

BPgroup Steppe

Maikop

Kumsay

GK2

Tutkaul

a f e c

BPgroup

Ukraine_N

Karelia

Karelia Çayönü

Caucasus–lower Volga cline

Ukraine–Eastern hunter-gatherer cline

Don–Volga–Central Asia cline

Remontnoye

Dnipro cline

Volga cline Labazy

Labazy

C. Anatolia

Masis Blur

Areni-1

Maikop

Serednii

Stih

Murzikha

Core

Yamnaya

Golubaya

Krinitsa

Khi Kmed Klo

Ekaterinovka

Mesopotamia–Caucasus cline

Caucasus–Don cline

Lebyazhinka

Syezzheye

Serednii Stih

Armenia_Aknashen_N

Armenia_C

Armenia_MasisBlur_N

Azerbaijan_C

Azerbaijan_N

BPgroup

Central_Anatolia_AssyrianColonyPeriod

Central_Anatolia_EBA_II

Central_Anatolia_OldHittitePeriod

CHG

CoreYamnaya

Csongrád_I5124

Ekaterinovka_o

GK1

GK2

Igren_o

Kazakhstan_Kumsay_EBA

Khi

KhlopkovBugor

Klo

Kmed

Krivyansky

Labazy

Lebyazhinka_HG

Maikop

Maximovka

Murzikha

PVgroup

Remontnoye

Russia_Don_EBA_Yamnaya

Russia_Karelia

Russia_Steppe_Maikop

Sharakhalsun_SA6010

SShi

SSlo

SSmed

Syezzheye

TTK

TUR_Aegean_BA

TUR_BlackSea_BA

TUR_BlackSea_ChL

TUR_C_ChL

TUR_E_BA

TUR_E_ChL

TUR_Hatay_BA

TUR_SE_BA

TUR_SE_Çayönü_PPN

TUR_SE_ChL

Ukraine_N

UpperVolga

Fig. 2 | The three cline s in the context of Eneolithic and Bron ze Age

admixture. Six three-source qpAdm mo dels elucidate a complex histor y

of admixture. a, Cauc asus and European hunter-gath erer admixtures in the

‘old steppe’: Kriv yansky on the lower Don received much mor e CHG-related

admixture tha n did upriver people of the middle Don at Golubaya Krini tsa. In

the middle and upper Volga and the Kama Rive r, populations had ne gligible

CHG-relate d influence. b, The Don–Volga differe nce. On the lower Volga and

North Cauc asus piedmont, the BPgro up received CHG-related anc estry, similar

to its western lower Don counterpart at Kriv yansky. But it also received ancestr y

from Central As ia, and this eastern influ ence was greater still in the Bronze Age

steppe Maikop. c, The Volga basin Ene olithic populations wi th regard to the

Don. Populati ons at Khvalynsk, Klopkov Bugor and Ekat erinovka form a Volga

cline betwe en the Berezhnvoka cluster on the lower Volga and the upriver EHG -

like population s of the middle Volga (Labaz y and Lebyazhinka). d, The Volga

basin Eneolit hic populations with rega rd to Central Asia. There is a slight

excess of Central As ian ancestry in the Khi subse t of Khvalynsk. e, The Dnipro

cline. The Core Yamnaya are on one end of a cline that also include s the Don

Yamnaya and Serednii Stih popul ations, formed by admixture of Nor th Pontic

hunter-gathere rs with those from the CLV cline of differential admix ture of

Neolithic Cau casus and BPgroup peopl e. The CLV cline includes diver se people

buried in kurgan s at Berezhnovka, Progres s-2, Remontnoye and theMaikop

sites ofKlady and Dlinnay a-Polyana, dating from around 500 0–3000 BC.

f, West Asia. CLV ancestry fir st appears in the Chalcolithic popu lation at Areni-1

in Armenia and is also pres ent in the Bronze Age at Maikop. Most of the ancestry

is from West Asian source s from the Mesopotamia– Caucasus (or Çayönü–

Masis Blur–Aknashen) cline . Chalcolithic and Bronze Age Anatoli ans lack CLV

ancestr y, but traces of it can be found in Bronze Age centr al Anatolians.

6 | Nature | www.nature.com

Article

The CLV cline reveals that the ancestors of Dnipro cline Serednii

Stih and Yamnaya were CLV cline people, similar to Remontnoye, who

had moved into the Dnipro–Don region and mixed with locals. The

actual sources for the Yamnaya may have differed from the sampled

Remont noye and SShi. The Dnipro clin e can be fit (Fig .2e) by a three-way

model in which a Dnipro or Don hunter-gatherer source mixed with

groups of mixed Aknashen and Berezhnovka ancestry. Either GK2 or

UNHG can fit as the northern riverine source, but we use GK2 in Fig.2e

because this model has a higher P-value (0.93) than the UNHG alterna-

tive (P = 0.04). The Yamnaya are inferred to have about one-fifth of

their ancestry from Dnipro/Don hunter-gatherers: either 22.5 ± 1.8%

GK2 or 17.7 ± 1.3% UNHG.

The CLV cline was the source from which Caucasus-derived ances

try flowed into the ancestors of the Yamnaya10. The Remontnoye +

SShi model predicts shared genetic drift with Neolithic Anatolians

well (Z = −0.8), unlike models lacking Anatolian Neolithic ancestry

(Fig.2a–d). Archaeology has established that trade in Balkan copper

during the late fifth millennium BC to north Caucasus farmer sites

(Svobodnoe) and the Volga (Khvalynsk) took place, and Neolithic pots

similar to those from Svobodnoe appeared in Dnipro–Don steppe

sites connected with the Seredni Stih culture (Novodanilovka). This

cultural exchange contextualizes the entry of BPgroup/Aknashen

mixed groups into the Dnipro–Don steppes.

CLV impact in Armenia and Anatolia

People of the CLV cline also went south (Fig.2f), explaining the steppe

ancestry found at Areni-1 in Chalcolithic Armenia from around 4000

BC13, w here lower Volga an cestry ( 26.9 ± 2.3% BPgroup) admixed

with a local Masis Blur-related Neolithic substratum (Supplemen-

tary Information section2). This contrasts with the north Caucasus

Maikop, where the substratum was Aknashen related. We can model

Masis Blur as 33.9 ± 8.6% Aknashen and 66.1 ± 8.6% Pre-Pottery Neo-

lithic of the Tigris Basin of Mesopotamia33 at Çayönü (P = 0.47),

part of a Neolithic Çayönü–Masis Blur–Aknashen cline. The popu-

lations of Armenia retained CHG differentially

: more (42.0 ± 3.8%)

in Aknashen than in Masis Blur (13.7 ± 4.0%). Some Anatolian Chal-

colithic and Bronze Age groups can be derived entirely from the

Caucasus–Mesopotamian cline (Fig.2f), w hereas others also have

ancestry from the Mesopotamian–Anatolian cline, lacking any steppe

ancestry10,15,34–36.

We s how th at C en tra l A nat oli an s

fro m th e Ea rly Bro nze Age ( 2750 –

2500 BC), Assyrian Colony (2000–1750 BC) and Old Hittite (1750–1500

BC) periods were unusual in the Anatolian landscape because they

had CLV ancestry combined with Mesopotamian (Çayönü) (Fig.2f,

Extended Data Fig.1 and Supplementary Information section 2).

The non-Mesopotamian ancestry varied, depending on the level of

CLV input: 10.8 ± 1.7% ancestry (P = 0.14) from BPgroup, 19.0 ± 2.4%

from Remontnoye (P = 0.19) or 33.5 ± 4.8% from Armenia_C (P = 0.10).

The exact source of the steppe ancestry in Anatolia cannot be pre-

cisely determined, but all fitting models involve some of it (Extended

Data Fig.1a). Some of the steppe-related sources are unlikely on

chronological or linguistic grounds; for example, the Core Yamnaya

(12.2 ± 2.0%; P = 0.10), as well as western Yamnaya-derived populations

from southeastern Europe, such as from Boyanovo or Mayaky Early

Bronze Age

(Extended Data Fig.1b). The Early Bronze Age Central

Anatolians from Ovaören

(275 0–25 00 BC) do temporally overlap the

late Yamnaya period, but the timing of the Yamnaya expansion is in

tension with the much-earlier linguistic split of Anatolian languages

that form an outgroup to those of the inner Indo-European Core37.

Fixing Çayönü as one source and adding pairs of steppe sources

(allowing ancestry to range freely along the Volga, Dnipro and CLV

clines), the hunter-gatherer contribution is negative on the Volga

cline (−3.4 ± 2.6% EHG) and on the Dnipro cline (−2.3 ± 2.7% UNHG

and −3.9 ± 3.5% GK2); thus, the admixing population had no more

EHG, UNHG or GK2 ancestry than did the BPgroup or Core Yamnaya

endpoints of these two clines (Supplementary Information sec-

tion2). Placing the admixing population on the CLV cline is successful

(P = 0.129), with a significant amount of BPgroup ancestry (8.8 ± 2.7%)

validating a CLV and north-of-the-Caucasus mountains Eneolithic ori

gin. Steppe + Mesopotamian models fit the Central Anatolian Bronze

Age but none of the Chalcolithic/Bronze Age Anatolian regional subsets

(P < 0.001; the BPgroup + Çayönü model is shown in Extended Data

Fig.1c): their success is not due to their general applicability. Moreover,

steppe ancestry in the Central Anatolian Bronze Age is observed across

individuals and periods (Extended Data Fig.1d), including Early Bronze

Age Ovaören south of the Kızılırmak river and Middle or Late Bronze Age

Kalehöyük just within the bend of the river

. This is consistent with an

Anatolian–Hattic linguistic boundary coinciding with the Kızılırmak,

a boundary breached before the conquest of Hattusa by the Hittites in

roughly 1730 BC

. Regardless of the (inherently unknowable) linguistic

identity of the sampled individuals, their unique blend of ancestries

demands an explanation.

Populations along the path to Central Anatolia can be modelled with

BPgroup ancestry and distinctive Caucasus–Mesopotamian substrata:

Aknashen related in the north Caucasus Maikop; Masis Blur related

in Chalcolithic Armenia; and Mesopotamian Neolithic in the Central

Anatolian Bronze Age (Extended Data Fig.1e,f). These admixtures had

begun by around 4300–4000 BC (the date range of the Armenia_C

population13) and we date them to 4382 ± 63 BC (Extended Data Fig.2e).

The Pre-Pottery Neolithic population of Çayönü was genetically halfway

between that of Mardin

, 200 km to the east, and the Central Anatolian

Potter y Neol ithic at Çatalhöy ük

along the Mesopotamian–Anatolian

cline. Chalcolithic/Bronze Age people from Southeastern and Central

Anatolia all stemmed from the same Çatalhöyük–Mardin continuum,

(Supplementary Information section2). If the proto-Anatolians came

from the east, their descendants may have been at the state of Armi,

the precise location of which is uncertain but whose Anatolian per-

sonal names are recorded by their neighbours in the kingdom of Ebla

in Syria5 in the 25th century BC, half a millennium before Anatolian

languages are attested, and just south of the proposed migratory path

(Extended Data Fig.1f). We therefore propose that people of the CLV

cline migrated southwards in around 4400 BC, a millennium before the

Yamnaya, admixed along the way, and finally reached Central Anatolia

from the east.

We find Y-chromosome evidence consistent with this reconstruc-

tion: there are sporadic instances of steppe-associated Y-chromosome

haplogroup R-V1636 in West Asia at Arslantepe

in eastern Anatolia and

inKalavan

in A rme ni a in the Ear ly B ronz e Ag e (a rou nd 33 00–2 500 BC)

among individuals without detectible steppe ancestryin the rest of

their genomes10,13. The R-V1636 individual (ART038) from Arslantepe

does not clearly have BPgroup ancestry (3.6 ± 3.1%), but ART027 from

the same site (3370–3100 BC) does (16.7 ± 3.5%; P = 0.171), preceding

the same mix in Early Bronze Age Central Anatolia by a few centu-

ries. R-V1636 in the Remontnoye male, both of those from Progress-2

(ref. 9), two of three from Berezhnovka and 11 individuals of the Volga

cline show it to be a prominent lineage of the pre-Yamnaya steppe, and

it also appeared as far away as northern Europe

39,40

. A single R-V1636

individual (SA6010; 2886–2671 BC) from Sharakhalsun

, consistent

with CLV ancestry (Fig.2), is found post-Yamnaya, a last hold-out of

this once pervasive lineage (Fig.3).

The Yamnaya expansion

We infer the average date of mixture in Core Yamnaya41 to be 4038 ±

48 BC (Extended Data Fig.2a), with sources related to UNHG/EHG

hunter-gatherers and West Asian/Caucasus-related people (Fig.1b).

Such a date does not preclude the possibility that the mixture began

earlier or continued afterwards, but it corresponds strikingly to the

burgeoning of the Serednii Stih culture. The ancestors of the Core Yam-

naya (Fig.1b and Extended Data Table2) must have been geographically

Nature | www.nature.com | 7

constrained

, contrasting with their later distribution from China to

Hungary (Extended Data Fig.3a, Extended Data Table2 and Supple-

mentary Table6), even while maintaining high genetic similarity (mean

FST = 0.005) (Extended Data Table3). The Don Yamnaya (Extended

Data Fig.3a) are modelled as 79.4 ± 1.1% Core Yamnaya and 20.6 ± 1.1%

UNHG. The non-Yamnaya component may be underestimated, if, as is

plausible, the Core Yamnaya admixed with a Serednii Stih population

of partial UNHG ancestry. We estimate that the Don Yamnaya formed

in the late fourth millennium BC (Extended Data Fig.2b), when, one

may assume, unmixed UNHG were rare.

The western expansion also brought Yamnaya into southeastern

Europe, reaching as far as Albania and Bulgaria3,10. Many of these clus-

ter with the Core Yamnaya, but others deviate towards Neolithic and

Chalcolithic populations of southeastern and central Europe (Extended

Data Fig.3b). Yamnaya admixture with these (Extended Data Table4)

occurred in the late fourth millennium BC (Extended Data Fig.2c), after

sporadic early Chalcolithic migrations into southeastern Europe from

the steppe3,25. By contrast, the Don Yamnaya expanded little, because

almost no individuals with high-quality data outside the Don are a clade

with them (Supplementary Information section2); the lower Don was

a cul-de-sac for the Yamnaya expansion.

Y-c hro mos om e h apl og rou p sh ar in g is no t in fo rm ati ve f or Cor e Ya m-

naya origins but shows that the Don Yamnaya, dominated by haplo-

group I-L699 (17 of 20 instances), had continuity with their Serednii

Stih and Neolithic hunter-gatherer ancestors (Fig.3 and Supplementary

Table7). The Core Yamnaya had R-M269 (49 of 51 instances), most of

which was the R-Z2103 (41 of 51) sublineage, which was undetected

before the Yamnaya period and related to R-L51, prevalent among Bell

Beaker burials7 and non-steppe Europe (Fig.3). Slightly more distant

is R-PF7563, found in Mycenaean Greece

. R-L23, formed at around

4450 BC (https://www.yfull.com/tree/R-L23/; v.12.04.00), unifies in

the Eneolithic Beakers, Yamnaya and Mycenaeans. Population diver

gences are lower than haplogroup ones, so these lineages may have

coexisted within the Yamnaya. Finding the R-L23 founder population

remains ch allenging, but our failu re to sample it thus far is not surp ris-

ing if it was small and isolated.

That the Core Yamnaya are part of the Dnipro cline may indicate an

origin in the Dnipro basin itself. However, the Dnipro cline is generated

by admixture with Dnipro–Don people (UNHG/GK2 related), and the

Yamnaya on the Don are also part of this cline, so an alternative origin

in the Don area cannot be excluded. Solutions further east are unlikely

because the Yamnaya are on neither the Volga nor the CLV cline. The

situation is similar for solutions west of the Dnipro: the Core Yamnaya

have little or no European farmer ancestry (from the west)17 (Fig.1b).

A more western origin of the Core Yamnaya would also bring their

latest ancestors in proximity with the likely founders of the Corded Ware

complex, whose origin is itself in question but who must have been in

the area of central eastern Europe occupied by the Globular Amphora

culture west of the Core Yamnaya. Most Corded Ware individuals, who

can be fit as tracing a large part of their ancestry to the Yamnaya

2,12

were formed by admixture concurrent with the Yamnaya expansion41

(Extended Data Fig.2d), shared identical-by-descent (IBD) segments

demonstrating genealogical timeframe connections43, and had a bal-

ance of ancestral components for their non-European farmer-related

ancestry that was indistinguishable from the Yamnaya

. The early-third

millennium BC history of the Corded Ware population is intertwined

with the Yamnaya expansion because it involved admixture with geneti-

cally, if not necessarily archaeologically, Yamnaya people. The Dnipro–

Don area of the Serednii Stih culture fits the genetic data, because

it explains the ancestry of the nascent Core Yamnaya. All ancestral

components found in the Serednii Stih and lacking elsewhere are found

in the Yamnaya (Extended Data Fig.4), and from the Dnipro–Don area,

both Corded Ware and southeastern European Yamnaya in the west,

and the Don Yamnaya in the east, could have emerged by admixture

of the Core Yamnaya with European farmers and UNHG descendants,

respectively.

Date BC

6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000

R-M417

R-L51

R-Z2103

Q-L56

J-M410

I-L699

G-P15

R-V1636

Dnipro–Don

Don Early Bronze Age

Early and Middle Bronze Age

Middle and Late Bronze Age

Fatyanovo

Volga

North Caucasus

Armenia

Turkey

Iran

Other European

Fig. 3 | Patriline al succession . Temporal distr ibution of key Y-chromos ome

haplogrou ps from Kazakhsta n, Kyrgyzs tan, Mongolia, Russia, Turkmenis tan,

Ukra ine, Uzbek istan an d compara tive regi ons of Europ e and West As ia,

6000–1000 BC. The Earl y and Middle Bronze Age group includes the Yamnaya,

Afanasievo, Polt avka, Catacomb, Chemurch ek and North Caucasus cultu res;

the Middle and Late Bronz e Age group includes individual s of diverse culture s

down to 1000 BC, including those of the Sint ashta, Andronovo, Potap ovka and

Srubnaya culture s. Information on which indiv iduals are plotted can be found

in Supplement ary Information (Supple mentary Table6).

8 | Nature | www.nature.com

Article

We estimated the population growth of Core Yamnaya using

HapNe-LD, which infers effective population-size fluctuations in

low-coverage ancient DNA data

. Core Yamnaya dating to the first

300 (n = 25) and later 300 (n = 26) years of our sampling produce 95%

confidence intervals of 3829–3374 BC and 3642–3145 BC for the time

before growth (Fig.4). For both, these correspond to growth from an

effective number of reproducing individuals of a few thousand. These

intervals overlap at 3642–3374 BC, which is the late Serednii Stih period.

Taken together with the admixture dating, a scenario emerges in which

Yam nay a an ce sto rs wer e fo rm ed by a dmi xt ure at aro und 40 00 BC, and

half a millennium later, a subgroup of them developed or adopted

cultural innovations, expanded dramatically and manifested archaeo-

logically around 3300 BC.

IBD43 genomic segments of at least 20 cM between pairs of individuals

did exist before the Yamnaya between regional populations (Fig.5a),

but they became much more common in the Yamnaya period (Fig.5b).

Segments shared across more than 500 km were extremely rare before

the Yamnaya (Fig.5c), but were a few percent between 500 and 5,000 km

(Fig. 5d) in the Yamnaya period. Close genetic relatives, sharing at least

three segments of at least 20 cM (about fifth-degree relatives)43 or a sum

of IBD of 100 cM or more, were found within 500 km in both periods,

and at a much higher rate within each cemetery (Fig.5e,f). Around

14.4% of Yamnaya–Afanasievo individual pairs within kurgans were

close relatives, and 7.4% of them across kurgans of the same cemetery,

which is much lower than the 29.0% in the tightly connected pedigree of

the Hazleton North chambered tomb in Neolithic Britain from around

3700 BC45 (P = 0.00075; Fisher’s exact test). Kurgans were therefore

not family tombs

of biological relatives; indeed, biological kinship

in them was mostly due to common descent centuries in the past, and

close kinship links within kurgans were largely non-biological.

The origin of Indo-Anatolian languages

The conventional view defines Indo-European as including Anatolian

languages as the first split

47,48

. Here we use a newer terminology that

denotes the entire group as Indo-Anatolian and restricts Indo-European

to the related non-Anatolian language families, including Tocharian,

Greek and Sanskrit

4,10

. The split of Indo-Anatolian is linguistically dated

to 4300–3500 BC

4,37,48,49

, pre-dating both the attestation of the Hit-

tite language in Central Anatolia (post-2000 BC

) and the Yamnaya

expansion. We identify the Yamnaya as the proto-Indo-Europeans for

several reasons: first,theformation of the Yamnaya around 4000 BC

and their expansionfrom the end of the fourth millennium BC corre-

sponds to the Indo-European–Anatolian split. Second, the Afanasievo

migration12, plausibly carrying languages ancestral to Tocharian,

iswidely recognized as the second, post-Anatolian, split

. The Yam-

nayacontributed,after 2500 BC,to Armenians and, since the Early

Bronze Age (Extended Data Table 2c),to the Balkans

3,10

, where Greek

and lesser-known BalkanIndo-European languages such as Illyrian

and Thracian were spoken

10,35,42

. For the remaining Indo-European lan-

guages, transmission was indirect via descendantcultures of mixed

Yamnaya–European farmer origins expanding well beyond the steppes.

It is from them thatthe vast majority of present-day Indo-Europeans

are descended. These include non-Balkan European (Italic, Celtic,

Germanic, Baltic, and Slavi c) speakers through the geographically

complementary Corded Ware2,12 and Beaker cultures of the third mil-

lennium BC

. The Indo-Iranians, the largest surviving Indo-European

group of Asia, were ultimately descended from the Corded Ware too,

viaa long chain of eastward migrations to Fatyanovo

and Si nta sht a

8,34

Yamnaya and Anatolians share CLV ancestry (Fig.2e,f), which

must stem from proto-Indo-Anatolian language speakers, except

for the possibility of an early transfer of language without admixture.

That the CLV ancestry in Central Anatolians during the Hittite pres-

ence included lower Volga-related ancestry implies an origin north

of the Caucasus (Fig.2f and Extended Data Fig.1). Long (30 cM or

longer) IBD segments shared by Igren-8 Serednii Stih and Areni-1 with

Berezhnovka-2 document Eneolithic links of lower Volga ancestry

(Extended Data Table5), and one link (15.2 cM) between the north

Caucasus Vonyucka-1 with early Bronze Age Ovaören (MA2213) ties

Central Anatolia to this once expansive network. Even so, only two

Indo-Anatolian descendant groups transmitted their languages to

posterity: the Yamnaya, aided by their horse-wagon technology

, and

Anatolian speakers, surviving long enough for their languages to be

committed to clay around 2000 BC

, vanishing in late antiquity and

fortuitously decyphered in the twentieth century. Our reconstruction,

based on genetics (Extended Data Fig.5), has traced both groups to

the CLV people north of the Caucasus, but it cannot discern who first

spoke pre-Indo-Anatolian languages.

Linguistic evidence has been advanced in favour of different solu-

tions to the problem of proto-Indo-European origins for more than

two centu ries, and here we review some rece nt proposals tha t are

relevant to our reconstruction of early Indo-Anatolian/Indo-European

history.

First, cereal terminology in Indo-Anatolian/Indo-European lan-

guages may restrict Indo-Anatolian origins to the easternmost extent

of agricultural subsistence during the Eneolithic, the Dnipro valley52.

Our findings do not contradict this, but they raise the possibility of a

3,0004,0005,000

Date BC

1,000

2,000

5,000

10,000

20,000

50,000

100,000

200,000 3150–2850 BC

2850–2550 BC

Fig. 4 | Trajectory of the Yamnaya expansio n. We used HapNe-LD to estimate

the changes in effec tive population size (Ne) over time of Yamnaya ancestors .

We carried out the compu tation separately for the indiv iduals from the earlier

300 years of our sampling and the lat er 300 years; shading shows confid ence

interval s (dark, 50%; light, 95%). Jointly disp laying these two trajec tories

reveals an extraordin ary population expans ion at 3642–3374 BC (intersection

of 95% confiden ce intervals for the two analyses for the minimu m), from when

the effect ive size is a few thousand to an order of magnitu de larger. The offset

on the x axis is due to the difference in samplin g time between the two groups .

Nature | www.nature.com | 9

Caucasus (rather than European) Neolithic source for this vocabulary

through the CLV cline.

Second, the attestation of Anatolian languages largely in central-

western Anatolia can be explained most parsimoniously by a western

entry (through the Balkans)

, but genetic data provide strong evidence

in favour of an eastern route

, because not only CLV but especially

Mesopotamian Neolithic, the two sources of the Central Anatolian

Bronze Age groups, are eastern. Further evidence comes from observ-

ing no European farmer or hunter-gatherer ancestry in Central Ana-

tolian Bronze Age groups, as might be expected from a Balkan route

from the west, although if these groups by-passed local Europeans, or

used a maritime route, we would not see European mixture. A weak-

ness of the eastern-entry hypothesis has always been that there is no

linguistic evidence of Anatolian speakers in eastern Anatolia along the

proposed migratory path. However, this argument does not add relative

weight to the western-entry hypothesis either, because no linguistic

evidence for migratory pre-Anatolian speakers has been found in the

southeastern European path proposed by that hypothesis. The lack of

linguistic traces in eastern Anatolia could be explained by the archaeo-

logically momentous expansion of the Kura-Araxes archaeological

a b

Pre-Yamnaya Yamnaya

At least one IBD segment ≥ 20 cM

Distance between individuals (km)

Fraction of individuals

sharing IBD

c d

0 1,000 2,000 3,000 4,000 5,000 6,000

Distance between individuals (km)

0 1,000 2,000 3,000 4,000 5,000 6,000

Distance between individuals (km)

0 1,000 2,000 3,000 4,000 5,000 6,000

Distance between individuals (km)

0 1,000 2,000 3,000 4,000 5,000 6,000

235

4 0 0

734

1,757

1,937 1,225

Pairs sharing IBD, IPairs sharing IBD, I

Total pairs, TTotal pairs, T

I/T I/T

Pairs sharing IBD, I

Total pairs, T

I/T

Pairs sharing IBD, I

Total pairs, T

I/T

At least one IBD segment ≥ 20 cM

257

120 54 21 46 53

34 17 18 5

283

2,880

2,374 2,023 972 1,204 1,113 1,681 571 737 358

80 0 0

734

1,757 1,937

Three IBD segments ≥ 20 cM or total IBD ≥ 100 cMThree IBD segments ≥ 20 cM or total IBD ≥ 100 cM

21 0 1 0000000

283

2,880 2,374 2,023 972 358

Fraction

Between cemeteries

(n = 1,106)

Between kurgans

(n = 95)

Within kurgans

(n = 125)

Hazleton North

(n = 435)

Neolithic Ireland

(n = 103)

(1) Three IBD segments ≥ 20 cM or total IBD ≥ 100 cM

(2) At least one IBD segment ≥ 20 cM but not in (1)

Neither (1) nor (2)

0.4

0.3

0.2

0.1

Fraction of individuals

sharing IBD

0.4

0.3

0.2

0.1

0.2

0.4

0.6

0.8

1.0

Fraction of individuals

sharing IBD

0.4

0.3

0.2

0.1

Fraction of individuals

sharing IBD

0.4

0.3

0.2

0.1

1,133

1,225 1,133 1,204 1,113 1,681 571 737

Fig. 5 | IBD analysis of the Yamnaya and their pre decessors . a,b, Pairs of

individual s linked by at least one IBD seg ment at least 20 cM in length reveal a

sparse but highly co nnected network in the pre-Yamnaya (Method s) (a) and

Yamnaya (b) groups. No detect able IBD is found in the pre-Yamnaya period

beyond the scale of 1,000 k m. c,d, Yamnaya share more IBD with eac h other

at short dista nce scalescompared wit hthe pre-Yamnaya people(c), but IBD

sharing exten ds all the way to the roughly 6,000 k m scale of their geographic al

distribution (d). e,f, However, closely related indiv iduals occur only at short

distanc e scales in both pre-Yamnaya (e) and Yamnaya (f) groups, indicating that

IBD sharing in the Yamnaya was a legacy of their comm on origin. In c–f, two-

sided 95% confi dence intervals are shown as a verti cal interval (at distance = 0)

or a rectangle (at dista nce ranges greater than 0); the fractio n of number of

pairs of individua ls sharing IBD (I)/total number of pairs of indiv iduals (T) is

shown in red. g, In a set of 9 Yamna ya cemeter ies and a tot al of 25 kurga ns,

close ly or distan tly relate d individu als are almos t absent in in ter-cemet ery

comparisons, more are found in inter-kurgan and within-cemetery comparis ons,

and even more are found in intra-kurgan comp arisons; nonethe less, most

Yamnaya individuals in all compar isons were unrelated. Kurg an burial of close

kin was less commo n than in the ca se of a local pa trilinea l dynast y, as at a

Neolithic lon g cairn at Neolithic Hazleton Nor th45, but was more common than

in Neolithic monu ments in Ireland55. Two-sided 95% confid ence intervals are

shown. The map was draw n using public-domain Natura l Earth data with the

rnaturale arth package in R54.

10 | Nature | www.nature.com

Article

culture in the Caucasus and eastern Anatolia after around 3000 BC,

which may have driven a wedge between steppe and West Asian speak-

ers of Indo-Anatolian languages, isolating them from each other and

perhaps explaining their survival in western Anatolia into recorded

history. That the expansion of the Kura-Araxes archaeological cul-

ture could have had a profound enough demographic impact to have

pushed out Anatolian speakers is directly attested by genetic evidence

showing that, in Armenia, the spread of the Kura-Araxes culture was

accompanied by the complete disappearance of CLV ancestry that had

appeared there in the Chalcolithic10,13 (Fig.2f).

The Kura-Araxes culture may not be the only reason for the Indo-

Anatolian split. Autosomal and Y-chromosome homogenization of the

Yamnaya ancestral population in the fourth millennium BC provides

another lens through which to understand its origins, with isolation

fostering linguistic divergence. This may have persisted after its expan-

sion: previous inhabitants largely disappear in the face of the Yamnaya

juggernaut, albeit with exceptions17. Perhaps mixing, which was avoided

by the kurgan elites, occurred between loca ls and Yamnaya not buried

in kurgans. The rise of the Yamnaya on the steppe at the expense of their

predecessors was followed by their demise after about 1,000 years,

displaced by descendants of people of the Corded Ware culture. Was

this the fall of the kurgan elites or of the population as a whole? The

steppe was dominated by many and diverse groups later still, such as

the Scythians and Sarmatian nomads of the Iron Age. These groups were

certainly diverse genetically, but their kurgans, found across the steppe,

attest to the persistence of at least some elements of culture that began

in the Caucasus–Volga area some 7,000 years ago before blooming, in

the Dnipro–Don area, into the Yamnaya culture that first united the

steppe and had an impact on most of Eurasia. For what symbolic pur-

pose the Yamnaya and their precursors erected these mounds we may

never fully know. If they aimed to preserve the memory of those buried

under them, they did achieve their goal, as the kurgans, dotting the

landscape of the Eurasian steppe, drew generations of archaeologists

and anthropologists to their study, enabling the genetic reconstruction

of their makers’ origins presented here.

Online content

Any methods, additional references, Nature Portfolio reporting summa-

ries, source data, extended data, supplementary information, acknowl-

edgements, peer review information; details of author contributions

and competing interests; and statements of data and code availability

are available at https://doi.org/10.1038/s41586-024-08531-5.

1. Jones, E. R. etal. Upper Palaeolithic genomes reveal deep roots of modern Eurasians. Nat.

Commun. 6, 8912 (2015).

2. Haak, W. etal. Massive migration from the steppe was a source for Indo-European

languages in Europe. Nature 522, 207–211 (2015).

3. Mathieson, I. etal. The genomic history of southeastern Europe. Nature 555, 197–203

(2018).

4. Kloekhorst, A. in The Indo-European Puzzle Revisited: Integrating Archaeology, Genetics,

and Linguistics (eds K risti ansen , K. eta l.) 42 –60 (C ambri dge Uni v. Press , 2023 ).

5. Kroonen, G., Barjamovic, G. & Peyrot, M. Linguistic supplement to Damgaard etal. 2018:

Early Indo-European languages, Anatolian, Tocharian and Indo-Iranian. Zenodo https://

doi.org/10.5281/zenodo.1240524 (2018) .

6. Anthony, D. W. The Horse, the Wheel, and Language: How Bronze-Age Riders from the

Eurasian Steppes Shaped the Modern World (Prin ceton Univ. Pr ess, 20 07).

7. Olalde, I. etal. The Beaker phenomenon and the genomic transformation of northwest

Europe. Nature 555, 190–196 (2018).

8. Narasimhan V. M. etal. The formation of human populations in South and Central Asia.

Science 365, eaat7487 (2019).

9. 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).

10. Lazaridis, I. etal. The genetic history of the Southern Arc: a bridge between West Asia and

Europe. Science 377, eabm4247 (2022).

11. Allentoft, M. E. etal. Population genomics of post-glacial western Eurasia. Nature 625,

301–311 (2024).

12. Allentoft, M. E. etal. Population genomics of Bronze Age Eurasia. Nature 522, 167–172

(2015).

13. Lazaridis, I. etal. Genomic insights into the origin of farming in the ancient Near East.

Nature 536, 419–424 (2016).

14. Lazaridis, I. etal. Ancient DNA from Mesopotamia suggests distinct Pre-Pottery and

Pottery Neol ithic migrations into An atolia. Science 377, 982–987 (2022).

15. Skourtanioti, E. etal. Genomic history of Neolithic to Bronze Age Anatolia, Northern

Levant and Southern Caucas us. Cell 181, 1158–1175 (2020).

16. Lazaridis, I. etal. Ancient human genomes suggest three ancestral populations for

present-day Europeans. Nature 513, 409–413 (2014).

17. Nikitin, A. G. etal. A genomic history of the North Pontic Region from the Neolithic to the

Bronze Age. https://doi.org/10.1038/s41586-024-08372-2 (202 5).

18. Reich, D. etal. Reconstruc ting Native American popu lation history. Nature 488, 370–374

(2012).

19. Zeng, T. C. etal. Postglacial genomes from foragers across Northern Eurasia reveal

prehistoric mobility associated with the spread of the Uralic and Yeniseian languages.

Nature (in p ress).

20. Posth, C. etal. Palaeogen omics of Upper Palaeolit hic to Neolithic Europea n hunter-

gatherers. Nature 615, 117–126 (2023).

21. Vybornov, A. etal. Diet and chronology of Neolithic–Eneolithic cultures (from 6500 to

4700 cal BC ) in the Lower Volga Basin . Radiocarbon 60, 1597–1610 (2018).

22. Gimbutas, M. The Prehistory of Eastern Europe (Peabod y Muse um, 195 6).

23. Fu, Q. etal. The gen etic history of Ice Age Europe. Nature 534, 200–205 (2016).

24. Anthony, D. W. etal. The Eneolithic cemetery at Khvalynsk on the Volga River. Praehist. Z.

97, 22–67 (2022).

25. Penske, S. etal. Early contact between late farming and pastoralist societies in southeastern

Europe. Nature 620, 358–365 (2023).

26. Nikitin, A. G., Ivanova, S., Culleton, B. J., Potekhina, I. & Reich, D. New radiocarbon and

stable isotope data from the Usatove culture site of Mayaky in Ukraine. SSRN https://

doi.org/10.2139/ssrn.4236123 (2023) .

27. Govedarica, B. & Manzura, I. The Giurgiulesti cemetery in chronological and cultural

context of Southeastern and Eastern Europe. Eurasia Antiqua 22, 1–39 (2016).

28. Skorobogatov, A. M. Pamyatniki Neolita I Eneolita v Ust’e Chernoi Kalitvy. Trudy

Voronezhskogo Oblastnogo Kraevedcheskogo Muzeiya Vyp. 3, 47–53 (2019).

29. Skorobogatov, A. M. & Smol’janinov, R. V. Srednestogovskie materialy v bassejne Verhnego

i Srednego Dona. Rossiyskaya ark heologiya 2013, 126–136 (2013).

30. Shishlina, N. I. etal. Paleoecology, subsistence, and 14C chronology of the Eurasian

Caspian steppe Bronze Age. Radiocarbon 51, 481–499 (2009).

31. Korenev skii, S. N. Rozhdenie Kurgana ( Taus, 20 12).

32. Zhur, K. V. etal. Human DNA from the oldest Eneolithic cemetery in Nalchik points the

spread of farming from the Caucasus to the Eastern European steppes. iScience https://

doi.org/10.1016/j.isci.2024.110963 (2024).

33. Altınışık, N. E. etal. A genomic snapshot of demographic and cultural dynamism in Upper

Mesopotamia during the Neolithic transition. Sci. Adv. https://doi.org/10.1126/sciadv.

abo3609 (202 2).

34. 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).

35. Lazaridis, I. etal. Genetic origins of the Minoans and Mycenaeans. Nature 548, 214–218

(2017).

36. Lazaridis, I. The evolutionary history of human populations in Europe. Curr. Opin. Genet.

Dev. 53, 21–27 (2018).

37. Kassian, A. S. etal. Rapid radiation of the inner Indo-European languages: an advanced

approach to Indo-European lexicostatistics. Linguistics 59, 949–979 (2021).

38. Yaka, R. etal. Variable kinship patterns in Neolithic Anatolia revealed by ancient genomes.

Curr. Biol. https://doi.org/10.1016/j.cub.2021.03.050 (2021 ).

39. Egfjord, A. F.-H. etal. Genomic steppe ancestry in skeletons from the Neolithic single

grave culture in Denmark. PLoS ONE 16, e0244872 (2021).

40. Mathieson, I. etal. Genome-wide patterns of selection in 230 ancient Eurasians. Nature

528, 499–503 (2015).

41. Chintalapati, M., Patterson, N. & Moorjani, P. The spatiotemporal patterns of major human

admixture events during the European Holocene. eLife 11, e77625 (2022).

42. A genetic probe into the ancient and medieval history of Southern Europe and West Asia.

Science 377, 940–951 (2022).

43. Ringbauer, H. etal. Accurate detection of identity-by-descent segments in human ancient

DNA. Nat. Genet. 56, 143–151 (2024).

44. Fournier, R., Tsangalidou, Z., Reich, D. & Palamara, P. F. Haplotype-based inference of

recent effecti ve population size in modern and ancient DNA samples. Nat. Commun. 14,

7945 (2023).

45. Fowler, C. etal. A high-resolution picture of kinship practices in an Early Neolithic tomb.

Nature 601, 584–587 (2022).

46. Shishlina, N. Reconstruction of the Bronze Age of the Caspian steppes: Life Styles and Life

Ways of Pasto ral Nomads (B ritish Archa eolog ical Re ports , 2008) .

47. Olander, T. Indo-European cladistic nomenclature. Indogermanische Forschungen 124,

231–244 (2019).

48. Anthony, D. W. & Ringe, D. The Indo-European homeland from linguistic and archaeological

perspectives. Annu. Rev. Linguist. 1, 199–219 (2015).

49. Pronk, T. C. Indo-European secondary products terminology and the dating of Proto-

Indo-Anatolian. J. Indo-E uropean S tud. 49, 141–170 (2022).

50. Ringe, D., Warnow, T. & Taylor, A. Indo-European and computational cladistics. Trans.

Philol. Soc. 100, 59–129 (2002).

51. Saag, L. etal. Genetic ancestry changes in Stone to Bronze Age transition in the East

European plain. Sci. Adv. 7, eabd6535 (2021).

52. Kroonen, G., Jakob, A., Palmér, A. I., van Sluis, P. & Wigman, A. Indo-European cereal

terminology suggests a Northwest Pontic homeland for the core Indo-European

languages. PLoS ONE 17, e0275744 (2022).

53. Kristiansen, K. in Dispersals and Diversification: Linguistic and Archaeological Perspectives

on the Early Stages of Indo-European (eds Se range li, M. & Oland er, T.) 157–16 5 (Bri ll, 20 19).

54. Massicotte, P. & South, A. rnaturalearth: World map data from Natural Earth. https://docs.

ropensci.org/rnaturalearth/ (2024 ).

55. Cassidy, L. M. etal. A dynastic elite in monumental Neolithic society. Nature 582, 384–388

(2020).

Nature | www.nature.com | 11

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in

published maps and institutional affiliations.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this

article under a publishing agreement with the author(s) or other rightsholder(s); author

self-archiving of the accepted manuscript version of this article is solely governed by the

terms of such publishing agreement and applicable law.

1Department of Human Evolutionary Biology, Harvard University, Cambridge, MA, USA.

2Department of Genetics, Harvard Medical School, Boston, MA, USA. 3Broad Institute of

Harvard and MIT, Cambridge, MA, USA. 4Hartwick College, Department of Anthropology,

Oneonta, NY, USA. 5Department of Biology and Ecology, Faculty of Science, University of

Ostrava, Ostrava, Czechia. 6Department of Statistics, University of Oxford, Oxford, UK.

7Department of Archaeogenetics, Max Planck Institute for Evolutionary Anthropology, Leipzig,

Germany. 8BIOMICs Research Group, Department of Zoology and Animal Cell Biology,

University of the Basque Country UPV/EHU, Vitoria-Gasteiz, Spain. 9Ikerbasque-Basque

Founda tion of Sc ience, Bi lbao, Spai n. 10Samara State University of Social Sciences and

Education, Samara, Russia. 11Center of Human Ecology, Institute of Ethnology and

Anthropology, Russian Academy of Science, Moscow, Russia. 12Department of Archaeology,

State History Museum, Moscow, Russia. 13”Gavrilă Simion” Eco-Museum Research Institute,

Tul cea , Ro man ia. 14Historical Ecological and Cultural Association Povolzje, Samara Regional

Public Organization, Samara, Russia. 15Azov History, Archaeology and Palaeontology

Museum-Reserve, Azov, Russia. 16Institute of Archaeology named after A.Kh Margulan,

Almaty, Kazakhstan. 17Department of Biological Anthropology, Institute of Biology, University

of Szeged, Szeged, Hungary. 18Research Institute and Museum of Anthropology, Moscow,

Russia. 19Institute of Archeology named after A. Kh. Khalikov Tatarstan Academy of Sciences,

Kazan, Russia. 20Orheiul Vechi Cultural-Natural Reserve, Institute of Bioarchaeological and

Ethnocultural Research, Chișinău, Republic of Moldova. 21Fr. I Rainer Institute of

Anthropology, University of Bucharest, Bucharest, Romania. 22Damjanich János Museum,

Szolnok, Hungary. 23Department of Archaeology, University of Szeged, Szeged, Hungary.

24Déri Museum, Debrecen, Hungary. 25Department of Regional Studies of Russia, National and

State-Confessional Relations, Altai State University, Barnaul, Russia. 26Department of

Anthropology, Hungarian Natural History Museum-Hungarian National Museum Public

Collection Centre, Budapest, Hungary. 27Research Institute GAUK RO “Don Heri tage”,

Rostov-on-Do n, Russia. 28Institute of Parasitology, Biology Centre of the Czech Academy of

Sciences, České Budějovice, Czechia. 29Prahova County Museum of History and Archaeology,

Ploiești, Romania. 30Department of Geography, Faculty of Humanities, University Valahia of

Târgoviște, Târgo vişt e, Rom ania . 31Department of Biological Anthropology, Institute of

Biology, Eötvös Loránd University, Budapest, Hungary. 32Department of Evolutionary

Anthropology, University of Vienna, Vienna, Austria. 33Human Evolution and Archaeological

Sciences, University of Vienna, Vienna, Austria. 34Department of Mountain and Highland

Archaeology, Institute of Archaeology and Ethnology, Polish Academy of Science, Kraków,

Poland. 35Slovak National Museum-Archaeological Museum, Bratislava, Slovak Republic.

36Peter the Great Mus eum of Anthropology and Eth nography, Department of Physica l

Anthropology, St. Petersburg, Russia. 37State Autonomous Cultural Institution of the Republic

of Khakassia “Khakassian National Museum of Local Lore named after L.R. Kyzlasova”, Abakan,

Russia. 38Institute of Archaeology, HUN-REN Research Centre for the Humanities, Budapest,

Hungary. 39Centre for Egyptological Studies of the Russian Academy of Sciences, Russian

Academy of Scien ces, Moscow, Russia. 40Department of Archaeology and History of the

Ancient World, Southern Federal University, Rostov-on-Don, Russia. 41Museum of Vojvodina,

Novi Sad, Serbia. 42Department of History of the Institute of Humanities, Ural Federal

University, Ekaterinburg, Russia. 43Institute of Plant and Animal Ecology, Urals Branch of the

Russian Academy of Sciences, Yekaterinbu rg, Russia. 44Institute of History, Archaeology and

Ethnography, Dagestan Branch of the Russian Academy of Science, Makhachkala, Dagestan,

Russia. 45Independent researcher, Philadelphia, PA, USA. 46Department of General History,

Historical and Literary Institute of the State University of Education, Ministry of Education

Moscow, Moscow, Russia. 47Centre for Applied Bioanthropology, Institute for Anthropological

Research, Zag reb, Croatia. 48Department of Archaeology and Heritage, Faculty of Humanities,

University of Primorska, Koper, Slovenia. 49Kalmyk Scientific Centre of the Russian Academy

of Sciences, Elista, Russia. 50National Agency for Archaeology, Chișinău, Republic of Moldova.

51National Research Tomsk State University, Tomsk, Russia. 52V.F. Voino-Yasenetsky

Krasnoyarsk State Medical University, Krasnoyarsk, Russia. 53Department of Archaeology,

Ethnography and Museology, Altai State University, Barnaul, Russia. 54Laboratory of Ancient

and Medieval Archaeology of Eurasia, Altai State University, Barnaul, Russia. 55Slovak National

Museum-Natural History Museum, Bratislava, Slovak Republic. 56Faculty of History, University

of Oxford, Oxford, UK. 57Center for Stone Age Archeology, Institute of History and Archaeology,

Ural Branch of the Russian Academy of Sciences, Ekaterinburg, Russia. 58Olga Necrasov

Centre for Anthropological Research, Romanian Academy, Iași Branch, Iași, Romania.

59Tyumen Scientific Center of the Siberian Branch of Russian Academy of Sciences, Institute

of Problems of Northern Development, Tyumen, Russia. 60Institute for the History of Material

Culture, Russian Academy of Sciences, St Petersburg, Russia. 61Institute of Ethnology and

Anthropology, Russian Academy of Sciences, Moscow, Russia. 62Howard Hughes Medical

Institute, Harvard Medical School, Boston, MA, USA. 63Museo delle Civiltà, Italian Ministry of

Culture, Rome, Italy. 64School of Archaeology, University College Dublin, Dublin, Ireland.

65Institute of Archaeogenomics, HUN-REN Research Centre for the Humanities, Budapest,

Hungary. 66Wellcom e Cent re for Hu man Ge netics , Univ ersity of Oxfo rd, Oxfo rd, UK. 67These

authors contributed equally: Iosif Lazaridis, Nick Patterson, David Anthony, Leonid Vyazov.

✉e-mail: lazaridis@fas.harvard.edu; nickp@broadinstitute.org; AnthonyD@hartwick.edu;

l.a.vyazov@gmail.com; ron.pinhasi@univie.ac.at; reich@genetics.med.harvard.edu

Article

Methods

Terminology for archaeological cultures and geographic

locations

For archaeological cultures and geographic locations that span more

than one modern country, we used the prevalent term in the archaeo-

logical and genetic literature; for example, Yamnaya is the common

term in Russia and most of Eastern Europe, instead of the Ukrainian

Yamna. For archaeological cultures and locations that are confined to

a single country, we generally used the local terminology; for example,

we referred to the archaeological cultures of Usatove, Trypillia and

Serednii Stih, and the river Dnipro, using the Ukrainian terms instead

of the corresponding Russian terms Usatovo, Tripolye, Sredni Stog

and Dniepr.

Previously published Eneolithic and Yamnaya/Afanasievo

individuals

We counted previously published Yamnaya/Afanasievo individuals

with genome-wide autosomal data (n = 75) from the archaeogenetic

literature2,3,8–10,12,33,56–62. We counted pre-Yamnaya Eneolithic individu-

als

3,9,11,17,20,40,51,63,64

with genome-wide data from the European steppe and

its environs (n = 82) by filtering individuals to the date range 5000–

3500 BC, the countries of Russia and Ukraine, and latitude west or equal

to 60° E and longitude south or equal to 60° N.

Sampling ancient individuals

The skeletal remains were all analysed with permission from local

authorities in each location from which they came. Every sample is

represented by stewards, such as archaeologists or museum curators,

who are either authors of this paper or are thanked in the Acknowl-

edgements. The remains were nearly all sampled in ancient-DNA clean

rooms, either at Harvard Medical School, the University of Vienna or

the Institute for Archaeogenomics in Budapest. We prioritized sam-

pling petrous bones if they were available and accessible, taking bone

powder from the cochlea by sandblasting and milling

, or drilling into

the cochlea directly after physical surface cleaning, or drilling through

the cranial base to minimize damage to intact skulls66. If we could not

sample from the cochlea, we sought to sample a tooth, prioritizing the

cementum layer after physical surface cleaning67. If neither a cochlea

nor a tooth was available, we sought to sample a dense cortical bone,

which we analysed by drilling and collecting powder after physical

surface cleaning. For some samples that could not leave the museum,

we sampled on site, either drilling directly into the cochlea, the tooth

root or bone after physical surface removal. We sometimes dislodged

auditory ossicles during sandblasting or drilling into the cochlea. When

this happened during the cleaning procedure, we generally stopped

the destructive sampling and collected the ossicle(s)68. As suggested

in the study

that recognized the high preservation of DNA in ossicles,

we cleaned the ossicle with 10% bleach and radiated it with UV light for

10 min before submerging it in extraction buffer without attempting

to produce powder.

Ancient-DNA data generation

The samples for which we report new data were processed between

2013 and 2023, and so were analysed using a changing set of pro-

tocols. Details and protocols used for each library can be found in

Supplementary Table2. At Harvard Medical School, where most of

the wet laboratory work was done, we initially carried out all DNA

extractions and Illumina library preparations manually, using small

batches of samples and silica columns for DNA clean-up69–71. Begin-

ning in 2018, we used automated liquid handlers (Agilent Bravo

Workstations) for both DNA extraction72 and library preparation

with magnetic beads (see the supplementary material in ref. 73 for

automated double-stranded library preparation, and ref. 74 for

automated single-stranded library preparation). We treated DNA

extracts with USER (NEB) during library preparation to cut DNA at

uracils; this treatment is inefficient at terminal uracils and leaves a

damage pattern expected for ancient DNA at the terminal bases that

can be filtered out for downstream analysis while allowing a library to

be authenticated as old. All libraries were either dual barcoded through

double-stranded ligation or dual indexed through indexing PCR at the

end of single-stranded library preparation to allow pooling before

sequencing.

Before 2015, we screened libraries for mitochondrial DNA (mtDNA)

before attempting to capture nuclear loci

. In the following two years,

we added an increasing number (between 10 and 4,000) of nuclear

single-nucleotide polymorphisms (SNPs) as targets for the screening

capture because mtDNA quality does not always correlate well with

nuclear DNA quality and quantity. We later increased the number of

targeted SNPs in our nuclear capture from about 390,000 (390k)

2,76

to about 1.24 million (1,240k)77 for libraries passing the mitochondrial

capture with nuclear spike-in. We later stopped using the screening

capture and added the mitochondrial probes to the 1240k probes

(1240k+). In 2022, we switched from the 1240k homebrew capture to

a kitted capture product available from Twist Biosciences78.

For ancient-DNA data generated at the Institute of Archaeogenomics

in Budapest, we followed the protocol described in ref. 79.

Bioinformatic processing

All ancient-DNA libraries were sequenced with paired-end readson

Illumina instruments. We then performed the following steps: preproc-

essing, alignment and post-alignment filtering for variant calling. The

goal of preprocessing was to take raw sequenced products and create

merged sequences for alignment. We demultiplexed reads, binned

these to whichever library each read belonged to using the identify-

ing barcodes and indices, trimmed these identifying markers as well

as any residual adapter sequences, and merged each paired-end read

into a single molecule using the overlap of the paired-end reads as a

guide, employing a modified version of SeqPrep (https://github.com/

jstjohn/SeqPrep). We aligned the resulting sequencesto both the hg19

human genome reference (https://w ww.internationalgenome.org/

category/grch37/) and the inferred ancestral Reconstructed Sapiens

Referenc e Sequence mitochondr ial sequen ce

using the samse aligner

of bwa

. We marked duplicate molecules by barcode bin, based on the

same start and stop positions and orientation. The computational

pipelines with specific parameters used are available on GitHub at

https://github.com/dReichLab/ADNA-Tools and https://github.com/

dReichLab/adna-workflow.

We used a pseudohaploid genotyping approach to determine a ran-

domly selected allele at SNP sets of interest. To represent the allele at

each SNP, we randomly selected sequences from a pool of all sequences

covering that position with a minimum data quality; our criteria were

a minimum mapping quality of at least 10 and a base quality of at least

20, after trimming sequences by 2 base pairs at both the 5′ and 3′ ends

to remove damage artefacts. We assessed ancient-DNA authenticity

by using contamMix-1.0.105182 to search for heterogeneity in mtDNA

sequences, which are expected to be non-variable in uncontaminated

individuals, and also ANGSD to test for heterogeneity in X-chromosome

sequences, which are expected to be homozygous in males83. We further

evaluated the authent icity of the ancient samples by using pmdtools

to measure the rate of cytosine-to-thymine mutations in the first and

last nucleotides (in untrimmed sequences), which is expected for genu-

ine ancient DNA

, and by computing the ratio of Y chromosomes to

the sum of X and Y chromosome sequences, which is expected to be

very low for females and to have a much higher value for males. We

determined a consensus for mtDNA using bcftools (https://github.com/

samtools/bcftools) and SAMTools 85, requiring a minimum of two-fold

coverage to call the nucleotide and a majority rule to determine its

value. We used HaploGrep2 to determine mitochondrial haplogroups

based on the phylotree database (mtDNA tree build 17)86,87.

PCA

We projected individuals in Fig.1b in smartpca

using parameters

newshrink: YES and lsqporject: YES on a PCA space with axes formed

by the following populations: OberkasselCluster (a set of trans-

Alpine WHG individuals identified in ref. 20), Russia_Firsovo_N, Iran_

HajjiFiruz_C

, Iran_C_SehGabi

, Iran_C_TepeHissar

, Israel_C

and

Germany_EN_LBK2,40,79,91. The coordinates of plotted points are shown

in Supplementary Table5.

FST estimation

We computed FST in smartpca88 with parameters inbreed: YES and

fstonly: YES92.

Drawing maps

We d rew th e ma ps in F ig s.1 a nd5, Extended Data Figs.1 and 5, and Sup-

plementary Information section2 using public-domain Natural Earth

data with the rnaturalearth package in R54. Digital elevation maps in

Supplementary Information section1 were drawn using the Coperni-

cus digital elevation model (https://doi.org/10.5270/ESA-c5d3d65).

Visualizing the three Eneolithic clines and preceding

populations

We fit models for Eneolithic cline populations (Fig.1c) using

qpAdm

and with the following set of right populations: OldAfrica,

Russia_AfontovaGora3, CHG, Iran_GanjDareh_N, Italy_Villabruna,

Russia_Sidelkino.SG and Turkey_N (Fig.1c). Diverse ternary models of

preceding, Eneolithic and Bronze Age populations are shown in Fig.2.

Individuals plotted at the triangle edge fit the simpler two-source

model (P > 0.05) (in some of these cases, the three-source models have

a negative coefficient from one of the three sources). The corners of

each tri angle represent t he sources. Unp lotted individu als all gave

fits at P < 0.05 and so should be viewed as poorly described by the

model.

Model competition with qpAdm/qpWave

We used qpWave/qpAdm methods

2,18

to characterize relationships

among diverse target and source populations from the steppe and

adjacent areas (Supplementary Information section2). We use

OldAfrica, Russia_AfontovaGora3, CHG, Iran_GanjDareh_N, Italy_

Villabruna, Russia_Sidelkino.SG and Turkey_N as the set of right popu-

lations for most ana lyses. For analysis of Anatolians, we expanded

this to OldAfrica, CHG, Iran_GanjDareh_N, Italy_Villabruna, Russia_

AfontovaGora3, Russia_Sidelkino.SG, TUR_Marmara_Barcın_N, TUR_C_

Boncuklu_PPN and TUR_C_Çatalhöyük_N, Natufian to gain leverage

for differentiating among West Asian sources. For faster computa-

tion, we ran qpWave/qpAdm on precomputed output from qpfstats

runs (https://github.com/DReichLab/AdmixTools/blob/master/

qpfs.pdf) with a poplistname that includes Han.DG, and all target,

source and right populations, and parameters allsnps: YES, inbreed:

NO. We performed separate qpWave/qpAdm runs directly on genotype

files as needed when the target or source populations were not present

in the qpfstats output with parameter basepop: Han.DG. We identi-

fied feasible models as having P > 0.05, all standard errors ≤0.1, and

admixture proportions ≤2 standard errors from 0 and 1. We removed

target or source populations from the right set. Competition of mod-

els A and B involves two qpWave/qpAdm runs in which all sources

of A\B and B\A (\ denotes set difference) are placed on the right set.

Details of all analyses can be found in Supplementary Information

section2.

Y-chromosome haplogroup inference

We used the methodology described in ref. 6, which used the YFull

YTree v.8.09 phylogeny (https://github.com/YFullTeam/ YTree/blob/

master/ytree/tree_8.09.0.json) to denote Y-chromosome haplogroups

in terminal notation93.

Estimates of dates of admixture

We u sed DATES 8,41 to estimate dates of admixture for the Core Yamnaya,

Don Yamnaya, Eastern European Yamnaya, Corded Ware and Caucasus–

Anatolian populations (Extended Data Fig.2). For the Core Yamnaya

and Caucasus–Anatolian populations, we used sets of diverse West

Asian and European hunter-gatherer populations as the two sources.

For the Don Yamnaya, we used the Core Yamnaya and UNHG as the

two sources. For the Eastern European Yamnaya, we used the Core

Yamnaya and a diverse set of Neolithic/Chalcolithic European farmers

from Extended Data Fig.3b. For the Corded Ware, we used the Core

Yamnaya and Globular Amphora as the two sources. It is more impor-

tant to use many source samples even if they are genetically somewhat

drifted to the true ones; picking the wrong sources does not bias the

date estimate41.

IBD segment detection

We used ancIBD

to detect IBD segments of length greater than or

equal to 8 cM. The pre-Yamnaya individuals plotted in Fig.5 are from

the period 5500–3500 BC.

Estimates of geographical distance

To study the decay of IBD with geographical distance, we estimated the

distance between sites on the basis of their latitude and longitude, given

in Supplementary Table4, using the Haversine distance as implemented

in distHaversine94 of the package geosphere in R.

Estimates of effective population size

We ran HapNe-LD (v.1.20230726)

using the default parameters and

providing pseudo-haploid genotypes as input. In brief, HapNe-LD uses

a summary statistic that measures long-range correlations between

markers to infer fluctuations in effective population size (defined as

the inverse of the coalescence rate) over time. We studied two distinct

sets of unrelated individuals, all of which had a coverage of at least 0.7×

on the target autosomal SNPs and with a standard deviation on their

estimated date smaller than 180 years (about 6 generations). The first

group consisted of 25 Core Yamnaya individuals with estimated dates

ranging between 4,500 and 4,800 years before present. The second

group contained 26 Core Yamnaya individuals ranging from 4,800 to

5,100 years before present.

If no evidence of effective population-size fluctuations can be

detected in the data, HapNe-LD produces a flat line. An output contain

ing fluctuations should thus be interpreted as the detection of changes

in historical effective population size. Recent admixture between

highly differentiated populations (F

> 0.1) might lead to biases in

LD-based analyses that induce fluctuations similar to a population

bottleneck. However, HapNe implements a test to flag the presence of

recent structure in the data, which was not detected in either sample

set (approximate P ≥ 0.1), indicating that the observed signal instead

reflected variation in the effective population size of these groups.

In our analyses, the effective population size was defined as the

inverse of the instantaneous coalescence rate. This quantity corre-

sponds to twice the number of breeding individuals in an idealized

population. As well as changes in the number of individuals in the

population (census size), several factors, such as changes in popula-

tion structure, selection and cultural practices

, can have an influ-

ence on the effective population size. These factors may in part be

responsible for the effective size fluctuations observed in the Core

Yamnaya.

We inferred approximate confidence intervals using bootstrap with

different chromosome arms as resampling units. We determined the

beginning of the expansion by using the location of the minimum of

each bootstrapped trajectory. We converted the results into years by

assuming 28.6 years per generation for the median minimum location,

and 25.6 and 31.5 years per generation for the lower and upper bounds,

respectively96. We used these values, corresponding to the estimated

Article

number of years per generation for males (31.5) and females (25.6), to

account for uncertainty in the conversion factor.

Ethics statement. The individuals studied in this work were all ana-

lysed with the goal of minimizing damage to their skeletal remains,

with permission from local authorities in each location from which

they came. Every sample is represented by stewards, such as archae-

ologists or museum curators, who are either authors or are thanked

in the Acknowledgements. Open-science principles require making

all the data used to support the conclusions of a study maximally

available, and we support these principles here by making publicly

available not only the digital copies of molecules (the uploaded

sequences), but also the molecular copies (the ancient-DNA libraries,

which constitute molecular data storage). Researchers who wish to

carry out deeper sequencing of the libraries published in this study

can make a request to corresponding author D.R. We commit to grant-

ing reasonable requests as long as the libraries remain preserved

in our laboratories, with no requirement that we be included as

collaborators or co-authors on any resulting publications.

Reporting summary

Furth er infor mation on researc h desig n is availa ble in th eNature Po rt-

folio Reporting Summary linked to this article.

Data availability

Genotype data for individuals included in this study can be obtained

from the Harvard Dataverse repository at https://doi.org/10.7910/

DVN/QGNMRH. The DNA seq uences re ported in this paper h ave

been deposited in the European Nucleotide Archive under acces-

sion number PRJEB81467. Other newly reported data, such as

radiocarbon dates and archaeological context information, are

included in this paper and the Supplementary Information.

56. Järve, M. etal. Shifts in the genetic landscape of the Western Eurasian steppe

associated with the beginning and end of the Scythian Dominance. Curr. Biol. 29,

2430– 2441 (2019 ).

57. Jeong, C. etal. A dynamic 6,000-year genetic history of Eurasia’s eastern steppe. Cell

183, 890–904 (2020).

58. Kumar, V. etal. Bronze and Iron Age population movements underlie Xinjiang population

history. Science 376, 62–69 (2022).

59. Patterson, N. etal. Large-scale migration into Britain during the Middle to Late Bronze

Age. Nature 601, 588–594 (2022).

60. Olalde, I. etal. The genomic history of the Iberian Peninsula over the past 8000 years.

Science 363, 1230–1234 (2019).

61. Zhang, F. etal. The genomic origins of the Bronze Age Tarim Basin mummies. Nature 599,

256–261 (2021).

62. Wang, C.-C. etal. Genomic insights into the formation of human populations in East Asia.

Nature 591, 413–419 (2021).

63. Gelabert, P. etal. Genomes from Verteba cave suggest diversity within the Trypillians in

Ukraine. Sci. Rep. 12, 7242 (2022).

64. Mattila, T. M. etal. Genetic continuity, isolation, and gene flow in Stone Age Central and

Eastern Europe. Commun. Biol. 6, 793 (2023).

65. Pinhasi, R., Fernandes, D. M., Sirak, K. & Cheronet, O. Isolating the human cochlea to

generate bone powder for ancient DNA analysis. Nat. Protoc. 14, 1194–1205 (2019).

66. Sirak, K. A. etal. A minimally-invasive method for sampling human petrous bones from

the cranial base for ancient DNA analysis. Biotechniques 62, 283–289 (2017).

67. Damgaard, P. B. etal. Improving access to endogenous DNA in ancient bones and teeth.

Sci Rep. 5, 11184 (2015).

68. Sirak, K. etal. Human auditory ossicles as an alternative optimal source of ancient DNA.

Genome Res. 30, 427–436 (2020).

69. 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,

15758–15763 (2013).

70. Korlević, P. etal. Reducing microbial and human contamination in DNA extractions from

ancient bones and teeth. Biotechniques 59, 87–93 (2015).

71. Rohland, N., Harney, E., Mallick, S., Nordenfelt, S. & Reich, D. Partial uracil-DNA-glycosylase

treatment for screening of ancient DNA. Philos. Trans. R. Soc. Lond. B Biol. Sci. 370,

20130624 (2015).

72. Rohland, N., Glocke, I., Aximu-Petri, A. & Meyer, M. Extraction of highly degraded DNA

from ancient bones, teeth and sediments for high-throughput sequencing. Nat. Protoc.

13, 2447–2461 (2018).

73. Prendergast, M. E. etal. Ancient DNA reveals a multistep spread of the first herders into

sub-Saharan Africa. Science 365, eaaw6275 (2019).

74. 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 oth er sources of highly degrad ed DNA. Nat. Protoc. 15, 2279–2300 (2020).

75. Maricic, T., Whitten, M. & Pääbo, S. Multiplexed DNA sequence capture of mitochondrial

genomes using PCR products. PLoS ONE 5, e14004 (2010).

76. Fu, Q. etal. DNA analysi s of an early moder n human from Tianyu an Cave, China. Proc.

Natl Acad. Sci. USA 110, 2223–2227 (2013).

77. Fu, Q. etal. An early mode rn human from Roma nia with a recent Nea nderthal ance stor.

Nature 524, 216–219 (2015).

78. Rohland, N. etal. Three assays for in-solution enrichment of ancient human DNA at more

than a million SNPs. Genome Res. 32, 2068–2078 (2022).

79. Lipson, M. etal. Parallel palaeogenomic transects reveal complex genetic history of early

European farmers. Nature 551, 368–372 (2017).

80. Behar, D. M. etal. A “Copernican” reassessment of the human mitochondrial DNA tree

from its root. Am. J. Hum. Genet. 90, 675–684 (2012).

81. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler transform.

Bioinformatics 25, 1754–1760 (2009).

82. Fu, Q. etal. A revised timescale for human evolution based on ancient mitochondrial

genomes. Curr. Biol. 23, 553–559 (2013).

83. Korneliussen, T. S., Albrechtsen, A. & Nielsen, R. ANGSD: analysis of next generation

sequencing data. BMC Bioinformatics 15, 356 (2014).

84. Skoglund, P. etal. Separating endogenous ancient DNA from modern day contamination

in a Siberian Neandertal. Proc. Natl Acad. Sci. USA 111, 2229–2234 (2014).

85. Li, H. etal. The sequence alignment/map format and SAMtools. Bioinformatics 25,

2078–2079 (2009).

86. Weis senste iner, H . etal. Haplo Grep 2: mitoc hondr ial hap logro up cla ssific ation in the era

of high-throughput sequencing. Nucleic Acids Res. 44, W58–W63 (2016).

87. van Oven, M. & Kayser, M. Updated comprehensive phylogenetic tree of global human

mitochondrial DNA variation. Hum. Mutat. 30, E386–E394 (2009).

88. Patterson, N., Price, A. L. & Reich, D. Population structure and eigenanalysis. PLoS Genet.

2, e190 (2006).

89. Shinde, V. etal. An ancient Harappan genome lacks ancestry from steppe pastoralists or

Iranian farmers. Cell 179, 729–735 (2019).

90. Harney, É. etal. Ancient DNA from Chalcolithic Israel reveals the role of population

mixture in cultural transformation. Nat. Commun. 9, 3336 (2018).

91. 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).

92. Reich, D., Thangaraj, K., Patterson, N., Price, A. L. & Singh, L. Reconstructing Indian

population history. Nature 461, 489–494 (2009).

93. Adamov, D., Gurianov, V. M., Karzhavin, S., Tagankin, V. & Urasin, V. Defining a new rate

constant for Y-chromosome SNPs based on full sequencing data. Russian J. Genet. Geneal.

7, 1920–2997 (2015).

94. Sinnott, R. W. Virtues of the Haversine. Sky Telescope 68, 158 (1984).

95. Charlesworth, B. Effective population size and patterns of molecular evolution and

variation. Nat. Rev. Genet. 10, 195–205 (2009).

96. Fenner, J. N. Cross-cultural estimation of the human generation interval for use in

genetics-based population divergence studies. Am. J. Phys. Anthropol. 128, 415–423

(2005).

Acknowledge ments We thank A. G. Nikitin for advice and feedback; N. Adamski, R. Bernardos,

N. Broomandkhoshbacht, D. Fernandes, M. Ferry, E. Harney, K. Mandl, S. Nordenfelt,

K. Stewardson, B. G. Mende and Z. Zhang for laboratory or bioinformatics work; L. Bembeeva,

B. Preda-Bălănică, I. Ecsedy, A. I. Gotlib, V. M. Heyd, S. A. Mikhailovich, N. Morgunova,

A. Soficaru, S. S. Tur and P. Włodarczak for anthropological work and comments; I. Vyazov

for help in adapting illustrati ons; andA. A . Shalapini n, A. F. Kochkina an d D. A. Stashen kov

for archa eological fieldwork a nd sample col lection. T.Ha. wa s supporte d by the Hungar ian

Research, Development and Innovation Office (grant FK128013), the Bolyai Scholarship of

the Hungarian Academy of Sciences, and by the ÚNKP-23-5 New National Excellence Program

of the Ministry for Culture and Innovation from the National Research, Development and

Innovation Fund. P.F. and L.V. were supported by the Czech Ministry of Education, Youth and

Sports (programme ERC CZ, project LL2103). P.F. was supported by the Czech Science

Founda tion (pro ject 21-2 7624S), and the Europ ean Union Operatio nal Progr amme Just

Transition (LERCO project CZ.10.03.01/00/22_003/0000003). We acknowledge support from

Polish scientific project grant NCN OPUS 2015/17/B/HS3/01327, and the Russian Science

Foundation grants 21-18-00026 to N.I.S. and 22-18-00470 to A.A.T. We acknowledge support

from the Museum of the Institute of Plant and Animal Ecology (UB RAS) to P. Kosintsev. K.N.S.

was supported by grant FWRZ-2021-0006. This study depended on support from the research

computing group at Harvard Medical School. The ancient-DNA data generation and analysis

was supported by the National Institutes of Health (R01-HG012287), the John Templeton

Foundation (grant 61220), a gift from J.-F. Clin, the Allen Discovery Center programme,

a Paul G. Allen Frontiers Group advised programme of the Paul G. Allen Family Foundation,

and the Howard Hughes Medical Institute (to D.R.). The accepted version of this article (before

the editing, proofreading and formatting changes following the paper being accepted) is

subject to the Howard Hughes Medical Institute (HHMI) Open Access to Publications policy;

HHMI lab heads have previously granted a non-exclusive CC BY 4.0 licence to the public and

a sublicensable licence to HHMI in their research articles. Pursuant to those licences, the

accepted manuscript can be made freely available under a CC BY 4.0 licence immediately

on publication.

Author contr ibutions I.L., N.P., D.A., L.V. and D.R. wrote the manuscript. A.S.-N., P.F.P., S.M.,

N.R., R.P. and D.R. supervised parts of the study. I.L. and N.P. carried out the main genetic

analyses. R.F., H.R., I.O. and P.F.P. contributed other genetic analyses. D.A. and L.V. edited

archaeological information. D.A., A.A.K., E.P.K., N.I.S., S.C.A., E.B., Z.B., A.B., P.C., A.A.C., I.C.,

M. Constantinescu, M. Csányi, J.D., S.É., A. Faifert, P.F., A. Frînculeasa, M.N.F., T.Ha., T.Hi, P. Jelínek,

V.I.K., V.K., A. Kitova, A. Korolev, G.K., N.A.N., B.B., P.K.D., P. Jarosz, E.N.K., A.V.K., J.K., P.Ko., P.Ku.,

R.M., M.Mi., E. Melis, V.M., E. Molnár, J.M., M.N., O.N., M.P.R., M.O.-G., G.P., S.P., T.M.S., N.N.S.,

A.Š., I.S., V.N.S., A. Simalcsik, K.S., K.N.S., J.T., A.A.T., V.T., S.V., O.F., A.S.-N., D.S.A., A.M.M., S.A.A.,

R.S.M., and R.P. sampled anthropological remains and/or contributed to the creation of the

archaeological supplement. N.I.S. shared previously unpublished radiocarbon dates. A.A.,

E.S.B., M. Mah, A.M. and S.M. processed bioinformatic data. K.C., F.C., O.C., E.C., L.I., A. Kearns,

D.K. , A .M. L., M. M ich el, J.O., L.Q. , J.N .W., F.Z . an d N. R. c arr ied out wet lab ora tor y wo rk.

Competing interests The authors declare no competing interests.

Additional informat ion

Supplementary information The online version contains supplementary material available at

https://doi.org/10.1038/s41586-024-08531-5.

Correspondence and requests for materials should be addressed to Iosif Lazaridis,

Nick Patterson, David Anthony, Leonid Vyazov, Ron Pinhasi or David Reich.

Peer review inform ation Nature thanks Kristian Kristiansen and the other, anonymous,

reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports

are available.

Reprint s and permiss ions informa tion is available at http://www.nature.com/reprints.

Article

Extende d Data Fig. 1 | See next page for caption.

Extende d Data Fig. 1 | The origin of Central Ana tolian Bronze Age peop le.

a, Models with eas tern steppe sources (inclu ding CLV and Serednii Stih). Fittin g

models includ e Mesopotamian (Çayönü) and stepp e ancestry. b, Models with

western sou rces, including Usatove and those fro m Southeastern Europ e fail

except those with Çayön ü and either Mayaky or Boyanovo EBA (both of which

are Yamnaya-derived). c, The st eppe (BPg roup)+Çayö nü model fa ils all

Chalcolithi c/Bronze Anatolian s exceptpeople of the Central Anatoli an Bronze

Age. d, Steppe (BPgrou p) ancestr y in the BPgroup+Çayönü model is obs erved

in all individuals of the Central Anatolian Bronze Age (mean and ±3 s.e. estimated

by qpAdm are shown for all Chalcolithic and Bronze Age indiv iduals from

Anatolia that f it the model at p > 0.05) as well as in individual ART027_d from

Chalcolithi c Arslantepe in Eastern An atolia. e, BPgroup-relat ed ancestry

admixed with dif ferent substrata: Ak nashen-related in the Nor th Caucasus

Maikop, Masis Blur-related in Chalcolit hic Armenia, and Mesop otamian-related

(Çayönü) in the ancestors of the Centra l Anatolian Bronze Age, following the

route (f) from the North Cauc asus to Anatolia; sites with BPgro up-related

ancestr y marked in bold. In all panels p-values estim ated by qpWave are shown.

Article

Extende d Data Fig. 2 | Admixture date estima tes. We estimate admixt ure

dates for the Core Yamnaya as a mixture of European hunte r-gatherer and West

Asian populat ions (a), for the Don Yamnaya as a mixture of Core Yamnaya and

UNHG (b), for the Bulgaria, Moldova, Roman ia, and Serbia (BMRS) Yamnaya as a

mixture of Core Yamnaya and European Ne olithic/Chalcolithi c farmers (c), for

the Corded Ware as a mixture of Core Yamnaya and Globula Amphor a (d), and

for a combined Cauca sus-Anatolia populatio n (Maikop-Armenia_C-TUR _C_BA)

a mixture of European hunt er-gatherer and West Asian populati ons which

occurred ca . 4400 BC (e). The Core Yamnaya were formed ca. 4000 BC, followed

by admixture ca. 33 50 BC with UNHG and Europe an farmers in the east and west

of the Dnipro-Don re gion and <3000 BC in central-eas tern Europe.

Extende d Data Fig. 3 | Population struc ture in people with a Yamnaya

cultural af filiati on. a,Individuals are project ed in the same space as in Fig.1,

showing that the Core Yamnaya cluste r (red fill symb ols) from diverse sites is

different iated from the Don Yamnaya (blue fill) who tend towards the UNH G.

b, Yamnaya ind ividuals in the West (Ukra ine, Hung ary, Slovakia, and

Southea stern Europe) include a tight cluste r of individua ls as well as others

that tend towards the direc tion of European Neolithi c and Chalcolithic groups

from Romania and Hunga ry. Individuals from Russia are shown in grey circle s in

panel b. Coordinate s of plotted points can be found in Supplem entary Table6.

Article

Extende d Data Fig. 4 | A 4-way model for the entire Dnipro-D on-Volga-Cauca sus region. Mean and ±1 st andard error estimated by qpAdm is shown .

Extende d Data Fig. 5 | The origin of Indo-Ana tolian and Indo-Eu ropean

languages. Genetic rec onstructio n of the ancestr y of Pontic-Caspian ste ppe

and West Asian populat ions points to the North Cauc asus-Lower Volga area as

the homeland of Indo-Anatolian languages and to the Serednii Stih archaeological

culture of the Dnipro-D on area as the homeland of Indo-Europ ean languages.

The Caucas us-Lower Volga people had diverse dist al roots, estimated usin g

the qpAdm softwa re on the left barplot, as Caucasus hun ter-gatherer (purple),

Central Asia n (red), Eastern hunt er-gatherer (pink), and West Asian Neolithic

(green). Cauc asus-Lower Volga expansions, es timated using qpAdm on the

right barplo t, disseminated Cauc asus Neolithic (blue)-Lowe r Volga Eneolith ic

(orange) proximal ancestr ies, mixing with the inhabi tants of the North Pontic

region (yellow), Volga region (yel low), and West Asia (green).

Article

Extended Data Table 1 | FST valu es amo ng sele ct pop ulations of the Dnipr o, Don, Volga , and C aucasu s area s

FST val ues a re sh own b elow the d iago nal and t heir stan dard erro rs ab ove i t.

Extended Data Table 2 | Extraordinary genetic homogeneity in the Core Yamnaya

We tested all populations and individuals for cladality with Samara Yamnaya. We list populations for which this is not rejected (qpWave p > 0.05) and populations that include individuals that fit

Core Yamnaya selection criteria (qpWave p > 0.2, at least 300k SNPs, and Yamnaya or Afanasievo culture).

Article

Extended Data Table 3 | FST valu es amo ng popu latio ns that include Core Yam naya individuals

FST val ues a re sh own b elow the d iago nal and t heir stan dard erro rs ab ove i t.

Extended Data Table 4 | qpAdm mode ls tha t fit no n-Core Yamnaya

We use the following sources to model Yamnaya-related populations other than the Core and Don Yamnaya: CoreYamnaya, Romania_C_Bodrogkeresztur, Romania_N,

Serbia_IronGates_Mesolithic, Trypillia, Ukraine_N, Usatove. The Baden individuals from Hungary represent a reburial into a kurgan54 and are predominantly of European farmer, not Yamnaya,

ancestry. The Riltsi individual is shown with Usatove ancestry here and can also be modeled with about half Remontnoye ancestry, as the Usatove have ancestry from the CLV cline17.

Article

Extended Data Table 5 | Cross-regional shared identity-by-descent (IBD) segments

We list all segments≥12 cM shared between individuals from two different regions defined as follows. “Dnipro cline”: CoreYamnaya, GK1, GK2, Russia_Don_EBA_Yamnaya, SShi, SSlo,

SSmed, Ukraine_N. Volga River basin ancestry gradients (downriver “Volga Cline” and upriver “European Hunter-Gatherer Cline”): Ekaterinovka, Khi, KhlopkovBugor, Klo, Kmed, Labazy,

Lebyazhinka_HG, Maximovka, Murzikha, Syezzheye, UpperVolga. “Caucasus-Lower Volga Eneolithic”: BPgroup, PVgroup. “CLV-South”: Remontnoye, Maikop, Unakozovskaya, Armenia_C,

TUR_C_Kalehöyük_MLBA, TUR_C_Ovaören_EBA.

A preview of this full-text is provided by Springer Nature.

Learn more

Content available from Nature

This content is subject to copyright. Terms and conditions apply.

A genomic history of the North Pontic Region from the Neolithic to the Bronze Age

Article

Full-text available

Feb 2025
NATURE

The North Pontic Region was the meeting point of the farmers of Old Europe and the foragers and pastoralists of the Eurasian steppe1,2, and the source of migrations deep into Europe3, 4–5. Here we report genome-wide data from 81 prehistoric North Pontic individuals to understand the genetic makeup of its people. North Pontic foragers had ancestry from Balkan and Eastern hunter-gatherers⁶ as well as European farmers and, occasionally, Caucasus hunter-gatherers. During the Eneolithic period, a wave of migrants from the Caucasus–Lower Volga area⁷ bypassed local foragers to mix in equal parts with Trypillian farmers, forming the people of the Usatove culture around 4500 bce. A temporally overlapping wave of migrants from the Caucasus–Lower Volga blended with foragers instead of farmers to form Serednii Stih people⁷. The third wave was the Yamna—descendants of the Serednii Stih who formed by mixture around 4000 bce and expanded during the Early Bronze Age (3300 bce). The temporal gap between Serednii Stih and the Yamna is bridged by a genetically Yamna individual from Mykhailivka, Ukraine (3635–3383 bce), a site of archaeological continuity across the Eneolithic–Bronze Age transition and a likely epicentre of Yamna formation. Each of these three waves of migration propagated distinctive ancestries while also incorporating outsiders, a flexible strategy that may explain the success of the peoples of the North Pontic in spreading their genes and culture across Eurasia3, 4–5,8, 9–10.

A Critical Analysis of “The Genetic Origin of the Indo-Europeans” as Published in the Journal Nature by Lazaridis et al on February 5, 2025.

Method

Full-text available

Feb 2025

Michael St. Clair

The exploration of language prehistory should utilize an empirical methodology rather than lexical reconstructions and statistical modeling.

Beating the retreat from the Steppe hypothesis

Preprint

Full-text available

Feb 2025

Paul Heggarty

Commentary by Paul Heggarty on: Lazaridis et al. (2025): 'The genetic origin of the Indo-Europeans' https://doi.org/10.1038/s41586-024-08531-5

Inferring domestic goat demographic history through ancient genome imputation

Preprint

Full-text available

Apr 2025

Goats were among the earliest managed animals, making them a natural model to explore the genetic consequences of domestication. However, a challenge in ancient genomic analysis is the relatively low genome coverage for most samples, limiting analysis to pseudohaploid genotypes. Genotype imputation offers potential to alleviate this limitation by improving information content and accuracy in low coverage genomes. To test this we used published high coverage (>8✕) goat palaeogenomes, imputing downsampled genomes using the VarGoats dataset (1,372 individuals) as a reference panel. Measuring concordance between imputed and high coverage genotypes, we find high concordance after filtering for common (>5%), high confidence variants, with 0.5✕ genomes reaching >0.97 concordance. There is a trade-off between coverage, genotype probability (GP) thresholds, and genotype recovery, where higher coverage and more lenient GP thresholds result in higher recovery, and a reduction in heterozygous false positive rates with stricter thresholds. We then imputed 36 goat palaeogenomes with ≥0.5✕ coverage to examine runs-of-homozygosity (ROH) and identity-by-descent (IBD) patterns. Using a novel approach combining ROH profiles across tools, we find that among Neolithic goats, ROH increases with distance from the Zagros Mountains, suggesting a large effect of the initial dispersal of managed herds. Inbreeding levels decrease across Southwest Asia in more recent periods. IBD mirrored this pattern, with less relatedness in the early herding site of Ganj Dareh compared to higher relatedness in goats from later in the dispersal process. These findings provide insights into the genetic consequences of early goat management on demography, and confirm the utility of imputation in leveraging low coverage palaeogenomes.

No collapse in sight: enclosures, the mortuary arena and the Big Other in the Trypillia world

Article

Apr 2025

Long-term hunter-gatherer continuity in the Rhine-Meuse region was disrupted by local formation of expansive Bell Beaker groups

Preprint

Full-text available

Mar 2025

The first phase of the ancient DNA revolution painted a broad-brush picture of European Holocene prehistory, whereby 6500-4000 BCE, farmers descending from western Anatolians mixed with local hunter-gatherers resulting in 70-100% ancestry turnover, then 3000-2500 BCE people associated with the Corded Ware complex spread steppe ancestry into north-central Europe. We document an exception to this pattern in the wider Rhine-Meuse area in communities in the wetlands, riverine areas, and coastal areas of the western and central Netherlands, Belgium and western Germany, where we assembled genome-wide data for 109 people 8500-1700 BCE. Here, a distinctive population with high hunter-gatherer ancestry (~50%) persisted up to three thousand years later than in continental European regions, reflecting limited incorporation of females of Early European Farmer ancestry into local communities. In the western Netherlands, the arrival of the Corded Ware complex was also exceptional: lowland individuals from settlements adopting Corded Ware pottery had hardly any steppe ancestry, despite a characteristic early Corded Ware Y-chromosome. The limited influx may reflect the unique ecology of the region's river-dominated landscapes, which were not amenable to wholesale adoption of the early Neolithic type of farming introduced by Linearbandkeramik, making it possible for previously established groups to thrive, and creating a persistent but permeable boundary that allowed transfer of ideas and low-level gene flow. This changed with the formation-through-mixture of Bell Beaker using populations ~2500 BCE by fusion of local Rhine-Meuse people (9-17%) and Corded Ware associated migrants of both sexes. Their expansion from the Rhine-Meuse region then had a disruptive impact across a much wider part of northwest Europe, including Britain where its arrival was the main source of a 90-100% replacement of local Neolithic peoples.

The legacy of Luca Cavalli-Sforza on human evolution

Article

Full-text available

Feb 2025

Archaeology and the branch of population genetics focusing on the human past have historically lived parallel lives, often having complicated encounters when it came to unravelling the origins and evolution of Homo sapiens. These interactions were proven invaluable to obtain a deeper and more complete understanding of our past. At the same time, they sometimes uncovered biases and misinterpretations, with serious consequences for our understanding of data, methods and, most importantly, the history of our species. Cavalli-Sforza pioneered a real multidisciplinary approach, bridging population genetics and statistics with archaeology, human origins, and other fields in the humanities, inspiring researchers from these fields and blazing a trail for today’s successful interactions and collaborations. His legacy showed that these interdisciplinary approaches are possible and of vital importance, and exposed areas that still need significant development today.

Genetic and geographical origins of Eurasia’s influential Yamna culture

Article

Feb 2025
NATURE

How one language family took over the world: ancient DNA traces its spread

Article

Feb 2025
NATURE

Ewen Callaway

A genomic history of the North Pontic Region from the Neolithic to the Bronze Age

Article

Full-text available

Feb 2025
NATURE

Human DNA from the oldest Eneolithic cemetery in Nalchik points the spread of farming from the Caucasus to the Eastern European steppes

Article

Full-text available

Oct 2024

The Darkveti-Meshoko culture (c.5000–3500/3300 BCE) is the earliest known farming community in the Northern Caucasus, but its contribution to the genetic profile of the neighboring steppe herders has remained unclear. We present analysis of human DNA from the Nalchik cemetery—the oldest Eneolithic site in the Northern Caucasus—which shows a link with the LowerVolga’s first herders of the Khvalynsk culture. The Nalchik male genotype combines the genes of the Caucasus hunter-gatherers, the Eastern hunter-gatherers and the Pre-Pottery Neolithic (PPN) farmers of western Asia. Improved comparative analysis suggests that the genetic profile of certain Khvalynsk individuals shares the genetic ancestry of the Unakozovo-Nalchik type population of the Northern Caucasus’ Eneolithic. Therefore, it seems that in the first half of the 5th millennium BCE, cultural and mating networks helped agriculture and pastoralism spread from West Asia across the Caucasian, into the steppes between the Don and the Volga in Eastern Europe.

Population genomics of post-glacial western Eurasia

Article

Full-text available

Jan 2024
NATURE

Western Eurasia witnessed several large-scale human migrations during the Holocene1–5. Here, to investigate the cross-continental effects of these migrations, we shotgun-sequenced 317 genomes—mainly from the Mesolithic and Neolithic periods—from across northern and western Eurasia. These were imputed alongside published data to obtain diploid genotypes from more than 1,600 ancient humans. Our analyses revealed a ‘great divide’ genomic boundary extending from the Black Sea to the Baltic. Mesolithic hunter-gatherers were highly genetically differentiated east and west of this zone, and the effect of the neolithization was equally disparate. Large-scale ancestry shifts occurred in the west as farming was introduced, including near-total replacement of hunter-gatherers in many areas, whereas no substantial ancestry shifts happened east of the zone during the same period. Similarly, relatedness decreased in the west from the Neolithic transition onwards, whereas, east of the Urals, relatedness remained high until around 4,000 bp, consistent with the persistence of localized groups of hunter-gatherers. The boundary dissolved when Yamnaya-related ancestry spread across western Eurasia around 5,000 bp, resulting in a second major turnover that reached most parts of Europe within a 1,000-year span. The genetic origin and fate of the Yamnaya have remained elusive, but we show that hunter-gatherers from the Middle Don region contributed ancestry to them. Yamnaya groups later admixed with individuals associated with the Globular Amphora culture before expanding into Europe. Similar turnovers occurred in western Siberia, where we report new genomic data from a ‘Neolithic steppe’ cline spanning the Siberian forest steppe to Lake Baikal. These prehistoric migrations had profound and lasting effects on the genetic diversity of Eurasian populations.

Accurate detection of identity-by-descent segments in human ancient DNA

Article

Full-text available

Dec 2023
Nat Genet

Long DNA segments shared between two individuals, known as identity-by-descent (IBD), reveal recent genealogical connections. Here we introduce ancIBD, a method for identifying IBD segments in ancient human DNA (aDNA) using a hidden Markov model and imputed genotype probabilities. We demonstrate that ancIBD accurately identifies IBD segments >8 cM for aDNA data with an average depth of >0.25× for whole-genome sequencing or >1× for 1240k single nucleotide polymorphism capture data. Applying ancIBD to 4,248 ancient Eurasian individuals, we identify relatives up to the sixth degree and genealogical connections between archaeological groups. Notably, we reveal long IBD sharing between Corded Ware and Yamnaya groups, indicating that the Yamnaya herders of the Pontic-Caspian Steppe and the Steppe-related ancestry in various European Corded Ware groups share substantial co-ancestry within only a few hundred years. These results show that detecting IBD segments can generate powerful insights into the growing aDNA record, both on a small scale relevant to life stories and on a large scale relevant to major cultural-historical events.

New radiocarbon and stable isotope data from the Usatove culture 1 site of Mayaky in Ukraine

Article

Full-text available

Dec 2023

21 22 The Usatove (Usatovo) culture provides a pivotal link between farmer and forager-pastoralist 23 cultures of the North Pontic region and adjacent areas during the Eneolithic-Early Bronze Age 24 (EN-EBA) transition in southeast Europe. Usatove is conventionally dated to the middle of the 25 4 th-early 3 rd millennium BCE. However, the AMS dating of human remains from the Mayaky 26 complex of the Usatove culture in Ukraine has produced dates beginning in the mid 5 th 27 millennium BCE. In this study, we report a series of new radiocarbon dates and stable isotope 28 ratios from human remains at Mayaky and present an attempt to evaluate this data in the 29 context of the Usatove culture chronology. The results suggest a diet of the individuals 30 attributed to the Usatove culture at Mayaky based on aquatic resources, potentially 31 contributing to a reservoir offset (RO) on the AMS dates obtained on human remains. The 32 presented data indicate a change in dietary preferences of the users of the Mayaky 33 archaeological complex in succeeding chronological periods. The presented as well as 34 previously published AMS radiocarbon dates establish the utilization of the Mayaky 35 archaeological site that precedes and postdates the Usatove culture. 36 37

Haplotype-based inference of recent effective population size in modern and ancient DNA samples

Article

Full-text available

Dec 2023

Individuals sharing recent ancestors are likely to co-inherit large identical-by-descent (IBD) genomic regions. The distribution of these IBD segments in a population may be used to reconstruct past demographic events such as effective population size variation, but accurate IBD detection is difficult in ancient DNA data and in underrepresented populations with limited reference data. In this work, we introduce an accurate method for inferring effective population size variation during the past ~2000 years in both modern and ancient DNA data, called HapNe. HapNe infers recent population size fluctuations using either IBD sharing (HapNe-IBD) or linkage disequilibrium (HapNe-LD), which does not require phasing and can be computed in low coverage data, including data sets with heterogeneous sampling times. HapNe shows improved accuracy in a range of simulated demographic scenarios compared to currently available methods for IBD-based and LD-based inference of recent effective population size, while requiring fewer computational resources. We apply HapNe to several modern populations from the 1,000 Genomes Project, the UK Biobank, the Allen Ancient DNA Resource, and recently published samples from Iron Age Britain, detecting multiple instances of recent effective population size variation across these groups.

Genetic continuity, isolation, and gene flow in Stone Age Central and Eastern Europe

Article

Full-text available

Aug 2023

The genomic landscape of Stone Age Europe was shaped by multiple migratory waves and population replacements, but different regions do not all show similar patterns. To refine our understanding of the population dynamics before and after the dawn of the Neolithic, we generated and analyzed genomic sequence data from human remains of 56 individuals from the Mesolithic, Neolithic, and Eneolithic across Central and Eastern Europe. We found that Mesolithic European populations formed a geographically widespread isolation-by-distance zone ranging from Central Europe to Siberia, which was already established 10,000 years ago. We found contrasting patterns of population continuity during the Neolithic transition: people around the lower Dnipro Valley region, Ukraine, showed continuity over 4000 years, from the Mesolithic to the end of the Neolithic, in contrast to almost all other parts of Europe where population turnover drove this cultural change, including vast areas of Central Europe and around the Danube River.

Early contact between late farming and pastoralist societies in southeastern Europe

Article

Full-text available

Jul 2023
NATURE

Archaeogenetic studies have described two main genetic turnover events in prehistoric western Eurasia: one associated with the spread of farming and a sedentary lifestyle starting around 7000–6000 BC (refs. 1–3) and a second with the expansion of pastoralist groups from the Eurasian steppes starting around 3300 BC (refs. 4,5). The period between these events saw new economies emerging on the basis of key innovations, including metallurgy, wheel and wagon and horse domestication(refs. 6–9). However, what happened between the demise of the Copper Age settlements around 4250 BC and the expansion of pastoralists remains poorly understood. To address this question, we analysed genome-wide data from 135 ancient individuals from the contact zone between southeastern Europe and the northwestern Black Sea region spanning this critical time period. While we observe genetic continuity between Neolithic and Copper Age groups from major sites in the same region, from around 4500 BC on, groups from the northwestern Black Sea region carried varying amounts of mixed ancestries derived from Copper Age groups and those from the forest/steppe zones, indicating genetic and cultural contact over a period of around 1,000 years earlier than anticipated. We propose that the transfer of critical innovations between farmers and transitional foragers/herders from different ecogeographic zones during this early contact was integral to the formation, rise and expansion of pastoralist groups around 3300 BC.

Palaeogenomics of Upper Palaeolithic to Neolithic European hunter-gatherers

Article

Full-text available

Mar 2023
NATURE

Modern humans have populated Europe for more than 45,000 years1,2. Our knowledge of the genetic relatedness and structure of ancient hunter-gatherers is however limited, owing to the scarceness and poor molecular preservation of human remains from that period³. Here we analyse 356 ancient hunter-gatherer genomes, including new genomic data for 116 individuals from 14 countries in western and central Eurasia, spanning between 35,000 and 5,000 years ago. We identify a genetic ancestry profile in individuals associated with Upper Palaeolithic Gravettian assemblages from western Europe that is distinct from contemporaneous groups related to this archaeological culture in central and southern Europe⁴, but resembles that of preceding individuals associated with the Aurignacian culture. This ancestry profile survived during the Last Glacial Maximum (25,000 to 19,000 years ago) in human populations from southwestern Europe associated with the Solutrean culture, and with the following Magdalenian culture that re-expanded northeastward after the Last Glacial Maximum. Conversely, we reveal a genetic turnover in southern Europe suggesting a local replacement of human groups around the time of the Last Glacial Maximum, accompanied by a north-to-south dispersal of populations associated with the Epigravettian culture. From at least 14,000 years ago, an ancestry related to this culture spread from the south across the rest of Europe, largely replacing the Magdalenian-associated gene pool. After a period of limited admixture that spanned the beginning of the Mesolithic, we find genetic interactions between western and eastern European hunter-gatherers, who were also characterized by marked differences in phenotypically relevant variants.

Proto-Indo-Anatolian, the “Anatolian Split” and the “Anatolian Trek”: A Comparative Linguistic Perspective

Chapter

May 2023

Alwin Kloekhorst

This book examines the impact of ancient DNA research and scientific evidence on our understanding of the emergence of Indo-European languages in prehistory. Offering cutting-edge contributions from an international team of scholars, it considers the driving forces behind the Indo-European migrations during the 3rd and 2nd millenia BC. The volume explores the rise of the world's first pastoral nomads the Yamnaya Culture in the Russian Pontic steppe including their social organization, expansions, and the transition from nomadism to semi-sedentism when entering Europe. It also traces the chariot conquest in the late Bronze Age and its impact on the expansion of the Indo-Iranian languages into Central Asia. In the final section, the volumes consider the development of hierarchical societies and the origins of slavery. A landmark synthesis of recent, exciting discoveries, the book also includes an extensive theoretical discussion regarding the integration of linguistics, genetics, and archaeology, and the importance of interdisciplinary research in the study of ancient migration.

The genetic origin of the Indo-Europeans

Abstract and Figures

Recommended publications