CulturechangeinEurope:ideasorpeople?

Gene6cevidenceofpopula6onexpansionandmigra6on

DNAevidencefrommodernpopula6ons

PC1fromclassicgene6cmarkers:EvidenceforNeolithicdemicexpansion?

EarlymtDNAresults(350bpseq)Evidenceagainstdemicexpansion?

Europeandistribu6onslooksimilar,anddifferentfromtheNearEast

LatermtDNA(completegenomes)Evidencefordemicexpansion?

Recentexpansion9KYA

Farmers?

Earlierexpansion20-10KYA

Hunter-gatherers?

LatermtDNA(completegenomes)Evidencefordemicexpansion?

Es6matedeffec6vepopula6onsize(Ne)oftypeH(red)andtypeUmtDNAhaplotypes(blue)forthecompleteEuropeanmtDNAdatasetaswellasforthesampleddataset.Fuetal.2012.[Notethatcolorsareflippedfrompreviousslide]

YchromosomedataStrongSE–NWcline

Ychromosomedata

Veryrecent,post-Neolithicexpansion

intoEurope?

TwohypothesesforspreadofIndo-Europeanlanguages

AnatolianhypothesisSpreadofearlyNeolithicfarmersfromAnatolia

KurganhypothesisSpreadofCopperandBronzeagepastoralistsfromthe

steppe

Diversifica6onofthemajorIndo-Europeansubfamilies

Linguis6csupportfortheAnatolianhypothesis.Bouckaertetal.2012

Hotoffthepress:170ancientEuropeanDNAgenomes!!!

Haaketal.2015

the Tyrolean Iceman3, 9 Late Copper/Early Bronze Age individuals(Yamnaya: ,3,300–2,700 BC), 15 Late Neolithic individuals (,2,500–2,200 BC), 9 Early Bronze Age individuals (,2,200–1,500 BC), two LateBronze Age individuals (,1,200–1,100 BC) and one Iron Age indivi-dual (,900 BC). Two individuals were excluded from analyses as theywere related to others from the same population. The average number ofSNPs covered at least once was 212,375 and the minimum was 22,869(Fig. 1).

We determined that 34 of the 69 newly analysed individuals weremale and used 2,258 Y chromosome SNPs targets included in the cap-ture to obtain high resolution Y chromosome haplogroup calls (Sup-plementary Information section 4). Outside Russia, and before the LateNeolithic period, only a single R1b individual was found (early NeolithicSpain) in the combined literature (n 5 70). By contrast, haplogroupsR1a and R1b were found in 60% of Late Neolithic/Bronze Age Europeansoutside Russia (n 5 10), and in 100% of the samples from EuropeanRussia from all periods (7,500–2,700 BC; n 5 9). R1a and R1b are themost common haplogroups in many European populations today18,19,and our results suggest that they spread into Europe from the East after3,000 BC. Two hunter-gatherers from Russia included in our study be-longed to R1a (Karelia) and R1b (Samara), the earliest documented ancientsamples of either haplogroup discovered to date. These two hunter-gatherers did not belong to the derived lineages M417 within R1a andM269 within R1b that are predominant in Europeans today18,19, but all7 Yamnaya males did belong to the M269 subclade18 of haplogroup R1b.

Principal components analysis (PCA) of all ancient individuals alongwith 777 present-day West Eurasians4 (Fig. 2a, Supplementary Infor-mation section 5) replicates the positioning of present-day Europeansbetween the Near East and European hunter-gatherers4,20, and the clus-tering of early farmers from across Europe with present day Sardinians3,4,suggesting that farming expansions across the Mediterranean to Spainand via the Danubian route to Hungary and Germany descended froma common stock. By adding samples from later periods and additionallocations, we also observe several new patterns. All samples from Russiahave affinity to the ,24,000-year-old MA1 (ref. 6), the type specimen for

the Ancient North Eurasians (ANE) who contributed to both Europeans4

and Native Americans4,6,8. The two hunter-gatherers from Russia (Kareliain the northwest of the country and Samara on the steppe near the Urals)form an ‘eastern European hunter-gatherer’ (EHG) cluster at one end ofa hunter-gatherer cline across Europe; people of hunter-gatherer ances-try from Luxembourg, Spain, and Hungary sit at the opposite ‘westernEuropean hunter-gatherer’4 (WHG) end, while the hunter-gatherersfrom Sweden4,8 (SHG) are intermediate. Against this background of dif-ferentiated European hunter-gatherers and homogeneous early farmers,multiple population turnovers transpired in all parts of Europe includedin our study. Middle Neolithic Europeans from Germany, Spain, Hungary,and Sweden from the period ,4,000–3,000 BC are intermediate betweenthe earlier farmers and the WHG, suggesting an increase of WHG ances-try throughout much of Europe. By contrast, in Russia, the later Yamnayasteppe herders of ,3,000 BC plot between the EHG and the present-dayNear East/Caucasus, suggesting a decrease of EHG ancestry during thesame time period. The Late Neolithic and Bronze Age samples fromGermany and Hungary2 are distinct from the preceding Middle Neo-lithic and plot between them and the Yamnaya. This pattern is alsoseen in ADMIXTURE analysis (Fig. 2b, Supplementary Informationsection 6), which implies that the Yamnaya have ancestry from popu-lations related to the Caucasus and South Asia that is largely absent in38 Early or Middle Neolithic farmers but present in all 25 Late Neo-lithic or Bronze Age individuals. This ancestry appears in CentralEurope for the first time in our series with the Corded Ware around2,500 BC (Supplementary Information section 6, Fig. 2b). The CordedWare shared elements of material culture with steppe groups such asthe Yamnaya although whether this reflects movements of people hasbeen contentious21. Our genetic data provide direct evidence of migra-tion and suggest that it was relatively sudden. The Corded Ware aregenetically closest to the Yamnaya ,2,600 km away, as inferred bothfrom PCA and ADMIXTURE (Fig. 2) and FST (0.011 6 0.002) (ExtendedData Table 3). If continuous gene flow from the east, rather than migra-tion, had occurred, we would expect successive cultures in Europeto become increasingly differentiated from the Middle Neolithic, but

0 50,000 100,000 150,000 200,000 250,000 300,000 350,000

Number of autosomal SNPs covered in 94 individuals

Maxim

um = 354,198

Minim

um = 22,869

Mean = 212,375

Median = 231,945

n = 69; this study (UDG treated)

n = 4; previous studies (UDG treated)

n = 21; previous studies (not UDG treated)

Time (ky BC) Group West Central East

43–22 Pleistocene hunter-gatherer

6–4.6 Holocene hunter-gatherer

6–5.5 Early Neolithic

4–3 Mid Neolithic

3.3–2.7 Late Copper Age (steppe)

2.5–2.2 Late Neolithic

2.2–1.6 Early Bronze Age1.1 Late Bronze Age

0.9 Iron Age

Ust Ishim (1)Kostenki14 (1)MA1 (1)

Karelia (1)Samara (1)Motala (7)

Sweden MHG (1)

Sweden NHG (3)

Loschbour (1)La Brana1 (1) Hungary HG (1)

Starcevo (1)

LBKT (1)Hungary EN (8)LBK (12)

Stuttgart (1)Els Trocs (5)

Iceman (1)La Mina (4)

Baalberge (3)Esperstedt (1)

Sweden MN (1)

Yamnaya (9)Hungary CA (1)

Corded Ware (4)Karsdorf (1)Bell Beaker (6)BenzigerodeHeimburg (3)Alberstedt (1)

Unetice (8)Hungary BA (2)

Halberstadt (1)

Hungary IA (1)

b

a

Figure 1 | Location and SNP coverage of samples included in this study.a, Geographic location and time-scale (central European chronology) of the 69newly analysed ancient individuals from this study (black outline) and 25 from

the literature for which shotgun sequencing data was available (no outline).b, Number of SNPs covered at least once in the analysis data set of 94individuals.

RESEARCH LETTER

G2015 Macmillan Publishers Limited. All rights reserved

2 0 8 | N A T U R E | V O L 5 2 2 | 1 1 J U N E 2 0 1 5

69newlysequencedindividuals25exis6ngindividuals

8000-3000BP20K-350KAutosomalSNPS

Haaketal.2015

Haaketal.2015By using a large number of outgroup populations we can fit the admix-

ture coefficients ai and estimate mixture proportions (SupplementaryInformation section 9, Extended Data Fig. 2). Using 15 outgroupsfrom Africa, Asia, Oceania and the Americas, we obtain good fits asassessed by a formal test (Supplementary Information section 10), andestimate that the Middle Neolithic populations of Germany and Spainhave ,18–34% more WHG-related ancestry than Early Neolithicpopulations and that the Late Neolithic and Early Bronze Age popula-tions of Germany have ,22–39% more EHG-related ancestry than theMiddle Neolithic ones (Supplementary Information section 9). If wemodel them as mixtures of Yamnaya-related and Middle Neolithicpopulations, the inferred degree of population turnover is doubled to48–80% (Supplementary Information sections 9 and 10).

To distinguish whether a Yamnaya or an EHG source fits the databetter, we added ancient samples as outgroups (Supplementary Infor-mation section 9). Adding any Early or Middle Neolithic farmer resultsin EHG-related genetic input into Late Neolithic populations being apoor fit to the data (Supplementary Information section 9); thus, LateNeolithic populations have ancestry that cannot be explained by a mix-ture of EHG and Middle Neolithic. When using Yamnaya instead ofEHG, however, we obtain a good fit (Supplementary Information sec-tions 9 and 10). These results can be explained if the new genetic materialthat arrived in Germany was a composite of two elements: EHG and atype of Near Eastern ancestry different from that which was introducedby early farmers (also suggested by PCA and ADMIXTURE; Fig. 2, Sup-plementary Information sections 5 and 6). We estimate that these twoelements each contributed about half the ancestry of the Yamnaya(Supplementary Information sections 6 and 9), explaining why thepopulation turnover inferred using Yamnaya as a source is about twiceas high compared to the undiluted EHG. The estimate of Yamnaya-related ancestry in the Corded Ware is consistent when using eitherpresent populations or ancient Europeans as outgroups (Supplemen-tary Information sections 9 and 10), and is 73.1 6 2.2% when both setsare combined (Supplementary Information section 10). The best pro-xies for ANE ancestry in Europe4 were initially Native Americans12,22,and then the Siberian MA1 (ref. 6), but both are geographically andtemporally too remote for what appears to be a recent migration intoEurope4. We can now add three new pieces to the puzzle of how ANEancestry was transmitted to Europe: first by the EHG, then the Yamnayaformed by mixture between EHG and a Near Eastern related popu-lation, and then the Corded Ware who were formed by a mixture of theYamnaya with Middle Neolithic Europeans. We caution that the sampledYamnaya individuals from Samara might not be directly ancestral toCorded Ware individuals from Germany. It is possible that a morewestern Yamnaya population, or an earlier (pre-Yamnaya) steppe popu-lation may have migrated into central Europe, and future work mayuncover more missing links in the chain of transmission of steppe ancestry.

By extending our model to a three-way mixture of WHG, Early Neolithicand Yamnaya, we estimate that the ancestry of the Corded Ware was79% Yamnaya-like, 4% WHG, and 17% Early Neolithic (Fig. 3). A smallcontribution of the first farmers is also consistent with uniparentallyinherited DNA: for example, mitochondrial DNA haplogroup N1a andY chromosome haplogroup G2a, common in early central Europeanfarmers14,23, almost disappear during the Late Neolithic and BronzeAge, when they are largely replaced by Y haplogroups R1a and R1b (Sup-plementary Information section 4) and mtDNA haplogroups I, T1, U2, U4,U5a, W, and subtypes of H14,23,24 (Supplementary Information section 2).The uniparental data not only confirm a link to the steppe populationsbut also suggest that both sexes participated in the migrations (Sup-plementary Information sections 2 and 4 and Extended Data Table 2).The magnitude of the population turnover that occurred becomes evenmore evident if one considers the fact that the steppe migrants may wellhave mixed with eastern European agriculturalists on their way to cen-tral Europe. Thus, we cannot exclude a scenario in which the CordedWare arriving in today’s Germany had no ancestry at all from localpopulations.

Our results support a view of European pre-history punctuated bytwo major migrations: first, the arrival of the first farmers during theEarly Neolithic from the Near East, and second, the arrival of Yamnayapastoralists during the Late Neolithic from the steppe. Our data furthershow that both migrations were followed by resurgences of the previousinhabitants: first, during the Middle Neolithic, when hunter-gathererancestry rose again after its Early Neolithic decline, and then betweenthe Late Neolithic and the present, when farmer and hunter-gathererancestry rose after its Late Neolithic decline. This second resurgencemust have started during the Late Neolithic/Bronze Age period itself,as the Bell Beaker and Unetice groups had reduced Yamnaya ancestrycompared to the earlier Corded Ware, and comparable levels to that insome present-day Europeans (Fig. 3). Today, Yamnaya related ances-try is lower in southern Europe and higher in northern Europe, and allEuropean populations can be modelled as a three-way mixture of WHG,Early Neolithic, and Yamnaya, whereas some outlier populations showevidence for additional admixture with populations from Siberia andthe Near East (Extended Data Fig. 3, Supplementary Information sec-tion 9). Further data are needed to determine whether the steppe ances-try arrived in southern Europe at the time of the Late Neolithic/BronzeAge, or is due to migrations in later times from northern Europe25,26.

Our results provide new data relevant to debates on the origin andexpansion of Indo-European languages in Europe (Supplementary Infor-mation section 11). Although the findings from ancient DNA are silenton the question of the languages spoken by preliterate populations,they do carry evidence about processes of migration which are invokedby theories on Indo-European language dispersals. Such theories makepredictions about movements of people to account for the spread of

Starcevo_ENLBKT_EN

Spain_ENBaalberge_MN

Esperstedt_MNSpain_MN

Stuttgart

HungaryGamba_CA

HungaryGamba_EN

Iceman

Sardinian

HungaryGamba_BA

AlbanianGreek

SpanishBergamo

BasqueTuscan

BulgarianSpanish_North

FrenchCroatian

French_SouthOrcadian

English

Unetice_EBA

UkrainianHungarian

Bell_Beaker_LN

BelarusianCzech

ScottishIcelandicEstonian

LithuanianNorwegian

Halberstadt_LBA

Alberstedt_LNBenzigerode_LN

Karsdorf_LNCorded_Ware_LN

Early Neolithic (LBK_EN)Western European hunter−gatherer (Loschbour)Yamnaya

0 0.2 0.4 0.6 0.8 1.0

Mod

ern

Anc

ient

Tim

e

Figure 3 | Admixture proportions. We estimate mixture proportionsusing a method that gives unbiased estimates even without an accuratemodel for the relationships between the test populations and the outgrouppopulations (Supplementary Information section 9). Population samplesare grouped according to chronology (ancient) and Yamnaya ancestry(present-day humans).

RESEARCH LETTER

G2015 Macmillan Publishers Limited. All rights reserved

2 1 0 | N A T U R E | V O L 5 2 2 | 1 1 J U N E 2 0 1 5

Admixturepropor6ons

Haaketal.2015

Extended Data Figure 4 | Geographic distribution of archaeologicalcultures and graphic illustration of proposed population movements /turnovers discussed in the main text. a, Proposed routes of migration by earlyfarmers into Europe ,9,00027000 years ago. b, Resurgence of hunter-gatherer

ancestry during the Middle Neolithic 7,00025,000 years ago. c, Arrival ofsteppe ancestry in central Europe during the Late Neolithic ,4,500 years ago.White arrows indicate the two possible scenarios of the arrival of Indo-European language groups. Symbols of samples are identical to those in Fig. 1.

LETTER RESEARCH

G2015 Macmillan Publishers Limited. All rights reserved

FarmerexpansionCardialWare(Mediterranean):6400-5500BCStarčevoculture(SEEurope):5500and4500BCLinearpokery(LBK)(CentralEurope):5500-4500BC

Mesolithichunter-gatherer/NeolithicFarmerFunnelBeaker(NorthernEurope):4300-2800BC

PastoralistexpansionYamnaya(Pon6csteppe):3,600–2,300BCCordedWare(CentralEurope):2,900–circa2,350BCBellBeaker(WesternEurope):c.2800–1800BC

Allentometal.2015

personhood13,14, rapidly stretching from Hungary to the Urals15. By2800 BC a new social and economic formation, variously namedCorded Ware, Single Grave or Battle Axe cultures developed in tem-perate Europe, possibly deriving from the Yamnaya background, andculturally replacing the remaining Neolithic farmers16,17 (Fig. 1). Inwestern and Central Asia, hunter-gatherers still dominated in EarlyBronze Age, except in the Altai Mountains and Minusinsk Basinwhere the Afanasievo culture existed with a close cultural affinity toYamnaya15 (Fig. 1). From the beginning of 2000 BC, a new class ofmaster artisans known as the Sintashta culture emerged in the Urals,building chariots, breeding and training horses (Fig. 1), and pro-ducing sophisticated new weapons18. These innovations quicklyspread across Europe and into Asia where they appeared to give riseto the Andronovo culture19,20 (Fig. 1). In the Late Bronze Age around1500 BC, the Andronovo culture was gradually replaced by theMezhovskaya, Karasuk, and Koryakova cultures21. It remains debatedif these major cultural shifts during the Bronze Age in Europe andAsia resulted from the migration of people or through cultural dif-fusion among settled groups15–17, and if the spread of the Indo-European languages was linked to these events or predates them15.

Archaeological samples and DNA retrievalGenomes obtained from ancient biological remains can provideinformation on past population histories that is not retrievable fromcontemporary individuals4,22. However, ancient genomic studies haveso far been restricted to single or a few individuals because of thedegraded nature of ancient DNA making sequencing costly and timeconsuming23. To overcome this, we increased the average output of

authentic endogenous DNA fourfold by: (1) targeting the outercementum layer in teeth rather than the inner dentine layer24,25,(2) adding a ‘pre-digestion’ step to remove surface contaminants24,26,and (3) developing a new binding buffer for ancient DNA extraction(Supplementary Information, section 3). This allowed us to obtainlow-coverage genome sequences (0.01–7.43 average depth, overallaverage equal to 0.73) of 101 Eurasian individuals spanning the entireBronze Age, including some Late Neolithic and Iron Age individuals(Fig. 1, Supplementary Information, sections 1 and 2). Our data setincludes 19 genomes, between 1.1–7.43 average depth, thereby doub-ling the number of existing Eurasian ancient genomes above 13coverage (ref. 27).

Bronze Age EuropeBy analysing our genomic data in relation to previously publishedancient and modern data (Supplementary Information, section 6),we find evidence for a genetically structured Europe during theBronze Age (Fig. 2; Extended Data Fig. 1; and Supplementary Figs 5and 6). Populations in northern and central Europe were composed ofa mixture of the earlier hunter-gatherer and Neolithic farmer10

groups, but received ‘Caucasian’ genetic input at the onset of theBronze Age (Fig. 2). This coincides with the archaeologically well-defined expansion of the Yamnaya culture from the Pontic-Caspiansteppe into Europe (Figs 1 and 2). This admixture event resulted in theformation of peoples of the Corded Ware and related cultures, assupported by negative ‘admixture’ f3 statistics when using Yamnayaas a source population (Extended Data Table 2, Supplementary Table12). Although European Late Neolithic and Bronze Age cultures such

OkunevoOkunevoKarasukKarasukAndronovoAndronovo

Early/mid secondEarly/mid second millennium BCmillennium BC

SintashtaSintashta

Corded WareCorded Ware

AfanasievoAfanasievo

Third millennium BCThird millennium BC

0 1,000 km

0 1,000 km

TocharianTocharian

YamnayaYamnaya

N

N

3400 BC 2600 BC 1800 BC 1000 BC 200 BC 600 AD

Iron AgeAfontova GoraNordic Late BAMezhovskayaKarasukLate BAAndronovoVatyaMiddle BASintashtaNordic LN-EBANordic LNMarosUneticeOkunevoBell BeakerStalingrad Q.Nordic MN BBAC, CWCYamnayaAfanasievoRemedello

Remedello, CA

Stalingrad quarry,CA-EBA

Yamnaya, CA-EBA

Afanasievo, EBA

Battle Axe andCorded Ware,CA-EBA

Nordic MN B

Bell Beaker, CA-EBA

Okunevo, EBA

Unetice, EBA

Maros, MBA

Sintashta, EBA

Nordic LN

Vatya, MBA

Nordic LN-EBA

Andronovo, BA

Middle BA

Karasuk, LBA

Mezhovskaya,LBA

Late BA

Nordic Late BA

Afontova Gora,LBA-IA

Velika Gruda,LBA-IA

Iron AgeN

0 1,000 km

Figure 1 | Distribution maps of ancient samples. Localities, culturalassociations, and approximate timeline of 101 sampled ancient individualsfrom Europe and Central Asia (left). Distribution of Early Bronze Age culturesYamnaya, Corded Ware, and Afanasievo with arrows showing the Yamnayaexpansions (top right). Middle and Late Bronze Age cultures Sintashta,Andronovo, Okunevo, and Karasuk with the eastward migration indicated

(bottom right). Black markers represent chariot burials (2000–1800 BC) withsimilar horse cheek pieces, as evidence of expanding cultures. Tocharian is thesecond-oldest branch of Indo-European languages, preserved in WesternChina. CA, Copper Age; MN, Middle Neolithic; LN, Late Neolithic; EBA, EarlyBronze Age; MBA, Middle Bronze Age; LBA, Late Bronze Age; IA, Iron Age;BAC, Battle Axe culture; CWC, Corded Ware culture.

1 6 8 | N A T U R E | V O L 5 2 2 | 1 1 J U N E 2 0 1 5

RESEARCH ARTICLE

G2015 Macmillan Publishers Limited. All rights reserved

SummaryofrecentaDNAevidence

Evidenceofselec6oninEurasians

Mathiesonetal.2015

230WestEurasianswholivedbetween6500and300BC.a,Loca6onscolour-codedbydate,witharandomjikeraddedforvisibility(8AfanasievoandAndronovosamplesliefurthereastandarenotshown).b,Principalcomponentanalysisof777modernWestEurasiansamples(grey),with221ancientsamplesprojectedontothefirsttwoprincipalcomponentaxesandlabeledbyculture.E/M/LN,Early/Middle/LateNeolithic;LBK,Linearbandkeramik;E/WHG,Eastern/Westernhunter-gatherer;EBA,EarlyBronzeAge;IA,IronAge;LNBA,LateNeolithicandBronzeAge.

Mathiesonetal.2015

0 0 M O N T H 2 0 1 5 | V O L 0 0 0 | N A T U R E | 3

ARTICLE RESEARCH

© 2015 Macmillan Publishers Limited. All rights reserved

samples into three groups based on which of these populations they clustered with most closely (Fig. 1b and Extended Data Table 1). We estimated mixture proportions for the present-day European ancestry populations and tested every SNP to evaluate whether its present-day frequencies were consistent with this model. We corrected for test statistic inflation by applying a genomic control correction analogous to that used to correct for population structure in genome-wide asso-ciation studies14. Of approximately one million non-monomorphic autosomal SNPs, the ~50,000 in the set of potentially functional SNPs were significantly more inconsistent with the model than neutral SNPs (Fig. 2), suggesting pervasive selection on polymorphisms of functional importance. Using a conservative significance threshold of P = 5.0 × 10−8, and a genomic control correction of 1.38, we iden-tified 12 loci that contained at least three SNPs achieving genome-wide significance within 1 Mb of the most associated SNP (Fig. 2, Extended Data Table 3, Extended Data Fig. 3 and Supplementary Data Table 3).

The strongest signal of selection is at the SNP (rs4988235) responsi-ble for lactase persistence in Europe15,16. Our data (Fig. 3) strengthens previous reports that an appreciable frequency of lactase persistence in Europe only dates to the last 4,000 years3,5,17. The allele’s earliest appearance in the dataset is in a central European Bell Beaker sample (individual I0112) dated to between 2450 and 2140 bc. Two other independent signals related to diet are located on chromosome 11 near FADS1 and DHCR7. FADS1 and FADS2 are involved in fatty acid metabolism, and variation at this locus is associated with plasma lipid and fatty acid concentration18. The selected allele of the most significant SNP (rs174546) is associated with decreased triglyceride levels18. This locus has experienced independent selection in non-Eu-ropean populations13,19,20 and is likely to be a critical component of adaptation to different diets. Variants at DHCR7 and NADSYN1 are associated with circulating vitamin D levels21 and the most associ-ated SNP in our analysis, rs7940244, is highly differentiated across closely related northern European populations22,23, suggesting selection related to variation in dietary or environmental sources of vitamin D.

Two signals have a potential link to coeliac disease. One occurs at the ergothioneine transporter SLC22A4 that is hypothesized to have expe-rienced a selective sweep to protect against ergothioneine deficiency in agricultural diets24. Common variants at this locus are associated with increased risk for ulcerative colitis, coeliac disease, and irritable bowel

disease and may have hitchhiked to high frequency as a result of this sweep24–26. However, the specific variant (rs1050152, L503F) that was thought to be the target did not reach high frequency until relatively recently (Extended Data Fig. 4). The signal at ATXN2/SH2B3—also associated with coeliac disease25—shows a similar pattern (Extended Data Fig. 4).

The second strongest signal in our analysis is at the derived allele of rs16891982 in SLC45A2, which contributes to light skin pigmentation and is almost fixed in present-day Europeans but occurred at much lower frequency in ancient populations. In con-trast, the derived allele of SLC24A5 that is the other major deter-minant of light skin pigmentation in modern Europe (and that is not significant in the genome-wide scan for selection) appears fixed in the Anatolian Neolithic, suggesting that its rapid increase in frequency to around 0.9 in Early Neolithic Europe was mostly due to migration (Extended Data Fig. 4). Another pigmenta-tion signal is at GRM5, where SNPs are associated with pigmen-tation possibly through a regulatory effect on nearby TYR27. We also find evidence of selection for the derived allele of rs12913832 at HERC2/OCA2, which is at 100% frequency in the European hunter- gatherers we analysed, and is the primary determinant of light eye colour in present-day Europeans28,29. In contrast to the other loci, the range of frequencies in modern populations is within that of ancient populations (Fig. 3). The frequency increases with higher latitude, suggesting a complex pattern of environmental selection.

The TLR1–TLR6–TLR10 gene cluster is a known target of selec-tion in Europe, possibly related to resistance to leprosy, tuberculosis or other mycobacteria30–32. There is also a strong signal of selection at the major histocompatibility complex (MHC) on chromosome 6. The strongest signal is at rs2269424 near the genes PPT2 and EGFL8, but there are at least six other apparently independent signals in the MHC (Extended Data Fig. 3); and the entire region is significantly more associated than the genome-wide average (residual inflation of 2.07 in the region on chromosome 6 between 29–34 Mb after genome-wide genomic control correction). This could be the result of multiple sweeps, balancing selection, or increased drift as a result of background selection reducing effective population size in this gene-rich region.

We find a surprising result in six Scandinavian hunter-gatherers (SHG) from Motala in Sweden. In three of six samples, we observe the haplotype carrying the derived allele of rs3827760 in the EDAR

Figure 2 | Genome-wide scan for selection. GC-corrected –log10 P value for each marker (Methods). The red dashed line represents a genome-wide significance level of 0.5 × 10−8. Genome-wide significant points filtered because there were fewer than two other genome-wide significant points within 1 Mb are shown in grey. Inset, quantile–quantile plots for corrected –log10 P values for different categories of potentially functional SNPs

(Methods). Truncated at − log10[P value] = 30. All curves are significantly different from neutral expectation. CMS, composite of multiple signals selection hits; HiDiff, highly differentiated between HapMap populations; Immune, immune-related; HLA, human leukocyte antigen type tag SNPs; eQTL, expression quantitative trait loci (see Methods).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 20 22

50

40

30

20

10

0

–log

10[P

val

ue]

Chromosome

30

25

20

15

10

5

0

0 1 2 3 4 5 6Expected –log10[P value]

Obs

erve

d –l

og10

[P v

alue

]

NeutralGWASCMSHiDiffImmuneHLAeQTL

LCT

TLR1-6-10

SLC

45A2

SLC

22A4 MH

CZK

SCAN3 DHCR7

FADS1-2

GRM5

ATX

N2

Chr

13:3

8.8

HER

C2

Mathiesonetal.2015

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 5

ARTICLERESEARCH

© 2015 Macmillan Publishers Limited. All rights reserved

gene (Extended Data Fig. 5), which affects tooth morphology and hair thickness33,34, has been the target of a selective sweep in East Asia35, and today is at high frequency in East Asians and Native Americans. The EDAR derived allele is largely absent in present-day Europe, except in Scandinavia, plausibly owing to Siberian move-ments into the region millennia after the date of the Motala samples. The SHG have no evidence of East Asian ancestry4,7, suggesting that the EDAR derived allele may not have originated in the main ances-tral population of East Asians as previously suggested35. A second surprise is that, unlike closely related WHGs, the Motala samples

have predominantly derived pigmentation alleles at SLC45A2 and SLC24A5.

Evidence of selection on heightWe also tested for selection on complex traits. The best-documented example of this process in humans is height, for which the differ-ences between northern and southern Europe have been driven by selection36. To test for this signal in our data, we used a statistic that tests whether trait-affecting alleles are both highly correlated and more differentiated, compared to randomly sampled alleles37. We predicted genetic heights for each population and applied the test to all populations together, as well as to pairs of populations (Fig. 4). Using 180 height-associated SNPs38 (restricted to 169 for which we successfully obtained genotypes from at least two individuals from each population), we detect a significant signal of directional selection on height (P = 0.002). Applying this to pairs of populations allows us to detect two independent signals. First, the Iberian Neolithic and Chalcolithic samples show selection for reduced height relative to both the Anatolian Neolithic (P = 0.042) and the central European Early and Middle Neolithic (P = 0.003). Second, we detect a signal for increased height in the steppe populations (P = 0.030 relative to the central European Early and Middle Neolithic). These results suggest that the modern South–North gradient in height across Europe is due to both increased steppe ancestry in northern popula-tions, and selection for decreased height in Early Neolithic migrants to southern Europe. We did not observe any other significant signals of polygenetic selection in five other complex traits we tested: body mass index39 (P = 0.20), waist-to-hip ratio40 (P = 0.51), type 2 dia-betes41 (P = 0.37), inflammatory bowel disease26 (P = 0.17) and lipid levels18 (P = 0.50).

Future studies of selection with ancient DNAOur results, which take advantage of the massive increase in sample size enabled by optimized techniques for sampling from the inner-ear regions of the petrous bone, as well as in-solution enrichment methods for targeted SNPs, show how ancient DNA can be used to perform a genome-wide scan for selection. Our results also directly document selection on loci related to pigmentation, diet and immu-nity, painting a picture of populations adapting to settled agricultural life at high latitudes. For most of the signals, allele frequencies of modern Europeans are outside the range of any ancient populations, indicating that phenotypically, Europeans of 4,000 years ago were different in important respects from Europeans today, despite having overall similar ancestry. An important direction for future research is to increase the sample size for European selection scans (Extended Data Fig. 6), and to apply this approach to regions beyond Europe and to other species.

−0.2

−0.1

0.0

0.1

0.2

0.3

0.4

Gen

etic

hei

ght (

GW

AS

effe

ct s

ize)

●

8,000 6,000 4,000 2,000 0

STP

HG

AN

CEMCLB

INC

CEU

IBS

Increasing height

Decreasing height

a

Steppe ancestryEarly farmersHunter-gatherers

Years before present

AN CEMCLB HG IN

CSTP IB

SCEU

CEU

IBS

STP

INC

HG

CLB

CEM

AN

−4

−3

−2

2

3

4

Zb

Figure 4 | Polygenic selection on height. a, Estimated genetic heights. Boxes show 0.05–0.95 posterior densities for population mean genetic height (Methods). Dots show the maximum likelihood point estimate. Arrows show major population relationships, dashed lines represent

ancestral populations. The symbols < and > label potentially independent selection events resulting in an increase or decrease in height. b, Z scores for the pairwise polygenic selection test. Positive if the column population is taller than the row population.

0.0

0.5

1.0

LCT

(rs49

8823

5)

0.0

0.5

1.0

FADS1−

2 (rs

1745

46)

0.0

0.5

1.0

SLC

45A2

(rs16

8919

82)

0.0

0.5

1.0

HER

C2/OCA2

(rs12

9138

32)

0.0

0.5

1.0

TLR1−

6−10

(rs4

8331

03)

Hunter-gatherer (HG)Early farmer (AN)Early farmer (CEM)Early farmer (INC)Steppe ancestry (CLB)Steppe ancestry (STP)Northwest Europe (CEU)Great Britain (GBR)Spain (IBS)Tuscan (TSI)

Figure 3 | Allele frequencies for five genome-wide significant signals of selection. Dots and solid lines show maximum likelihood frequency estimates and a 1.9-log-likelihood support interval for the derived allele frequency in each ancient population. Horizontal dashed lines show allele frequencies in the four modern 1000 Genomes populations. AN, Anatolian Neolithic; HG, hunter-gatherer; CEM, central European Early and Middle Neolithic; INC, Iberian Neolithic and Chalcolithic; CLB, central European Late Neolithic and Bronze Age; STP, steppe; CEU, Utah residents with northern and western European ancestry; IBS, Iberian population in Spain. The hunter-gatherer, early farmer and steppe ancestry classifications correspond approximately to the three populations used in the genome-wide scan with some differences (See Extended Data Table 1 for details).

Lactasepersistenceallele

Melaninsynthesis

Toll-likereceptor1

FakyAcidDesaturaseEyecolor

Mathiesonetal.2015

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 5

ARTICLERESEARCH

© 2015 Macmillan Publishers Limited. All rights reserved

gene (Extended Data Fig. 5), which affects tooth morphology and hair thickness33,34, has been the target of a selective sweep in East Asia35, and today is at high frequency in East Asians and Native Americans. The EDAR derived allele is largely absent in present-day Europe, except in Scandinavia, plausibly owing to Siberian move-ments into the region millennia after the date of the Motala samples. The SHG have no evidence of East Asian ancestry4,7, suggesting that the EDAR derived allele may not have originated in the main ances-tral population of East Asians as previously suggested35. A second surprise is that, unlike closely related WHGs, the Motala samples

have predominantly derived pigmentation alleles at SLC45A2 and SLC24A5.

Evidence of selection on heightWe also tested for selection on complex traits. The best-documented example of this process in humans is height, for which the differ-ences between northern and southern Europe have been driven by selection36. To test for this signal in our data, we used a statistic that tests whether trait-affecting alleles are both highly correlated and more differentiated, compared to randomly sampled alleles37. We predicted genetic heights for each population and applied the test to all populations together, as well as to pairs of populations (Fig. 4). Using 180 height-associated SNPs38 (restricted to 169 for which we successfully obtained genotypes from at least two individuals from each population), we detect a significant signal of directional selection on height (P = 0.002). Applying this to pairs of populations allows us to detect two independent signals. First, the Iberian Neolithic and Chalcolithic samples show selection for reduced height relative to both the Anatolian Neolithic (P = 0.042) and the central European Early and Middle Neolithic (P = 0.003). Second, we detect a signal for increased height in the steppe populations (P = 0.030 relative to the central European Early and Middle Neolithic). These results suggest that the modern South–North gradient in height across Europe is due to both increased steppe ancestry in northern popula-tions, and selection for decreased height in Early Neolithic migrants to southern Europe. We did not observe any other significant signals of polygenetic selection in five other complex traits we tested: body mass index39 (P = 0.20), waist-to-hip ratio40 (P = 0.51), type 2 dia-betes41 (P = 0.37), inflammatory bowel disease26 (P = 0.17) and lipid levels18 (P = 0.50).

Future studies of selection with ancient DNAOur results, which take advantage of the massive increase in sample size enabled by optimized techniques for sampling from the inner-ear regions of the petrous bone, as well as in-solution enrichment methods for targeted SNPs, show how ancient DNA can be used to perform a genome-wide scan for selection. Our results also directly document selection on loci related to pigmentation, diet and immu-nity, painting a picture of populations adapting to settled agricultural life at high latitudes. For most of the signals, allele frequencies of modern Europeans are outside the range of any ancient populations, indicating that phenotypically, Europeans of 4,000 years ago were different in important respects from Europeans today, despite having overall similar ancestry. An important direction for future research is to increase the sample size for European selection scans (Extended Data Fig. 6), and to apply this approach to regions beyond Europe and to other species.

−0.2

−0.1

0.0

0.1

0.2

0.3

0.4

Gen

etic

hei

ght (

GW

AS

effe

ct s

ize)

●

8,000 6,000 4,000 2,000 0

STP

HG

AN

CEMCLB

INC

CEU

IBS

Increasing height

Decreasing height

a

Steppe ancestryEarly farmersHunter-gatherers

Years before present

AN CEMCLB HG IN

CSTP IB

SCEU

CEU

IBS

STP

INC

HG

CLB

CEM

AN

−4

−3

−2

2

3

4

Zb

Figure 4 | Polygenic selection on height. a, Estimated genetic heights. Boxes show 0.05–0.95 posterior densities for population mean genetic height (Methods). Dots show the maximum likelihood point estimate. Arrows show major population relationships, dashed lines represent

ancestral populations. The symbols < and > label potentially independent selection events resulting in an increase or decrease in height. b, Z scores for the pairwise polygenic selection test. Positive if the column population is taller than the row population.

0.0

0.5

1.0

LCT

(rs49

8823

5)

0.0

0.5

1.0

FADS1−

2 (rs

1745

46)

0.0

0.5

1.0

SLC

45A2

(rs16

8919

82)

0.0

0.5

1.0

HER

C2/OCA2

(rs12

9138

32)

0.0

0.5

1.0

TLR1−

6−10

(rs4

8331

03)

Hunter-gatherer (HG)Early farmer (AN)Early farmer (CEM)Early farmer (INC)Steppe ancestry (CLB)Steppe ancestry (STP)Northwest Europe (CEU)Great Britain (GBR)Spain (IBS)Tuscan (TSI)

Figure 3 | Allele frequencies for five genome-wide significant signals of selection. Dots and solid lines show maximum likelihood frequency estimates and a 1.9-log-likelihood support interval for the derived allele frequency in each ancient population. Horizontal dashed lines show allele frequencies in the four modern 1000 Genomes populations. AN, Anatolian Neolithic; HG, hunter-gatherer; CEM, central European Early and Middle Neolithic; INC, Iberian Neolithic and Chalcolithic; CLB, central European Late Neolithic and Bronze Age; STP, steppe; CEU, Utah residents with northern and western European ancestry; IBS, Iberian population in Spain. The hunter-gatherer, early farmer and steppe ancestry classifications correspond approximately to the three populations used in the genome-wide scan with some differences (See Extended Data Table 1 for details).