ARTICLE doi:10.1038/nature13810 Genome sequence of a 45,000-year-old modern human from western Siberia QiaomeiFu1,2,HengLi3,4,PriyaMoorjani3,5,FloraJay6,SergeyM.Slepchenko7,AlekseiA.Bondarev8,PhilipL.F.Johnson9, AyinuerAximu-Petri2,KayPru¨fer2,CesaredeFilippo2,MatthiasMeyer2,NicolasZwyns10,11,DomingoC.Salazar-Garc´ıa10,12,13,14, YaroslavV.Kuzmin15,SusanG.Keates15,PavelA.Kosintsev16,DmitryI.Razhev7,MichaelP.Richards10,17,NikolaiV.Peristov18, MichaelLachmann2,19,KaterinaDouka20,ThomasF.G.Higham20,MontgomerySlatkin6,Jean-JacquesHublin10, DavidReich3,4,21,JanetKelso2,T.BenceViola2,10&SvantePa¨a¨bo2 Wepresentthehigh-qualitygenomesequenceofa 45,000-year-oldmodernhumanmalefromSiberia.Thisindividual derivesfromapopulationthatlivedbefore—orsimultaneouslywith—theseparationofthepopulationsinwesternand easternEurasiaandcarriesasimilaramountofNeanderthalancestryaspresent-dayEurasians.However,thegenomic segmentsofNeanderthalancestryaresubstantiallylongerthanthoseobservedinpresent-dayindividuals,indicating that Neanderthal gene flow into the ancestors of this individual occurred 7,000–13,000 years before he lived. We estimatean autosomalmutation rate of0.431029to0.631029per siteper year,aYchromosomal mutation rateof 0.731029to0.931029persiteperyearbasedontheadditionalsubstitutionsthathaveoccurredinpresent-daynon- Africanscomparedtothisgenome,andamitochondrialmutationrateof1.831028to3.231028persiteperyearbased ontheageofthebone. In2008,arelativelycompletelefthumanfemoraldiaphysiswasdiscov- yearsBP(46,880–43,210 calBPat 95.4%probability,Supplementary eredonthebanksoftheriverIrtysh(Fig.1a,c,d),nearthesettlement Informationsection1).TheUst’-Ishimindividualisthereforetheoldest ofUst’-IshiminwesternSiberia(OmskOblast,RussianFederation). directlyradiocarbon-datedmodernhumanoutsideAfricaandtheMid- Althoughtheexactlocalityisunclear,thefemurwaserodingoutofal- dleEast(Fig.1b).Carbonandnitrogenisotoperatiosindicatethatthe luvialdepositsontheleftbankoftheriver,northofUst’-Ishim.Here, dietoftheUst’-Ishimindividual(SupplementaryInformationsection4) LatePleistoceneandprobablyredepositedMiddlePleistocenefossils wasbasedonterrestrialC plantsandanimalsthatconsumedthem,but 3 arefoundinsandandgravellayersthatareabout50,000–30,000years alsothatanimportantpartofhisdietaryproteinmayhavecomefrom old(thatis,fromMarineOxygenIsotopeStage3). aquaticfoods,probablyfreshwaterfish,somethingthathasbeenob- servedinotherearlyUpperPalaeolithichumansfromEurope3. Morphology,datinganddiet Theproximalendoftheboneshowsalargeglutealbuttressandgluteal DNAretrievalandsequencing tuberosity,whilethemidshaftisdominatedbyamarkedlineaaspera, Ninesamplesofbetween41and130mgofbonematerialwereremoved resultinginateardrop-shapedcross-section(Fig.1e,f)(fordetails,see fromthedistalpartofthefemurandusedtoconstructDNAlibraries SupplementaryInformationsection3).Themorphologyoftheprox- usingaprotocoldesignedtofacilitatetheretrievalofshortanddamaged imalendoftheshaftissimilartoUpperPaleolithicmodernhumansand DNA4.ThepercentageofDNAfragmentsintheselibrariesthatcould distinctfromNeanderthals(SupplementaryTable3.1,Supplementary bemappedtothehumangenomevariedbetween1.8%and10.0%(Sup- Fig.3.2.),whiletheteardrop-shapedcrosssectionofthemidshaftissim- plementaryTable1.1).Fromtheextractcontainingthehighestpropor- ilartomostUpperPaleolithichumansandearlyanatomicallymodern tionofhumanDNA,eightfurtherlibrarieswereconstructed.Eachof humans1.Takentogether,thissuggeststhattheUst’-Ishimfemurderives theselibrarieswastreatedwithuracil-DNAglycosylaseandendonucle- fromamodernhuman. aseVIIItoremovedeaminatedcytosineresidues,andlibrarymolecules Twosamplesof890mgand450mgofthebonewereremovedonsep- withinsertsshorterthanapproximately35basepairs(bp)weredepleted arateoccasionsfordating.Collagenpreservationsatisfiedallcriteriafor bypreparativeacrylamidegelelectrophoresisbeforesequencingonthe dating2andafterultrafiltrationweobtainedagesof41,40061,300years IlluminaHiSeqplatform(SupplementaryInformationsection6).Intotal, beforepresent(BP)(OxA-25516)and41,40061,400BP(OxA-30190). 42-foldsequencecoverageofthe,1.86gigabases(Gb)oftheautosomal Thesetwodates,whencombinedandcorrectedforfluctuationsofatmo- genometowhichshortfragmentscanbeconfidentlymappedwasgen- spheric14Cthroughtime,correspondtoanageofabout45,000calibrated erated.ThecoverageoftheXandYchromosomeswasapproximately 1KeyLaboratoryofVertebrateEvolutionandHumanOriginsofChineseAcademyofSciences,IVPP,CAS,Beijing100044,China.2DepartmentofEvolutionaryGenetics,MaxPlanckInstituteforEvolutionary Anthropology,D-04103Leipzig,Germany.3BroadInstituteofMITandHarvard,Cambridge,Massachusetts02142,USA.4DepartmentofGenetics,HarvardMedicalSchool,Boston,Massachusetts02115, USA.5DepartmentofBiologicalSciences,ColumbiaUniversity,NewYork,NewYork10027,USA.6DepartmentofIntegrativeBiology,UniversityofCalifornia,Berkeley,California94720-3140,USA. 7InstituteforProblemsoftheDevelopmentoftheNorth,SiberianBranchoftheRussianAcademyofSciences,Tyumen625026,Russia.8ExpertCriminalisticsCenter,OmskDivisionoftheMinistryof InternalAffairs,Omsk644007,Russia.9DepartmentofBiology,EmoryUniversity,Atlanta,Georgia30322,USA.10DepartmentofHumanEvolution,MaxPlanckInstituteforEvolutionaryAnthropology, D-04103Leipzig,Germany.11DepartmentofAnthropology,UniversityofCalifornia,Davis,California95616,USA.12DepartmentofArchaeology,UniversityofCapeTown,CapeTown7701,SouthAfrica. 13DepartamentdePrehisto`riaiArqueologia,UniversitatdeVale`ncia,Valencia46010,Spain.14ResearchGrouponPlantFoodsinHomininDietaryEcology,Max-PlanckInstituteforEvolutionary Anthropology,D-04103Leipzig,Germany.15InstituteofGeologyandMineralogy,SiberianBranchoftheRussianAcademyofSciences,Novosibirsk630090,Russia.16InstituteofPlantandAnimalEcology, UralsBranchoftheRussianAcademyofSciences,Yekaterinburg620144,Russia.17LaboratoryofArchaeology,DepartmentofAnthropology,UniversityofBritishColumbia,Vancouver,BritishColumbia V6T1Z1,Canada.18SiberianCulturalCenter,Omsk644010,Russia.19SantaFeInstitute,SantaFe,NewMexico87501,USA.20OxfordRadiocarbonAcceleratorUnit,ResearchLaboratoryforArchaeology andtheHistoryofArt,UniversityofOxford,OxfordOX13QY,UK.21HowardHughesMedicalInstitute,HarvardMedicalSchool,Boston,Massachusetts02115,USA. 23 OCTOBER 2014 | VOL 514 | NATURE | 445 ©2014Macmillan Publishers Limited. All rights reserved RESEARCH ARTICLE a 60° E 70° E 80° E 90° E 100° E 110° E c d 60° N 60° N e 1 50° N 3 2 4 50° N 5 f 70° E 80° E 90° E 100° E b –35 H5 GI 12 H4 O 18RIP –40 G N –45 Europe Kent’s Cavern 4 (Model age) 5 0 Cavallo B (Model age) m Pestera cu Oase 1 (GrA-22810) m Kostenki 14 (OxA-X-2395-15) Kostenki 1 (OxA-15055) Siberia Ust’-Ishim (OxA-25516 & 30190) China Tianyuan Cave (BA-03222) 50,000 45,000 40,000 35,000 30,000 Calibrated date (cal BP) Figure1|Geographiclocation,morphologyanddating. a,MapofSiberia dated(OxCalv4.2.3(ref.33);r:5IntCal13atmosphericcurve34).H5:Heinrich5 withmajorarchaeologicalsites.Redtriangles:Neanderthalfossils;white event,H4:Heinrich4event,GI12:GreenlandInterstadial12.Foramore circlewithinared(Neanderthal)triangle:Denisovanfossils;bluesquare: extensivecomparisonseeSupplementaryInformationFig.2.1.c–f,TheUst’- InitialUpperPalaeolithicsites;yellowasterisk:Ust’-Ishim.1:Ust’-Ishim;2: Ishim1femur.c,Lateralview.d,Posteriorview.e,Cross-sectionatthe80 ChagyrskayaCave;3:OkladnikovCave;4:DenisovaCave;5:Kara-Bom. percentlevel.f,Cross-sectionatthemidshaft.Forotherviewssee b,RadiocarbonagesofearlymodernhumanfossilsinnorthernEurasiaandthe SupplementaryFig.3.1. NGRIPd18Opalaeotemperaturerecord.Specimensinlightgreyareindirectly halfthatoftheautosomes(,22-fold),indicatingthatthebonecomes branchleadingtotheUst’-IshimmtDNAislowerthanthenumbers fromamale.Alikelihoodmethodestimatedpresent-dayhumanmito- inferredtohaveoccurredonthebranchesleadingtorelatedpresent- chondrialDNA(mtDNA)contamination5to0.50%(95%confidence daymtDNAs(SupplementaryFig.8.1).Usingthisobservationandnine interval(CI)0.26–0.94%),whereasamethodthatusesthefrequencyof directlycarbon-datedancientmodernhumanmtDNAsascalibration non-consensusbasesinautosomalsequencesestimatedthecontam- points5,7inarelaxedmolecularclockmodel,weestimatetheageofthe inationtobelessthan0.13%(SupplementaryInformationsection7). Ust’-Ishimbonetobe,49,000yearsBP(95%highestposteriorden- Thus,lessthan1%ofthehomininDNAfragmentssequencedareesti- sity:31,000–66,000yearsBP),consistentwiththeradiocarbondate. matedtobeextraneoustothebone.Afterconsensusgenotypecalling, InaprincipalcomponentanalysisoftheUst’-Ishimautosomalge- suchlowlevelsofcontaminationwilltendtobeeliminated. nomealongwithgenotypingdatafrom922present-dayindividuals from53populations8(Fig.2a),theUst’-Ishimindividualclusterswith Populationrelationships non-AfricansratherthanAfricans.Whenonlynon-Africanpopula- About 7.7 positions per 10,000 are heterozygous in the Ust’-Ishim tionsareanalysed(Fig.2b),theUst’-Ishimindividualfallsclosetozero genome,whereasbetween9.6and10.5positionsareheterozygousin onthetwofirstprincipalcomponentaxes,suggestingthatitdoesnotshare present-dayAfricansand5.5and7.7inpresent-daynon-Africans(Sup- muchmoreancestrywithanyparticulargroupofpresent-dayhumans. plementaryInformationsection12).Thus,withrespecttogeneticdi- TodeterminehowtheUst’-Ishimgenomeisrelatedtothegenomesof versity,thepopulationtowhichtheUst’-Ishimindividualbelongedwas present-dayhumans,wetested,usingDstatistics8,whetheritsharesmore moresimilartopresent-dayEurasiansthantopresent-dayAfricans, derivedalleleswithonemodernhumanthanwithanothermodernhuman whichprobablyreflectstheout-of-Africabottlenecksharedbynon- usingpairsofhumangenomesfromdifferentpartsoftheworld(Fig.3). Africanpopulations.TheUst’-IshimmtDNAsequencefallsattheroot Basedongenotypingdatafor87Africanand108non-Africanindivi- ofalargegroupofrelatedmtDNAs(the‘Rhaplogroup’),whichoccurs duals(SupplementaryInformationsection11),theUst’-Ishimgenome todayacrossEurasia(SupplementaryInformationsection8).TheY sharesmorealleleswithnon-Africansthanwithsub-SaharanAfricans chromosomesequenceoftheUst’-Ishimindividualissimilarlyinferred (jZj541–89),consistentwiththeprincipalcomponentanalysis,mtDNA tobeancestraltoagroupofrelatedYchromosomes(haplogroupK(xLT)) andYchromosomeresults.Thus,theUst’-Ishimindividualrepresents thatoccursacrossEurasiatoday6(SupplementaryInformationsection9). apopulationderivedfrom,orrelatedto,thepopulationinvolvedinthe Asexpected,thenumberofmutationsinferredtohaveoccurredonthe dispersalofmodernhumansoutofAfrica.Amongthenon-Africans, 446 | NATURE | VOL 514 | 23 OCTOBER 2014 ©2014Macmillan Publishers Limited. All rights reserved ARTICLE RESEARCH a b 0.10 of variance 0.025 ECNEuaeosrnrottth prAa eAsl isaainad South Asia of variance 0.05 % South East Asia % genvector2 3.94 –00..002050 AOMUNmsocidtree_tdahIrlsin ecAhi aaEifmraicsat genvector2 1.79 –00..0050 Ei Sub-Saharan Africa Ei –0.10 –0.050 0.00 0.05 0.10 –0.025 0.000 0.025 0.050 Eigenvector1 6.44 % of variance Eigenvector1 5.26 % of variance Figure2|PrincipalComponents(PC)analysisexploringtherelationshipof usingEurasianindividualsandtheUst’-Ishimindividual.Thepercentagesof Ust’-Ishimtopresent-dayhumans. a,PCanalysisusing922present-day thetotalvarianceexplainedbyeacheigenvectoraregiven. individualsfrom53populationsandtheUst’-Ishimindividual.b,PCanalysis theUst’-Ishimgenomesharesmorederivedalleleswithpresent-day WestEurasianandEastEurasianpopulationsbefore—orsimultaneously peoplefromEastAsiathanwithpresent-dayEuropeans(jZj52.1–6.4). with—theirdivergencefromeachother.ThefindingthattheUst’-Ishim However,whenan,8,000-year-oldgenomefromwesternEurope(La individualisequallycloselyrelatedtopresent-dayAsiansandto8,000- Bran˜a)9ora24,000-year-oldgenomefromSiberia(Mal’ta1)10were to24,000-year-oldindividualsfromwesternEurasia,butnottopresent- analysed,thereisnoevidencethattheUst’-Ishimgenomesharesmore day Europeans, is compatible with the hypothesis that present-day derivedalleleswithpresent-dayEastAsiansthanwiththeseprehistoric Europeansderivesomeoftheirancestryfromapopulationthatdid individuals(jZj,2).ThissuggeststhatthepopulationtowhichtheUst’- notparticipateintheinitialdispersalsofmodernhumansintoEurope Ishimindividualbelongeddivergedfromtheancestorsofpresent-day andAsia11. Mutationrateestimates Y X Mbuti Onge Thehigh-qualityUst’-Ishimgenomesequence,incombinationwithits Mbuti Karitiana Mbuti Han radiocarbondate,allowsustogaugetherateofmutationsbyestimating Mbuti La Braña Mbuti Mal’ta thenumbersofmutationsthatare‘missing’intheUst’-Ishimindividual African YYoorruubbaa OKanrgiteiana Non-African relativetopresent-dayhumans.Thisresultsinamutationrateestimate YYoorruubbaa LHaa nBraña of0.4431029to0.6331029persiteperyearusingthehigh-coverage Yoruba Mal’ta genomesof14present-dayhumans.Achallengeforinferringthemuta- Dinka Onge Dinka Karitiana tionrateinthiswayisthatdifferencesinerrorratesamonggenome Dinka Han Dinka La Braña sequencescanconfoundtheinference(seediscussioninref.12).We Dinka Mal’ta thereforedevelopedanalternativeapproachthatleveragesthePairwise OOnnggee HKaarnitiana SequentiallyMarkovianCoalescent(PSMC),amethodwhichestimates OOnnggee MLaa Bl’traaña Other thedistributionofcoalescencetimesbetweenthetwochromosomes Other KKaarriittiiaannaa LHaa nBraña non-African acrossadiploidgenometoestimatepastchangesinpopulationsize13, non-African Karitiana Mal’ta andwhichislessinfluencedbydifferencesinerrorrates.WhentheUst’- Han La Braña Han Mal’ta Ishimgenomealongwith25present-dayhumangenomesareanalysed La Braña Mal’ta byPSMC,arecentreductioninpopulationsizesimilartothatseenfor French Onge 11present-daynon-AfricansisinferredfortheUst’-Ishimgenome.How- French Karitiana French Han ever,theapparentageofthissizereductionismorerecentthaninpresent- Present-day FFrreenncchh MLaa Bl’traaña Other dayhumans,consistentwiththeUst’-Ishimgenomebeingolder(Fig.4). European Sardinian Onge non-African Sardinian Karitiana Wethencomputethenumberofadditionalsubstitutionsthatareneeded Sardinian Han Sardinian La Braña tobestfittheUst’-IshimPSMCcurvetothoseofothernon-African Sardinian Mal’ta genomes.Assumingthatthiscorrespondstothenumberofmutations −0.10 0.00 0.10 0.20 0.30 0.40 thathaveaccumulatedoveraround45,000years,weestimateamuta- tionrateof0.4331029persiteperyear(95%CI0.3831029to0.493 D(X Y; Ust’−Ishim Chimpanzee) 1029)thatisconsistentacrossallnon-Africangenomesregardlessof Figure3|StatisticstestingwhethertheUst’-Ishimgenomesharesmore theircoverage(SupplementaryInformationsection14).Thisoverall derivedalleleswithoneortheotheroftwomodernhumangenomes(X,Y). rate,aswellastherelativeratesinferredfordifferentmutationalclasses WecomputedDstatisticsoftheformD(X,Y,Ust’-Ishim,Chimpanzee)using (transversions,non-CpGtransitions,andCpGtransitions),issimilarto asubsetofthegenome-wideSNParraydatafromtheAffymetrixHuman therateobservedfordenovoestimatesfromhumanpedigrees(,0.53 Originsarrayandrestrictingtheanalysistotransversions.Errorbars 1029persiteperyear14,15)andtothedirectestimateofbranchshortening correspondtothreestandarderrors.RedbarsindicatethattheDstatisticis significantlydifferentfrom0(|Z|.2),suchthattheUst’-Ishimgenome (SupplementaryInformationsection10).Asdiscussedelsewhere14,16,17, sharesmorederivedalleleswiththegenomeontheright(X)thantheleft(Y). theseratesareslowerthanthoseestimatedusingcalibrationsbasedon Ancientgenomesaregiveninitalics. thefossilrecordandthussuggestolderdatesforthesplitsofmodern 23 OCTOBER 2014 | VOL 514 | NATURE | 447 ©2014Macmillan Publishers Limited. All rights reserved RESEARCH ARTICLE 10 kya 100 kya 1 Mya 10 Mya oftheDNAsegmentscontributedbyNeanderthalstopresent-daynon- 310) 3 San Africans21.Thus,theUst’-Ishimindividualcouldpre-datetheNeander- × Mbuti μunits of 4N e 2 .25 MSaanYFrddoDreeirniHunnnibdacaknhnaaa tNo(hSfeuaNalpnaepdadlnmeemdritexhertnatuhltraaaerln.yaFddrItomnhmfeioxrtUthmusertaee’t-xiiInotsenhtnhitsmeeoUcfgtsseihotn’an-oIrmsi1nh6egi)ms,owsfiidnemederiislivatviiredmdutoaaatllpelterotlehebsseeebnp2etr.t-3owdp6eaoeyrn0te.it3aoh%snet ed in 1.5 KaritDinaai Asians(1.7–2.1%)andpresent-dayEuropeans(1.6–1.8%).Thus,admix- e (scal 1 UUssPtt’’a--IIpssuhhaiimmn tcuornetrwaistth,wNeefaanildteordtheatelscthaandyaclorenatdriybuotcicounrfrreodmbyD4en5,i0so0v0aynesa,rasltahgoou.gInh n siz corrected suchacontributionexistsinpresent-daypeoplenotonlyinOceania22,23, atio 0.5 buttoalesserextentalsoinmainlandeastAsia12,24(Supplementary ul Pop 0 Informationsection17). 10–5 10–4 10–3 10–2 TheDNAsegmentscontributedbyNeanderthalstotheUst’-Ishim Time (scaled in units of 2μT) individualareexpectedtobelongerthansuchsegmentsinpresent- Figure4|Inferredpopulationsizechangesovertime. ‘Time’onthexaxis daypeopleastheUst’-Ishimindividuallivedcloserintimetowhenthe referstothepairwiseper-sitesequencedivergence.Ifweerroneouslyassume admixtureoccurred,sotherewaslesstimeforthesegmentstobefrag- thatUst’-Ishimlivedtoday,itsinferredpopulationsizehistoryincludesan mentedbyrecombination.Totestifthisisindeedthecase,weidentified out-of-Africa-likepopulationbottleneckthatismorerecentthanthatseenin putativeNeanderthalDNAsegmentsintheUst’-Ishimandpresent- present-daynon-Africans(redboldcurve).ByshiftingtheUst’-Ishimcurve daygenomesbasedonderivedallelessharedwiththeNeanderthalge- toalignwiththoseinpresent-daynon-Africans(blueboldcurve),and nomeatpositionswhereAfricansarefixedforancestralalleles.Figure5 assumingthatthenumberofmutationsnecessarytodothiscorrespondsto 45,000years,weestimatetheautosomalmutationratetobe0.3831029to showsthatfragmentsofputativeNeanderthaloriginintheUst’-Ishim 0.4931029persiteperyear.Thetimesindicatedonthetopofthefigure individualaresubstantiallylongerthanthoseinpresent-dayhumans. arebasedonthismutationrate. WeusethecovarianceinsuchderivedallelesofputativeNeanderthal originacrosstheUst’-Ishimgenometoinferthatmeanfragmentsizes humanandarchaicpopulations.Wecaution,however,thatratesmay intheUst’-Ishimgenomeareintheorderof,1.8–4.2timeslongerthan havechangedovertimeandmaydifferbetweenhumanpopulations. inpresent-daygenomesandthattheNeanderthalgeneflowoccurred However,weexpect thismutationrateestimatetoapply atleast to 232–430generationsbeforetheUst’-Ishimindividuallived(Supplemen- non-Africanpopulationsoverthepast45,000years. taryInformationsection18;Fig.6).Underthesimplifyingassumption Wealsoestimatedaphylogenyrelatingthenon-recombiningpartof thatthegeneflowoccurredasasingleevent,andassumingageneration theUst’-IshimYchromosometothoseof23present-daymales.Using timeof29years16,25,weestimatethattheadmixturebetweentheances- thisphylogeny,wemeasuredthenumberof‘missing’mutationsinthe torsoftheUst’-IshimindividualandNeanderthalsoccurredapproxi- Ust’-IshimYchromosomallineagerelativetothemostcloselyrelated mately50,000to60,000yearsBP,whichisclosetothetimeofthemajor present-dayYchromosomeanalysed.Thisresultsinanestimateofthe expansionofmodernhumansoutofAfricaandtheMiddleEast.How- Ychromosomemutationrateof0.7631029persiteperyear(95%CI ever,wealsonotethatthepresenceofsomelongerfragments(Fig.5)may 0.6731029to0.8631029)(SupplementaryInformationsection9), indicatethatadditionaladmixtureoccurredevenlater.Nevertheless, significantlyhigherthantheautosomalmutationrate,consistentwith theseresultssuggestthatthebulkoftheNeanderthalcontributionto mutationratesinmalesbeinghigherthaninfemales18–20.Finally,using present-daypeopleoutsideAfricadoesnotgobacktomixturebetween theradiocarbondateoftheUst’-IshimfemurtogetherwiththemtDNAs Neanderthalsandtheanatomicallymodernhumanswholivedinthe of311present-dayhumans,weestimatedthemutationrateofthecom- MiddleEastatearliertimes;forexample,themodernhumanswhose pletemtDNAtobe2.5331028substitutionspersiteperyear(95% remainshavebeenfoundatSkhulandQafzeh26,27. highestposteriordensity:1.7631028to3.2331028)(Supplementary Information section 8) for mtDNA, in agreement with a previous AnInitialUpperPaleolithicindividual? study5. AcommonmodelforthemodernhumancolonizationofAsia23,28assumes thatanearlycoastalmigrationgaverisetothepresent-daypeopleof Neanderthaladmixture Oceania,whilealatermorenorthernmigrationgaverisetoEuropeans ThetimeofadmixturebetweenmodernhumansandNeanderthalshas andmainlandAsians.Thefactthatthe45,000-year-oldindividualfrom previouslybeenestimatedto37,000–86,000yearsBPbasedonthesize Siberia is not more closely related to the Onge from the Andaman Ust’-Ishim French_A French_B Sardinian_A Sardinian_B Han_A Han_B Dai_A Dai_B Karitiana_A Karitiana_B Mixe_B Papuan_A Papuan_B Australian_B1 Australian_B2 Figure5|RegionsofNeanderthalancestryonchromosome12inthe Neanderthalgenomecarriesthederivedallele.Homozygousancestralalleles Ust’-Ishimindividualandfifteenpresent-daynon-Africans. Theanalysisis areblack,heterozygousderivedallelesyellow,andhomozygousderivedalleles basedonSNPswhereAfricangenomescarrytheancestralalleleandthe blue. 448 | NATURE | VOL 514 | 23 OCTOBER 2014 ©2014Macmillan Publishers Limited. All rights reserved ARTICLE RESEARCH 13. Li,H.&Durbin,R.Inferenceofhumanpopulationhistoryfromindividual 0.10 Ust’−Ishim whole-genomesequences.Nature475,493–496(2011). French 14. Scally,A.&Durbin,R.Revisingthehumanmutationrate:implicationsfor Sardinian understandinghumanevolution.NatureRev.Genet.13,745–753(2012). ent 0.08 Han 15. Kong,A.etal.Rateofdenovomutationsandtheimportanceoffather’sageto fici Dai diseaserisk.Nature488,471–475(2012). ef 16. Langergraber,K.E.etal.Generationtimesinwildchimpanzeesandgorillas o c suggestearlierdivergencetimesingreatapeandhumanevolution.Proc.Natl ce 0.06 Acad.Sci.USA109,15716–15721(2012). an 17. Pru¨fer,K.etal.Thebonobogenomecomparedwiththechimpanzeeandhuman ari genomes.Nature486,527–531(2012). ov 0.04 18. Xue,Y.etal.HumanYchromosomebase-substitutionmutationratemeasuredby c d directsequencinginadeep-rootingpedigree.Curr.Biol.19,1453–1457(2009). hte 19. Kuroki,Y.etal.ComparativeanalysisofchimpanzeeandhumanYchromosomes eig 0.02 unveilscomplexevolutionarypathway.NatureGenet.38,158–167(2006). W 20. Wang,J.,Fan,H.C.,Behr,B.&Quake,S.R.Genome-widesingle-cellanalysisof recombinationactivityanddenovomutationratesinhumansperm.Cell150, 402–412(2012). 0.00 21. Sankararaman,S.,Patterson,N.,Li,H.,Pa¨a¨bo,S.&Reich,D.Thedateof interbreedingbetweenNeandertalsandmodernhumans.PLoSGenet.8, 0.0 0.2 0.4 0.6 0.8 1.0 e1002947(2012). Genetic distance (cM) 22. Reich,D.etal.GenetichistoryofanarchaichominingroupfromDenisovaCavein Siberia.Nature468,1053–1060(2010). Figure6|DatingtheNeandertaladmixtureinUst’-Ishimandpresent-day 23. Reich,D.etal.Denisovaadmixtureandthefirstmodernhumandispersalsinto non-Africans. Exponentiallyfittedcurvesshowingthedecayofpairwise SoutheastAsiaandOceania.Am.J.Hum.Genet.89,516–528(2011). covarianceforvariablepositionswhereAfricanscarryancestralallelesandthe 24. Skoglund,P.&Jakobsson,M.ArchaichumanancestryinEastAsia.Proc.NatlAcad. Neanderthalgenomecarriesderivedalleles. Sci.USA108,18301–18306(2011). 25. Fenner,J.N.Cross-culturalestimationofthehumangenerationintervalforusein genetics-basedpopulationdivergencestudies.Am.J.Phys.Anthropol.128, Islands(putativedescendantsofanearlycoastalmigration)thanheis 415–423(2005). topresent-dayEastAsiansorNativeAmericans(putativedescendants 26. McCown,T.D.&Keith,A.TheStoneAgeofMountCarmelVol.2(Clarendon,Oxford, 1939). ofanorthernmigration)(Fig.3)showsthatatleastoneothergroupto 27. Vandermeersch,B.LesHommesFossilesdeQafzeh(Israel)319(E´ditionsduCNRS, whichtheancestorsoftheUst’-Ishimindividualbelongedcolonized 1981). Asiabefore45,000yearsago.Interestingly,theUst’-Ishimindividual 28. Rasmussen,M.etal.AnAboriginalAustraliangenomerevealsseparatehuman probablylivedduringawarmperiod(GreenlandInterstadial12)that dispersalsintoAsia.Science334,94–98(2011). 29. Hublin,J.J.TheearliestmodernhumancolonizationofEurope.Proc.NatlAcad.Sci. hasbeenproposedtobeatimeofexpansionofmodernhumansinto USA109,13471–13472(2012). Europe29,30.However,thelatterhypothesisisbasedonlyontheappear- 30. Mu¨ller,U.C.etal.TheroleofclimateinthespreadofmodernhumansintoEurope. anceoftheso-called‘InitialUpperPaleolithic’industries(Supplemen- Quat.Sci.Rev.30,273–279(2011). 31. Goebel,T.A.,Derevianko,A.P.&Petrin,V.T.DatingtheMiddletoUpperPaleolithic taryInformationsection5),andnotontheidentificationofmodern transitionatKara-Bom.Curr.Anthropol.34,452–458(1993). humanremains31,32.ItispossiblethattheUst’-Ishimindividualwasasso- 32. Kuhn,S.L.&Zwyns,N.RethinkingtheinitialUpperPaleolithic.Quat.Int.http:// ciatedwiththeAsianvariantofInitialUpperPaleolithicindustry,doc- dx.doi.org/10.1016/j.quaint.2014.05.040(2014). 33. BronkRamsey,C.,Scott,M.&vanderPlicht,H.Calibrationforarchaeologicaland umentedatsitessuchasKara-BomintheAltaiMountainsatabout environmentalterrestrialsamplesinthetimerange26–50kacalBP.Radiocarbon 47,000yearsBP.Thisindividualwouldthenrepresentanearlymodern 55,2021–2027(2013). humanradiationintoEuropeandCentralAsiathatmayhavefailedto 34. Reimer,P.J.etal.IntCal13andMarine13radiocarbonagecalibrationcurves leavedescendantsamongpresent-daypopulations29. 0–50,000YearscalBP.Radiocarbon55,1869–1887(2009). SupplementaryInformationisavailableintheonlineversionofthepaper. OnlineContentMethods,alongwithanyadditionalExtendedDatadisplayitems andSourceData,areavailableintheonlineversionofthepaper;referencesunique AcknowledgementsWearegratefultoP.Gunz,M.Kircher,A.I.Krivoshapkin,P.Nigst, tothesesectionsappearonlyintheonlinepaper. M.Ongyerth,N.Patterson,G.Renaud,U.Stenzel,M.StonekingandS.Talamofor valuableinput,commentsandhelp;T.PfistererandH.Temmingfortechnical Received15May;accepted29August2014. assistance.Q.F.isfundedinpartbytheChineseAcademyofSciences(XDA05130202) andtheMinistryofScienceandTechnologyofChina(2007FY110200);P.A.K.byUrals Branch,RussianAcademyofSciences(12-C-4-1014)andY.V.K.bytheRussian 1. Trinkaus,E.&Ruff,C.B.Diaphysealcross-sectionalgeometryofNearEastern FoundationforBasicSciences(12-06-00045);F.J.andM.S.bytheNationalInstitutes MiddlePaleolithichumans:thefemur.J.Archaeol.Sci.26,409–424(1999). ofHealthoftheUSA(R01-GM40282);P.J.bytheNIH(K99-GM104158);and 2. Brock,F.etal.Reliabilityofnitrogencontent(%N)andcarbon:nitrogenatomic T.F.G.H.byERCadvancedgrant324139.D.R.isaHowardHughesMedicalInstitute ratios(C:N)asindicatorsofcollagenpreservationsuitableforradiocarbondating. InvestigatorandsupportedbytheNationalScienceFoundation(1032255)andthe Radiocarbon54,879–886(2012). NIH(GM100233).MajorfundingforthisworkwasprovidedbythePresidential 3. Richards,M.P.&Trinkaus,E.OutofAfrica:modernhumanoriginsspecialfeature: InnovationFundoftheMaxPlanckSociety. isotopicevidenceforthedietsofEuropeanNeanderthalsandearlymodern humans.Proc.NatlAcad.Sci.USA106,16034–16039(2009). AuthorContributionsQ.F.,S.M.S.,A.A.B.,Y.V.K.,J.K.,T.B.V.andS.P.designedthe 4. Meyer,M.etal.Ahigh-coveragegenomesequencefromanarchaicDenisovan research.A.A.P.andQ.F.performedtheexperiments;Q.F.,H.L.,P.M.,F.J.,P.L.F.J.,K.P., individual.Science338,222–226(2012). C.d.F.,M.M.,M.L.,M.S.,D.R.,J.K.andS.P.analysedgeneticdata;K.D.andT.F.G.H. 5. Fu,Q.etal.Arevisedtimescaleforhumanevolutionbasedonancient performed14Cdating;D.C.S.-G.andM.P.R.analysedstableisotopedata;N.V.P.,P.A.K. mitochondrialgenomes.Curr.Biol.23,553–559(2013). andD.I.R.contributedsamplesanddata;S.M.S.,A.A.B.,N.Z.,Y.V.K.,S.G.K.,J.-J.H.and 6. TheYChromosomeConsortiumAnomenclaturesystemforthetreeofhuman T.B.V.analysedarchaeologicalandanthropologicaldata;Q.F.,J.K.,T.B.V.andS.P.wrote Y-chromosomalbinaryhaplogroups.GenomeRes.12,339–348(2002). andeditedthemanuscriptwithinputfromallauthors. 7. Shapiro,B.etal.ABayesianphylogeneticmethodtoestimateunknownsequence ages.Mol.Biol.Evol.28,879–887(2011). AuthorInformationAllsequencedatahavebeensubmittedtotheEuropean 8. Patterson,N.etal.Ancientadmixtureinhumanhistory.Genetics192,1065–1093 NucleotideArchive(ENA)andareavailableunderthefollowingUst’-Ishim (2012). accessionnumber:PRJEB6622.Thedatafromthe25present-dayhuman 9. Olalde,I.etal.Derivedimmuneandancestralpigmentationallelesina7,000-year- genomesareavailablefrom(http://www.simonsfoundation.org/life-sciences/ oldMesolithicEuropean.Nature507,225–228(2014). simons-genome-diversity-project/)andfrom(http://cdna.eva.mpg.de/neandertal/ 10. Raghavan,M.etal.UpperPalaeolithicSiberiangenomerevealsdualancestryof altai/).Reprintsandpermissionsinformationisavailableatwww.nature.com/reprints. NativeAmericans.Nature505,87–91(2014). Theauthorsdeclarenocompetingfinancialinterests.Readersarewelcometo 11. Lazaridis,I.etal.Ancienthumangenomessuggestthreeancestralpopulationsfor commentontheonlineversionofthepaper.Correspondenceandrequestsfor present-dayEuropeans.Nature513,409–413(2014). materialsshouldbeaddressedtoQ.F.([email protected]),D.R. 12. Pru¨fer,K.etal.ThecompletegenomesequenceofaNeanderthalfromtheAltai ([email protected]),J.K.([email protected])orT.B.V. Mountains.Nature505,43–49(2014). ([email protected]). 23 OCTOBER 2014 | VOL 514 | NATURE | 449 ©2014Macmillan Publishers Limited. All rights reserved RESEARCH ARTICLE METHODS genomicregionsthatfallwithinthe95%coveragedistribution(SupplementaryIn- formationsection7)andwhereatleast99%ofoverlapping35merscoveringapos- AllsequencingwasperformedontheIlluminaHiSeq2000andbase-callingwas itionmapuniquely,allowingonemismatch. carriedoutusingIbis1.1.69(ref.35).Readsweremergedandremainingadaptor sequencestrimmedbeforebeingalignedtotheHumanreferencegenome(GRCh37/ 35. Kircher,M.,Stenzel,U.&Kelso,J.ImprovedbasecallingfortheIlluminaGenome 1000Genomes)usingBWA(version0.5.10)36.GATKversion1.3(v1.3-14-g348f2b) Analyzerusingmachinelearningstrategies.GenomeBiol.10,R83(2009). wasusedtoproducegenotypecallsforeachsite.Weexcludedfromanalysistan- 36. Li,H.&Durbin,R.FastandaccurateshortreadalignmentwithBurrows–Wheeler demrepeatsandregionsofthegenomethatarenotunique.Weconsideredonly transform.Bioinformatics25,1754–1760(2009). ©2014Macmillan Publishers Limited. All rights reserved