Genetic history of an archaic hominin group from Denisova Cave in Siberia


previous item | next item

Citation
Genetic history of an archaic hominin group from Denisova Cave in Siberia

Material Information

Title:
Genetic history of an archaic hominin group from Denisova Cave in Siberia
Series Title:
Nature
Creator:
Reich, David
Green, Richard E.
Kircher, Martin
Krause, Johannes
Patterson, Nick
Durand, Eric Y.
Viola, Bence
Briggs, Adrian W.
Stenzel, Udo
Johnson, Philip L. F.
Maricic, Tomislav
Good, Jeffrey M.
Marques-Bonet, Tomas
Alkan, Can
Fu, Qiaomei
Mallick, Swapan
Li, Heng
Meyer, Matthias
Eichler, Evan E.
Stoneking, Mark
Richards, Michael
Talamo, Sahra
Shunkov, Michael V.
Derevianko, Anatoli P.
Hublin, Jean-Jacques
Kelso, Janet
Slatkin, Montgomery
Pääbo, Svante
Language:
English

Subjects

Subjects / Keywords:
Gentics Of Hominin Group ( local )
Hominins ( local )
Hominin ( local )
Homin Group From Desinova Cave ( local )
Genetic Of Archaic Hominin Group ( local )
Archaic Hominin Group ( local )
Neanderthals ( local )
Desinovans ( local )
Melanesians ( local )
Genre:
serial ( sobekcm )

Notes

Abstract:
Using DNA extracted from a finger bone found in Denisova Cave in southern Siberia, we have sequenced the genome of an archaic hominin to about 1.9-fold coverage. This individual is from a group that shares a common origin with Neanderthals. This population was not involved in the putative gene flow from Neanderthals into Eurasians; however, the data suggest that it contributed 4–6% of its genetic material to the genomes of present-day Melanesians. We designate this hominin population ‘Denisovans’ and suggest that it may have been widespread in Asia during the Late Pleistocene epoch. A tooth found in Denisova Cave carries a mitochondrial genome highly similar to that of the finger bone. This tooth shares no derived morphological features with Neanderthals or modern humans, further indicating that Denisovans have an evolutionary history distinct from Neanderthals and modern humans.
Original Version:
Nature, Vol. 468.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
This item is licensed with the Creative Commons Attribution Non-Commercial Share Alike License. This license lets others remix, tweak, and build upon this work non-commercially, as long as they credit the author and license their new creations under the identical terms.
Resource Identifier:
K26-00023 ( USFLDC: LOCAL DOI )
k26.23 ( USFLDC: LOCAL Handle )

USFLDC Membership

Aggregations:
University of South Florida
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

ARTICLEdoi:10.1038/nature09710GenetichistoryofanarchaichominingroupfromDenisovaCaveinSiberiaDavidReich1,2*,RichardE.Green3,4*,MartinKircher3*,JohannesKrause3,5*,NickPatterson2*,EricY.Durand6*,BenceViola3,7*,AdrianW.Briggs1,3,UdoStenzel3,PhilipL.F.Johnson8,TomislavMaricic3,JeffreyM.Good9,TomasMarques-Bonet10,11,CanAlkan10,QiaomeiFu3,12,SwapanMallick1,2,HengLi2,MatthiasMeyer3,EvanE.Eichler10,MarkStoneking3,MichaelRichards7,13,SahraTalamo7,MichaelV.Shunkov14,AnatoliP.Derevianko14,Jean-JacquesHublin7,JanetKelso3,MontgomerySlatkin6&SvantePa¨a¨bo3UsingDNAextractedfromafingerbonefoundinDenisovaCaveinsouthernSiberia,wehavesequencedthegenomeofanarchaichominintoabout1.9-foldcoverage.ThisindividualisfromagroupthatsharesacommonoriginwithNeanderthals.ThispopulationwasnotinvolvedintheputativegeneflowfromNeanderthalsintoEurasians;however,thedatasuggestthatitcontributed4…6%ofitsgeneticmaterialtothegenomesofpresent-dayMelanesians.WedesignatethishomininpopulationDenisovansandsuggestthatitmayhavebeenwidespreadinAsiaduringtheLatePleistoceneepoch.AtoothfoundinDenisovaCavecarriesamitochondrialgenomehighlysimilartothatofthefingerbone.ThistoothsharesnoderivedmorphologicalfeatureswithNeanderthalsormodernhumans,furtherindicatingthatDenisovanshaveanevolutionaryhistorydistinctfromNeanderthalsandmodernhumans.Lessthan200,000yearsago,anatomicallymodernhumans(thatis,humanswithskeletonssimilartothoseofpresent-dayhumans)appearedinAfrica.Atthattime,aswellaslaterwhenmodernhumansappearedinEurasia,otherarchaichomininswerealreadypresentinEurasia.InEuropeandwesternAsia,homininsdefinedasNeanderthalsonthebasisoftheirskeletalmorphologylivedfromatleast230,000yearsagobeforedisappearingfromthefossilrecordabout30,000yearsago1.IneasternAsia,noconsensusexistsaboutwhichgroupswerepresent.Forexample,inChina,somehaveemphasizedmorphologicalaffinitiesbetweenNeanderthalsandthespecimenofMaba2,orbetweenHomoheidelbergensisandtheDaliskull3.However,othersclassifythesespecimensasearlyHomosapiens4.Inaddition,untilatleast17,000yearsago,Homofloresiensis,ashort-staturedhomininthatseemstorepresentanearlydivergencefromthelineageleadingtopresent-dayhumans5…7,waspresentontheislandofFloresinIndonesiaandpossiblyelsewhere.DNAsequencesretrievedfromhomininremainsofferanapproachcomplementarytomorphologyforunderstandinghomininrelation-ships.ForNeanderthals,thenucleargenomewasrecentlydeterminedtoabout1.3-foldcoverage8.ThisrevealedthatNeanderthalDNAsequencesandthoseofpresent-dayhumanssharecommonancestorsonaverageabout800,000yearsagoandthatthepopulationsplitofNeanderthalandmodernhumanancestorsoccurred270,000…440,000yearsago.ItalsoshowedthatNeanderthalssharedmoregeneticvariantswithpresent-dayhumansinEurasiathanwithpre-sent-dayhumansinsub-SaharanAfrica,indicatingthatgeneflowfromNeanderthalsintotheancestorsofnon-Africansoccurredtoanextentthat1…4%ofthegenomesofpeopleoutsideAfricaarederivedfromNeanderthals8.Inaddition,tenpartialandsixcompletemitochondrial(mt)DNAsequenceshavebeendeterminedfromNeanderthals9…17.ThishasshownthatallNeanderthalsstudiedsofarshareacommonmtDNAancestorontheorderof100,000yearsago10,andinturn,shareacommonancestorwiththemtDNAsofpresent-dayhumansabout500,000yearsago10,18,19(asexpected,thisisolderthantheNeanderthal…modernhumanpopulationsplittimeof270,000…440,000yearsagoestimatedfromthenucleargenome8).OneofthesemtDNAsequenceshasalsoshownthathomininscarryingmtDNAstypicalofNeanderthalswerepresentasfareastastheAltaiMountainsinsouthernSiberia13.In2008,thedistalmanualphalanxofajuvenilehomininwasexca-vatedatDenisovaCave.ThissiteislocatedintheAltaiMountainsinsouthernSiberia,andisareferencesitefortheMiddletoUpperPalaeolithicoftheregionwheresystematicexcavationsoverthepast25yearshaveuncoveredculturallayersindicatingthathumanoccu-pationatthesitestartedupto280,000yearsago20.Thephalanxwasfoundinlayer11,whichhasbeendatedto50,000to30,000yearsago.ThislayercontainsmicrobladesandbodyornamentsofpolishedstonetypicaloftheUpperPalaeolithicindustrygenerallythoughttobeassociatedwithmodernhumans,butalsostonetoolsthataremorecharacteristicoftheearlierMiddlePalaeolithic,suchasside-scrapersandLevalloisblanks21…23.Recently,weusedaDNAcaptureapproach10incombinationwithhigh-throughputsequencingtodetermineacompletemtDNAgenomefromtheDenisovaphalanx.Surprisingly,thismtDNAdivergedfromthecommonlineageleadingtomodernhumanandNeanderthalmtDNAsaboutonemillionyearsago19,thatis,abouttwiceasfarbackintimeasthedivergencebetweenNeanderthalandmodernhumanmtDNAs.However,mtDNAismaternallyinheritedasasingleunit*Theseauthorscontributedequallytothiswork. 1DepartmentofGenetics,HarvardMedicalSchool,Boston,Massachusetts02115,USA.2BroadInstituteofMITandHarvard,Cambridge,Massachusetts02142,USA.3DepartmentofEvolutionaryGenetics,MaxPlanckInstituteforEvolutionaryAnthropology,Leipzig04103,Germany.4DepartmentofBiomolecularEngineering,UniversityofCalifornia,SantaCruz95064,USA.5Institutfu¨rNaturwissenschaftlicheArcha¨ologie,UniversityofTu¨bingen,Tu¨bingen72070,Germany.6DepartmentofIntegrativeBiology,UniversityofCalifornia,Berkeley,California94720,USA.7DepartmentofHumanEvolution,MaxPlanckInstituteforEvolutionaryAnthropology,Leipzig04103,Germany.8DepartmentofBiology,EmoryUniversity,Atlanta,Georgia30322,USA.9DivisionofBiologicalSciences,UniversityofMontana,Missoula,Montana59812,USA.10HowardHughesMedicalInstitute,DepartmentofGenomeSciences,UniversityofWashington,Seattle,Washington98195,USA.11InstituteofEvolutionaryBiology(UPF-CSIC),08003Barcelona,Spain.12CAS-MPSJointLaboratoryforHumanEvolutionandArcheometry,InstituteofVertebratePaleontologyandPaleoanthropologyofChineseAcademyofSciences,Beijing100044,China.13DepartmentofAnthropology,UniversityofBritishColumbia,Vancouver,BritishColumbiaV6T1Z1,Canada.14PalaeolithicDepartment,InstituteofArchaeology&Ethnography,RussianAcademyofSciences,SiberianBranch,Novosibirsk630090,Russia.23/30DECEMBER2010|VOL468|NATURE|1053Macmillan Publishers Limited. All rights reserved©2010

PAGE 2

withoutrecombination,andthereforeissubjecttochanceeventssuchasgeneticdrift,aswellasgeneflowandpositiveselection.Incontrast,thenucleargenomecomprisestensofthousandsofunlinked,mostlyneutrallyevolvingloci.Thisallowsforanalysesofgeneticrelationshipsthatarerobusttothestochasticityofgeneticdrift,andaremuchlessaffectedbypositiveselection.ToclarifytherelationshipoftheDenisovaindividualtootherhominingroups,wehavethereforesequencedtheDenisovanucleargenomeandanalyseditsgenomicrelationshipstoNeanderthalsandpresent-dayhumans.Wehavealsoattemptedtoclarifythechronologyofhomininoccupationofthecaveandhaveidentifiedatoothfromthisgroupofhomininsamongmaterialexca-vatedinDenisovaCave.DNAsequencedeterminationTheentireinternalportionofthephalanxsamplewasusedforDNAextractioninourclean-roomfacility,whereprocedurestominimizecontaminationfrompresent-dayhumanDNAarerigorouslyimple-mented24,25(SupplementaryInformationsection1).TheDNAwastreatedwithtwoenzymes:uracil-DNA-glycosylase,whichremovesuracilresiduesfromDNAtoleaveabasicsites26,andendonucleaseVIII,whichcutsDNAatthe59and39sidesofabasicsites.SubsequentincubationwithT4polynucleotidekinaseandT4DNApolymerasewasusedtogenerate59-phosphorylatedbluntendsthatareamenabletoadaptorligation.BecausethegreatmajorityofuracilresiduesoccurclosetotheendsofancientDNAmolecules,thisprocedureleadstoonlyamoderatereductioninaveragelengthofthemoleculesinthelibrary,butaseveral-foldreductioninuracil-derivednucleotidemisincorporation27.Twoindependentsequencinglibraries(SL3003andSL3004)werecreatedfromtheDNA,usingamodifiedIlluminaprotocol28whereapolymerasechainreaction(PCR)isusedtoadda7-nucleotideindex(inthiscase59-GTCGACT-39)tothelibrarymolecules.Thisindexensuresthatthelibrariesarenotcontaminatedbyothersequencinglibrarieswhentheyaretakenoutofthecleanroomtobesequenced29.ThelibrariesweresequencedontheIlluminaGenomeAnalyserIIxplatformfor101cyclesfromeachendofthemoleculesandanaddi-tional7cyclesfordeterminationoftheindexuntilalmosteveryuniquesequenceinthelibrarieshadbeenseenmultipletimes,thatis,almosteveryclonepresentinthelibrarieshasbeensequenced(SupplementaryInformationsection1).Baseswerecalledusingthemachine-learningalgorithmIbis30andanoverlapofatleast11baseswasrequiredforpaired-endreadstobefusedtofull-molecule-sizeDNAsequencesthatwerefurtheranalysed.Thisresultsinagreatlyreducederrorrate27,althoughitremovesthefewmoleculesthatareabove191nucleotidesinlengthfromanalysis(,0.1%inSL3003and,0.2%inSL3004).SequencesweremappedusingtheprogramBWA31tothehuman(hg18/NCBI36)andthechimpanzee(panTro2/CGSC2.1)genomesaswellastotheinferredancestralgenomeofthesespecies(fromthesix-wayEnredo-Pecan-Ortheusalignment)32.PCRduplicateswereidentifiedandusedtofurtherincreasesequenceaccuracybycallingconsensussequences.Atotalof82,227,320sequencesmappeduniquely(mappingquality$30)tothehumangenome,yieldingabout5.2gigabasesofDNAsequences(1.9-foldgenomiccoverage),and72,304,848sequencesmappeduniquelytothechimpanzeegenome.WhenthesubstitutionsinferredtohaveoccurredontheDenisovaandthepresent-dayhumanlineageswerecompared,therelativenumbersofdifferentclassesofnucleotidesubstitutionsareremarkablysimilar,andtheexcessnumberofcandidatesubstitutionsontheDenisovalineagerelativetothepresent-dayhumanlineageisonly1.7-fold(SupplementaryFig.2.2andSupplementaryTable2.4).ThisreflectsanimprovementinerrorrateovertheNeanderthalgenomebyoveranorderofmagnitude8andismainlyduetotheenzymaticremovalofuracilresiduesfromtheDenisovaDNA27.WeestimatethatmosterrorsintheDenisovaDNAsequencesareduetolowgenomiccoverageratherthantoanyfeaturestypicalofancientDNA.HumanDNAcontaminationestimatesAlthoughrigorousmeasurestopreventcontaminationoftheexperi-mentsbyDNAfrompresent-dayhumanswereimplementedatalllaboratorysteps,itisimpossibletocompletelypreventcontaminationbecausebonesamplesaswellasreagentsmaybecontaminatedbeforetheyentertheclean-roomfacility.Toestimatethelevelsofcontami-nationinthesequencesproducedweusedthreeapproaches(Sup-plementaryInformationsection3).First,weestimatedthelevelofmtDNAcontaminationusing276sequencepositionswheretheDenisovamtDNAdiffersfrom.99%ofpresent-dayhumanmtDNAs.ForlibrarySL3003,weobserved7,433uniquesequencesthatcoveredsuchpositionsand7,421wereoftheDenisovatype.ForlibrarySL3004thecorrespondingnumberswere5,042and5,036,indicatingthatthemtDNAcontaminationinthelibrariesisontheorderof0.2%(95%confidenceinterval(CI):0.1…0.3%)and0.1%(CI:0.1…0.3%),respectively.Second,weidentifiedsequencesthatareuniquetotheYchro-mosome8.Iftheindividualfromwhomthephalanxderivesisfemale,thenumberofsuchsequencesrepresentstheextentofmaleDNAcontamination.WefoundzeroandthreesuchYchromosomalsequencesinthetwolibraries,respectively,whereas1,449and696areexpectediftheindividualismale.Thus,thebonederivesfromafemaleandmaleDNAcontaminationinthetwolibrariesisontheorderof0.00%(CI:0.00…0.25%)and0.43%(CI:0.09…1.26%),respectively.Third,toestimatetheextentofnuclearDNAcontaminationweusedonelibrarytoidentifypositionswheretheDenisovaindividualcarriesanancestral,thatis,chimpanzee-like,sequencevariantthatamongpresent-dayhumansisderivedandnotknowntovary.Wethenexaminedsequencesthatmapatthesepositionsintheotherlibraryanddeterminediftheycarrytheancestralsequenceorthederivedsequence.Observationofaderivedsequenceinthesecondlibrarycouldbeduetooneofthreepossibilities:thattheDNAfrag-mentinquestioncomesfrompresent-dayhumancontamination;thattheDenisovaindividualisheterozygousatthepositioninques-tion;orthattherehasbeenasequencingerror.Weimplementedamaximumlikelihoodmethodthatusesthenumberofindependentobservationsofancestralandderivedstatesacrosspositionstoco-estimatecontaminationalongwithheterozygosityandsequencingerrorasnuisanceparameters(SupplementaryInformationsection3).Fromthisanalysis,bothlibrariesareinferredtohavecontaminationratesoflessthan1%.AncestralfeaturesandduplicationsTheDenisovadraftgenomesequenceallowsfeaturesthatareancestralintheDenisovagenomeandderivedinpresent-dayhumanstobeidentified.Wepreviouslydescribedasetof10.5millionsinglenucleo-tidedifferencesandabouthalfamillioninsertion/deletions(indels)inferredtobeduetochangesthatoccurredonthehumanlineagesincethesplitfromthecommonancestorwiththechimpanzee8.Ofthese,4,267,431(40.5%)singlenucleotidedifferencesand105,372(22.0%)indelsarecoveredbytheDenisovasequences.Weidentified129inferredaminosubstitutionsand14indelsinthecodingsequencesofgeneswheretheDenisovaindividualcarriestheancestralallelesatpositionswherepresent-dayhumanscarryderivedallelesandarenotknowntovary(SupplementaryInformationsection4).Wealsoiden-tified90suchsitesin59untranslatedregions(UTRs),392in39UTRs,twoinmicroRNAgenesand104inhumanacceleratedregions.WhenwecomparedtheDenisovaandNeanderthalgenomeswefoundthattheycarrythesameassignedstateatsinglenucleotidedifferencesin87.9%oftheancestralpositionsand97.7%ofthederivedpositions.Theresultsforindelsaresimilar:87.6%forancestralstatesand98.6%forthederivedstates(SupplementaryTable4.3).WeanalysedthesegmentalduplicationcontentoftheDenisovagenomebydetectingregionswithanexcessreaddepth(Supplemen-taryInformationsection5).Inathree-waycomparisonofDenisova,Neanderthalandpresent-dayhumangenomes,wefoundanexcessof RESEARCH ARTICLE 1054|NATURE|VOL468|23/30DECEMBER2010Macmillan Publishers Limited. All rights reserved©2010

PAGE 3

privateDenisovaduplications(2.27megabases(Mb))comparedwithduplicationsthatwereprivateinNeanderthals(0.60Mb)orpresent-dayhumans(1.32Mb).Theseregionswereidentifiedbasedonsig-naturesofbothexcessreaddepthandincreasedsequencedivergence,makingthemunlikelytobeartefacts.WealsoidentifiedtworegionswheretheduplicationarchitectureofDenisovaismoresimilartothatofchimpanzeethantothatofeitherNeanderthalsorpresent-dayhumans,includingtwochromosomalregionsassociatedwithneuro-logicaldiseaseinhumans:spinalmuscularatrophyon5q13(includ-ingSMN2,oneofthemostrecentgeneduplicationsinthehumanlineage)andneuropsychiatricdiseaseon16p12.1.RelationshiptoNeanderthalsandmodernhumansAfundamentalquestioniswhethertheDenisovaindividualisanout-grouptoNeanderthalsandmodernhumans,asthemtDNAsuggests19,whetheritisasistergrouptoNeanderthalsortomodernhumans,orwhetheritfallswithintherangeofvariationofeitherofthesetwogroups.WeaddressedthisbyestimatingthedivergencebetweentheDenisovaandthehumangenomereferencesequenceasafractionofthedivergencebetweenpresent-dayhumansandthecommonancestorsharedwiththechimpanzee.Todothis,wescoredthefrequencywithwhichtheDenisovagenomecarriesthehumanversusthechimpanzeestateatpositionswherethehumanandchimpanzeereferencegenomesdiffer;assumingconstantevolutionaryrates(SupplementaryInforma-tionsection2).WerestrictedthisanalysistothepartsofthehumanreferencegenomethatareofAfricanancestry33asgeneflowfromNeanderthalstonon-Africans8couldotherwisecomplicatetheseana-lyses.TheDenisovagenomedivergedfromthereferencehumangenome11.7%(CI:11.4…12.0%)ofthewaybackalongthelineagetothehuman…chimpanzeeancestor.FortheVindijaNeanderthal,thedivergenceis12.2%(CI:11.9…12.5%).Thus,whereasthedivergenceoftheDenisovamtDNAtopresent-dayhumanmtDNAsisabouttwiceasdeepasthatofNeanderthalmtDNA19,theaveragedivergenceoftheDenisovanucleargenomefrompresent-dayhumansissimilartothatofNeanderthals.ApossibleexplanationforthesimilardivergenceoftheDenisovaindividualandNeanderthalsfrompresent-dayAfricansisthattheybothdescendfromacommonancestralpopulationthatseparatedearlierfromancestorsofpresent-dayhumans.SuchascenariowouldpredictacloserrelationshipbetweentheDenisovaindividualandNeanderthalsthanbetweeneitherofthemandpresent-dayhumans.Totestthisprediction,weestimatedthedivergencebetweenpairsofsevenancientandmoderngenomes(Denisova,Neanderthals,French,Han,Papuan,YorubaandSan),usinganapproachwherewecorrectforerrorratesineachgenomebasedontheassumptionthateachhasthesamenumberoftruedifferencesfromchimpanzee(SupplementaryInformationsection6).TheaveragedivergencebetweenDenisovaandVindijaNeanderthalsisestimatedtobe9.84%ofthewaytothechimpanzee…humanancestor;thatis,lessthantheaverage12.38%divergenceofbothfrompresent-dayAfricans.Assuming6.5millionyearsforhuman…chimpanzeedivergence,thisimpliesthatDNAsequencesofNeanderthalsandtheDenisovaindividualdivergedonaverage640,000yearsago,andfrompresent-dayAfricans804,000yearsago.ToanalysefurthertherelationshipoftheDenisovaindividualandNeanderthals,wealignedDenisova,NeanderthalandYorubasequencestothechimpanzeegenome,pickedasinglesequenceatrandomtorep-resenteachgroup,andexaminedsiteswheretwocopiesofaderivedandonecopyofanancestralallelewereobserved.Sequencingerrorsareexpectedtomakeanegligiblecontributionatsuchsites.ThenumberofsiteswheretheDenisovaindividualandNeanderthalclustertotheexclu-sionoftheYorubaandchimpanzeeis46,362,comparedwithanaverageof22,012sitesfortheothertwopossiblepatterns(YorubaandDenisova,orYorubaandNeanderthal).ThisexcessofsiteswhereDenisovaandNeanderthalclustersupportstheviewthattheDenisovaindividualandNeanderthalsshareacommonhistorysinceseparatingfromtheancestorsofmodernhumans(SupplementaryInformationsection6).ANeanderthal-specificbottleneckThefactthattheDenisovanucleargenomeonaveragesharesamorerecentcommonancestorwithNeanderthalthanwithpresent-dayhumansraisesthequestionofwhethertheoverallDNAsequencedivergenceoftheDenisovaindividualfallsinsidethegroupmorpho-logicallyandgeographicallydefinedasNeanderthals,orifitrepre-sentsasistergrouptoNeanderthals.Toinvestigatethisquestion,wetookadvantageofthefactthatinadditiontothethreeindividualsfromVindijaCave,Croatia,fromwhichmostoftheNeanderthalgenomesequenceswereproduced,wehavedeterminednuclearDNAsequencesfromthreefurtherNeanderthalindividualsfromRussia,SpainandGermany8.Ofthese,the60,000…70,000-year-oldskeletonofaNeanderthalchildfoundinMezmaiskayaCave,Russia,isbotholdestandgeographicallyclosesttotheDenisovaindividual.Usingthe56MbofautosomalDNAsequencesdeterminedfromthisspecimen8,weestimatethattheDNAsequencedivergencebetweentheVindijaandMezmaiskayaNeanderthalscorrespondstoadateof140,000633,000yearsago(SupplementaryInformationsection6)(Fig.1).Thisremarkablylowdivergence„whichisaboutone-thirdoftheclosestpairofpresent-dayhumansthatweanalysed„isinagreementwiththeobservationthatdiversityamongNeanderthalmtDNAsislowrelativetopresent-dayhumans10andindicatesthattheVindijaandMezmaiskayaNeanderthalsdescendfromacommonancestralpopulationthatexperiencedadrasticbottlenecksincesepa-ratingfromtheancestorsoftheDenisovaindividual.TounderstandfurtherthebottleneckinthehistoryofVindijaandMezmaiskayaNeanderthals,weexaminedfour-wayalignmentsoftheVindijaNeanderthalgenomesequence,theMezmaiskayaNeanderthal,theDenisovaindividualandthechimpanzeegenome.Attransversionsubstitutionswheretwocopiesofthederivedallelesareobserved,wedetect924substitutionsthatclustertheVindijaandMezmaiskayaNeanderthals,80thatclusterVindijaandDenisova,and81thatclusterMezmaiskayaandDenisova.Thiscorrespondstoatleasta65%probabilitythattheDNAsequencesintheNeanderthalsshareacommonancestormorerecentlythantheirsplitfromtheancestoroftheDenisovaindividual(SupplementaryInformationsection7).Itismuchhigherthanthe15…20%probabilityassociatedwiththeOutofAfricabottleneckcommontopresent-daynon-Africans34.IfwereplacetheMezmaiskayaNeanderthalinthisanalysiswithaNeanderthalfromElSidron,Spain,orfromFeldhofer,Germany,resultsarequalitativelysimilaralthoughnumbersaresmaller(SupplementaryInformationsection7).Thus,weconcludethatlateNeanderthalsacrossabroadgeographicalrangehaveapopulationhistorydistinctfromthatoftheDenisovaindividualinthattheyshareastrongpopulationbottlenecknotexperiencedbytheancestorsoftheDenisovaindividual.Wecall DenisovaMezmaiskaya 1Vindija 33.26Vindija 33.25Vindija 33.16SanHanPapuanFrenchYoruba Figure1|Aneighbour-joiningtreebasedonpairwiseautosomalDNAsequencedivergencesforfiveancientandfivepresent-dayhominins.Vindija33.16,Vindija33.25andVindija33.26refertothecataloguenumbersoftheNeanderthalbones. ARTICLE RESEARCH23/30DECEMBER2010|VOL468|NATURE|1055Macmillan Publishers Limited. All rights reserved©2010

PAGE 4

thegrouptowhichthisindividualbelongedDenisovansinanalogytoNeanderthals,asDenisovansaredescribedforthefirsttimebasedonmoleculardatafromDenisovaCavejustasNeanderthalswerefirstdescribedbasedonskeletalremainsretrievedintheNeanderValleyinGermany.NoDenisovangeneflowintoallEurasiansWehavepreviouslyshownthatVindijaNeanderthalssharemorederivedalleleswithnon-AfricansthanwithAfricans,consistentwithNeanderthalscontributing1…4%ofthegenomesofpresent-dayhumansacrossEurasia8.ToinvestigatetheextenttowhichtheDenisovaindividualsharesthispattern,weexaminedalignmentsofsetsoffourgenomes,eachconsistingofanAfrican(YorubaorSan),aEurasian(FrenchorHan),anarchaichominin(NeanderthalorDenisovan)andthechimpanzee.Werandomlysampledoneallelefromeachofthethreehominins,andcountedalltransversiondifferencesbetweentheAfricanandtheEurasianwherethearchaicindividualcarriesthederivedallele(theDstatisticsofref.8).NeanderthalsmatchtheFrenchgenomeonaverage4.660.7%moreoftenthantheymatchtheYorubagenome(Table1).AlthoughtheDenisovaindividualalsomatchestheFrenchmorethantheYorubagenome,thisskewissignifi-cantlylessstrongat1.860.5%.TheestimatesofDstatisticswerequantitativelyconsistent(withintwostandarddeviations)forallotherchoicesofEurasianandAfricanpopulations(Table1).ThesefindingsindicatethatthearchaiccomponentoftheEurasiangenepoolislesscloselyrelatedtotheDenisovaindividualthantoNeanderthals.Wealsoexamined13genomicregionsthatwepreviouslyidentifiedascandidatesforacontributionofarchaicgeneticmaterialintonon-Africans,basedontheirdeepergeneticdivergencesinnon-AfricansthaninAfricans8.UsingtagSNPsthatareinformativeaboutwhetherahaplotypeisfromthelineageuniquetonon-Africans,wefindthattheDenisovaindividualmatchesthedeeplydivergednon-Africanhaplotypein6cases,whereasNeanderthalsdosoin11cases(Su-pplementaryInformationsection7).Thus,bothNeanderthalsandDenisovansaremorerelatedthanwouldbeexpectedbychancetothesegenomicsegments,butthesignalinDenisovansisweaker.TheseanalysesindicatethatNeanderthalsaremorecloselyrelatedthanDenisovanstothepopulationthatcontributedtothegenepooloftheancestorsofpresent-dayEurasians.ThefactthatEurasianssharesomeadditionalaffinitywiththeDenisovaindividualrelativetoAfricansiscompatiblewithascenarioinwhichDenisovanssharedsomeoftheirhistorywithNeanderthalsbeforethegeneflowfromNeanderthalsintomodernhumansoccurred.DenisovangeneflowintotheancestorsofMelanesiansAlthoughtheDenisovaindividualderivesfromapopulationthatwasnotdirectlyinvolvedinthegeneflowfromNeanderthalstoEurasians,itispossiblethatDenisovansadmixedwiththeancestorsofpresent-daypeopleinsomepartsoftheOldWorld.Toinvestigatethis,weanalysedtherelationshipoftheDenisovagenometothegenomesof938present-dayhumansfrom53populationsthathavebeengeno-typedat642,690singlenucleotidepolymorphisms(SNPs)35.We Table1|Sharingofderivedallelesbetweenpresent-dayandarchaichomininsSampleH1SampleH2SourceofdataforH1andH2D(H1,H2,Neanderthal,chimpanzee)D(H1,H2,Denisova,chimpanzee)nBABAnABBAD(%)s.e.(%)Z-scorenBABAnABBAD(%)s.e.(%)Z-scoreEurasian/Eurasian*FrenchHanRef.817,21417,60221.10.821.427,25027,2650.00.60.0KaritianaSardinianThisstudy1,1161,0851.42.10.71,5591,62722.11.821.2KaritianaCambodianThisstudy1,6831,70720.71.820.42,3712,46021.81.521.2KaritianaMongolianThisstudy1,1281,19522.92.221.31,7651,7420.71.80.4SardinianCambodianThisstudy2,5922,67021.51.521.03,9353,9250.11.20.1SardinianMongolianThisstudy1,9662,02721.51.620.93,0363,05720.31.320.3CambodianMongolianThisstudy2,8112,8040.11.40.14,4424,3421.11.21.0African/African*SanYorubaRef.823,69023,85520.30.620.639,04239,0190.00.50.1Melanesian/Melanesian*Papuan2BougainvilleThisstudy3,3513,2841.01.30.85,3195,1401.71.11.5Eurasian/African*FrenchSanRef.825,24222,9824.70.67.6{39,83838,4951.70.53.4{FrenchYorubaRef.821,79419,8904.60.76.9{34,26233,0781.80.53.6{HanSanRef.825,08122,4705.50.68.5{38,81537,4391.80.53.4{HanYorubaRef.821,74119,4125.70.77.9{33,18232,1841.50.52.8KaritianaMbutiThisstudy1,5771,4733.41.91.82,3682,3600.21.50.1SardinianMbutiThisstudy2,5622,4003.31.52.24,0283,7843.11.22.6CambodianMbutiThisstudy4,2353,6417.51.26.5{6,3295,8503.91.04.0{MongolianMbutiThisstudy3,0772,7655.31.43.9{4,5144,5050.11.10.1Eurasian/Melanesian*FrenchPapuan1Ref.815,52315,54820.10.820.123,50925,47024.00.725.7{HanPapuan1Ref.815,05914,6771.30.91.522,26224,19824.20.725.8{KaritianaPapuan2Thisstudy1,5221,65824.31.922.22,2012,64129.11.625.8{KaritianaBougainvilleThisstudy1,5771,71724.31.822.42,2292,67129.01.525.9{SardinianPapuan2Thisstudy2,4472,64723.91.522.63,7144,15025.51.224.5{SardinianBougainvilleThisstudy2,5312,76224.41.523.03,8774,33625.61.124.9{CambodianPapuan2Thisstudy3,7133,89122.31.321.85,4576,27226.91.126.5{CambodianBougainvilleThisstudy3,8473,99421.91.221.65,7516,33324.81.024.7{MongolianPapuan2Thisstudy2,7832,85221.21.520.84,1924,75826.31.225.3{MongolianBougainvilleThisstudy2,8133,06624.31.522.94,2344,84726.81.126.0{Melanesian/African*Papuan1SanRef.821,98520,3663.80.75.1{35,92332,8414.50.67.2{Papuan1YorubaRef.819,10717,6464.00.84.9{30,99528,1864.70.67.4{Papuan2MbutiThisstudy3,8323,3247.11.35.4{6,1245,2337.81.17.2{BougainvilleMbutiThisstudy4,2163,5967.91.26.8{6,4985,6337.11.16.7{WepresenttheDstatisticD(H1,H2,X,chimpanzee),thenormalizeddifferencebetweenthenumberofsitesatwhichthederivedalleleinanarchaicreadfromXmatcheshumansampleH1(nBABA)andhumansampleH2(nBABA);thus,itsvalueisD5(nBABA2nABBA)/(nBABA1nABBA).Werestricttoautosomaltransversionsubstitutions,computestandarderrors(s.e.)fromablockjackknife,andhighlight(daggersymbol)theDstatisticsthataremorethanZ.3s.d.fromzero.BothNeanderthalsandDenisovansmatchEurasiansmorethantheAfricans,butthesignalsareconsistentlyandsignificantlystrongerwhenX5NeanderthalthanwhenX5Denisova.TheslightnumericaldifferenceswithTable4ofref.8areduetodifferencesinthedatafiltering.Herewerestricttocomparisonsofpresent-dayhumansamplesthatweresequencedbythesameprotocol(thefiveindividualssequencedinref.8,ortheseveninthisstudy);SupplementaryTable8.2presentsthecompletesetofpairwisecomparisons.*Comparison.{DstatisticsthataremorethanZ.3s.d.fromzero. RESEARCH ARTICLE 1056|NATURE|VOL468|23/30DECEMBER2010Macmillan Publishers Limited. All rights reserved©2010

PAGE 5

scoredeachofthesepresent-dayhumansbasedontheirrelativeproximitytoNeanderthalsandtheDenisovaindividualatpositionswherewehavehigh-qualitydataforboththeNeanderthalandDenisovagenomes(SupplementaryInformationsection8).Usingthemeansofthe53populations,thefirsttwoprincipalcomponentsseparatethepopulationsintothreegroups(Fig.2):first,the7sub-SaharanAfricanpopulations;second,agroupof44non-AfricanpopulationsaswellasonenorthAfricangroup;andthird,PapuanandBougainvillepopulationsfromMelanesia.Whenindividualsfromselectedpopulationsareanalysedseparately,thePapuanandBougainvilleislandersremaindistinctfromalmostallindividualsoutsideAfrica(SupplementaryFig.8.1b).Thus,withrespecttotheirrelationshiptoNeanderthalsandDenisovans,theMelanesianpopu-lationsstandoutrelativetoothernon-Africanpopulations.Toexplorethisfurther,weanalysedtherelationshipoftheDenisovagenometothegenomesoffivepresent-dayhumansthatwepreviouslysequencedtoaboutfivefoldcoverage8(aYorubaandaSangenomefromAfrica,aFrenchgenomefromEurope,aHangenomefromChinaandaPapuangenomefromMelanesia),aswellassevenpresent-dayhumansthatwesequencedto1…2-foldcoverageforthisstudy(aMbutigenomefromAfrica,aSardiniangenomefromEurope,aMongoliangenomefromCentralAsia,aCambodiangenomefromSouth-EastAsia,anadditionalPapuangenomefromMelanesia,aBougainvilleislandergenomefromMelanesia,andaKaritianagenomefromSouthAmerica)(SupplementaryInformationsection9).WeusedtheDstatistic8totestifvariouspairsofpresent-dayhumansshareequalnumbersofderivedalleleswiththeDenisovaindividual.Todothis,werestrictedcomparisonstopairsofpresent-dayhumanssequencedatthesametimetominimizethechancethatdifferencesinsampleprocessingcouldaffecttheresults.WefindthatthefivefoldcoveragePapuanindividualshares4.060.7%morealleleswiththeDenisovaindividualthandoestheFrenchindividual,andweobservedasimilarskewinall10comparisonsofMelanesianandothernon-Africanpopulations(Table1).Whenwestratifiedthedatabybasesubstitutionclassandchromosome,theDstatisticsarequalitativelyunchanged(SupplementaryInformationsection10).Similarly,theDstatisticsareconsistentforalldepthsofreadcoverage,indicatingthatmappingerrors,forexampleduetosegmentalduplications,arenotlikelytoexplaintheseresults.Finally,differencesinsequencingerrorrateacrosssamplescannotexplaintheobservedDstatistics(SupplementaryInformationsection10).Undertheassumptionthatgeneflowexplainstheseobservations,wedeterminedthedirectionofthisgeneflowbyaskingwhetherMelanesiansandotherEurasianssharederivedalleleswithAfricansequallyoften.IfthegeneflowwasentirelyintotheancestorsoftheDenisovanindividual,wewouldnotexpectthistoaffecttherelation-shipofAfricanstoMelanesiansandotherEurasiansandthuswewouldexpectthemtosharederivedallelesequallyoftenwithAfricans.However,wefindthatderivedallelesinAfricansmatchMelanesians3.460.4%lessoftenthanothernon-Africans(Z510.8).BecausethisskewisseenwithoutusingDenisovandataitcannotbeexplainedbygeneflowintoDenisovansor,forexample,bycontaminationoftheDenisovasamplebypresent-dayMelanesianDNA.Thus,atleastsomeoftheputativegeneflowmusthavebeenintoMelanesians(Sup-plementaryInformationsection8).Whenwecomparetheskewinthefractionofderivedallelessharedwiththetwoarchaichomininstowhatwouldbeexpectedforindividualsof100%NeanderthalorDenisovaancestry,respectively(Supplemen-taryInformationsection8andref.8),weestimatethat2.560.6%ofthegenomesofnon-AfricanpopulationsderivefromNeanderthals,inagreementwithourpreviousestimateof1…4%8.Inaddition,weestimatethat4.860.5%ofthegenomesofMelanesiansderivefromDenisovans.Altogether,asmuchas7.460.8%ofthegenomesofMelanesiansmaythusderivefromrecentadmixturewitharchaichominins.AmodelofpopulationhistoryTounderstandtheimplicationsoftherelationshipsobservedamongtheDenisovaindividual,theNeanderthalsandpresent-dayhumans,wefittheDstatisticsdescribedintheprevioussectionstoaparameterizedmodelofpopulationhistory.TheDstatisticsfortheDenisovaindi-vidualdifferintwoimportantwaysfromthosefortheNeanderthal.First,theDenisovaindividualsharesfewerderivedalleleswitheithertheFrenchorHanChinesepopulationsthandotheNeanderthals.Second,theDenisovaindividualsharesmorederivedalleleswiththePapuansthandotheNeanderthals.WeareabletofitthedatawithamodelthatassumestheDenisovansareasistergroupofNeanderthalswithapopulationdivergencetimeofone-halftotwo-thirdsofthetimetothecommonancestorofNeanderthalsandhumans.Afterthediver-genceoftheDenisovansfromNeanderthals,therewasgeneflowfromNeanderthalsintotheancestorsofallpresent-daynon-Africans.LatertherewasadmixturebetweentheDenisovansandtheancestorsofMelanesiansthatdidnotaffectothernon-Africanpopulations.ThismodelisillustratedinFig.3andisdescribedindetailinSupplementaryInformationsection11.Other,morecomplexmodelscouldalsoexplainthedata.Forexample,amodelthatinvokesonlygeneflowfromDenisovanstoMelanesianancestorsoutsideAfricaandassumesfoursubpopulationsinAfricathatexistedbetweenthetimesoftheoriginofDenisovanandNeanderthalancestorsandtheancestorsofpresent-dayEurasianscouldalsofitthedata(SupplementaryFig.11.4).However,becausebarrierstogeneflowbetweensuchsubpopulationswouldhavetopersistforhundredsofthousandsofyearstocreatetheobservedpatterns,suchamodelislessplausibleonbiologicalgroundsthanamodelthatinvokestwoinstancesofgeneflowoutsideAfrica.DiscordanceofmtDNAandnuclearhistoriesThepopulationhistoryindicatedbythenucleargenomeisdifferentfromthatindicatedbythemtDNAphylogeny.Therearetwopossibleexplanationsforthis.OneisthatthemtDNAlineagewasintroducedintoDenisovanancestorsbyadmixturefromanotherhomininlineageforwhichwehavenodata.Theotheristhatthediscordanceistheresultofincompletelineagesorting,thatis,therandomassortment Ð0.03Ð0.025Ð0.02Ð0.015Ð0.01Ð0.005 0Ð0.06Ð0.055Ð0.05Ð0.045Ð0.04Ð0.035Eigenvector 2Ei g envector 1 Papuan Bougainville Non-African African Figure2|Relationshipofpresent-daypopulationstotheDenisovaindividualandNeanderthalsbasedon255,077SNPs.Principalcomponentanalysisofthemeansof53present-dayhumanpopulationsprojectedontothetoptwoprincipalcomponentsdefinedbyDenisova,Neanderthalandchimpanzee.ThesevenAfricanpopulationsareSan,Mbuti,Biaka,BantuKenya,BantuSouthAfrica,YorubaandMandenka;theNon-Africanpopulationsare44diversegroupsfromoutsideAfricaexceptforPapuanandBougainvilleislanders. ARTICLE RESEARCH23/30DECEMBER2010|VOL468|NATURE|1057Macmillan Publishers Limited. All rights reserved©2010

PAGE 6

ofgeneticlineagesduetogeneticdriftwhichmayhaveallowedadivergentmtDNAlineagetosurviveinDenisovansbychancewhilebecominglostinNeanderthalsandmodernhumans.Alargeancestralpopulationsizemakesincompletelineagesortingmorelikelytooccur.InSupplementaryInformationsection11,weshowthatgivenreasonableassumptionsaboutthesizeoftheancestralpopulations,thediscordanceofthemtDNAphylogenywiththatindicatedbythenuclearDNAcanbeexplainedeitherbyasmallamountofadmixturefromanotherarchaichomininorbyincompletelineagesorting.Thus,thedatadonotallowustofavouronehypothesisovertheother.AtoothfromDenisovaCaveIn2000,ahominintoothwasdiscoveredinlayer11.1ofthesouthgalleryofDenisovaCave(Fig.4a,b).Thetoothisfromayoungadultandthereforefromanotherindividualthanthephalanxwhichstemsfromajuvenile(SupplementaryInformationsection12).Toelucidatetherelationshipofthetoothtotheindividualfromwhichthephalanxisderived,weextractedDNAfrom50mgofdentinfromtherootofthetoothandpreparedasequencinglibrary(SupplementaryInformationsection13).About0.17%ofrandomDNAsequencesdeterminedfromthislibraryalignedtothehumangenome,whereastherestislikelytorepresentmicrobialcontaminationcommoninancientbones.WethereforeusedanovelDNAcaptureapproach36toisolatemtDNAsequencesfromthesequencinglibrary.Atotalof15,094sequenceswereidentifiedwhichallowedthecompletemtDNAgenometobeassembledatanaveragecoverageof58-fold.ThissequencediffersattwopositionsfromthemtDNAofthephalanxwhereasitdiffersatabout380positionsfrombothNeanderthalandpresent-dayhumans.ThetimesincethemostrecentcommonancestorofthetwomtDNAsfromDenisovaCaveisestimatedtobe7,500years,witha95%upperboundof16,000years(SupplementaryInformationsection13).Weconcludethatthetoothandthephalanxderivefromtwodifferentindividualsthatareprobablyfromthesamehomininpopulation.MorphologyoftheDenisovamolarThetoothisanalmostcompleteleft,probablythird,butpossiblysecond,uppermolar(Fig.4b).Thecrownistrapezoidalandtapersstronglydistally,withbulginglingualandbuccalwallsgivingthetoothaninflatedappearance(SupplementaryInformationsection12).Therootsareshortbutrobustandstronglyflaring.Overall,thetoothisverylarge(mesiodistaldiameter,13.1mm;buccolingual,14.7mm).Asathirdmolar,itisoutsidetherangeofnormalsizevariationofallfossiltaxaofthegenusHomo,withtheexceptionofH.habilisandH.rudolfensis,andcomparabletoAustralopithecines(Fig.4c).Comparedtosecondmolars,itislarger Timef2f1African (Afr)FrenchHanMelanesian Vindija Neanderthal (V)Denisova (D)tAfrtVtDNNtGF1tGF2N0 Figure3|Amodelofpopulationhistorycompatiblewiththedata.Ndenoteseffectivepopulationsize,tdenotestimeofpopulationseparation,fdenotesamountofgeneflowandtGFdenotestimeofgeneflow. abc10 mm17M3 mesiodistalM3 buccolingual161514131211109879101112131415Australopithecus afarensis (n = 14)Australopithecus africanus (n = 12)Homo habilis (n = 7)DmanisiAfrican Homo erectusIndonesian Homo erectusChinese Homo erectus (n = 7)Early AMH (n = 6)Neanderthals (n = 14)Upper Palaeolithic (n = 11)Oase 2DenisovaEuropean Homo heidelbergensisAtapuerca SH16171819 Figure4|MorphologyoftheDenisovamolar.a,b,Occlusal(a)andmesial(b)views.c,ComparisonoftheDenisovamolartodiversethirdmolars,inabiplotofthemesiodistalandbuccolinguallengths(inmm).AMH,anatomicallymodernhumans;SH,SimadelosHuesos.SupplementaryFig.12.1presentsasimilarcomparisontosecondmolars. RESEARCH ARTICLE 1058|NATURE|VOL468|23/30DECEMBER2010Macmillan Publishers Limited. All rights reserved©2010

PAGE 7

thanNeanderthalsorearlymodernhumans,butsimilartoH.erectusandH.habilis(SupplementaryFig.12.1).Besidessize,itisalsodistinguishedfrommostNeanderthalthirdmolarsbytheabsenceofhypoconereduction,andfrombothsecondandthirdNeanderthalmolarsbythepresenceofalargetalonbasinandthestrongflareofthecrown.Furthermore,itlacksthelingualhypoconeprojectionseeninallNeanderthalfirstandmanysecondmolars,andhasstronglydivergingroots,unlikethecloselyspacedandfrequentlyfusedrootsofNeanderthals.ItisofparticularinteresttocomparetheDenisovamolartoMiddlePleistocenehomininsfromChina,whereH.erectusandotherarchaicforms,sometimesinterpretedasH.heidelbergensis,mayhavesurviveduntilrecently.Unfortunately,veryfewofthesefossilspreservethirduppermolars.Ofthefewexamplesthatareavailable,mostdifferfromtheDenisovamolarbytheirstronglyreducedsize.Secondmolarsaremorefrequentthanthirdmolars,andmosthaveatrapezoidalshapelikeDenisova,buttheydonothavethelinguallyskewedpositionofthehypoconeandmetaconeandthestrongbasalflareofthecrown.TheDenisovamolarsupportstheDNAevidencethattheDenisovanpopulationisdistinctfromlateNeanderthalsaswellasfrommodernhumans.Infact,theprimitivetraitsoftheDenisovatoothsuggestthatDenisovansmayhavebeenseparatedfromtheNeanderthallineagebeforeNeanderthaldentalfeaturesaredocumentedinWesternEurasia(.300,000yearsBP)(SupplementaryInformationsection12),althoughwecannotexcludethepossibilitythattheDenisovandentalmorphologyresultsfromareversion.StratigraphyanddatingThesmallsizeofboththephalanxandthetoothprecludesdirectradiocarbondating.Weinsteaddatedsevenbonefragmentsfoundclosetothehomininremainsinlayer11intheeastandsouthgalleries.Toensurethattheywereassociatedwithhumanoccupationofthecavewechosebonesthathaveevidenceofhumanmodification,includingaribwithregularincisionsandaboneprojectilepointblankgenerallyassociatedwithUpperPalaeolithicculturalassemblages.Inthesouthgallery,wheremodifiedboneswerenotavailable,weusedherbivorebones(SupplementaryInformationsection12).Fourofthesevendatesareinfinitedatesolderthan50,000yearsBP(uncalibrated),whereasthreearefinitedatesbetween16,000and30,000yearsBP(SupplementaryTable12.1).Theribwithincisionsandtheprojectilepointblankareabout30,000and23,000yearsBP,respectively.Togetherwiththreepreviousdates23thisshowsthatlayer11containsculturalremainsfromatleasttwodifferenttimeperiods,oneperiodolderthan50,000yearsBPandonemorerecentperiod.However,thestratigraphyiscomplicatedbythediscoveryofawedge-shapedareaclosetotheareawherethephalanxwasfoundthatislikelytobedisturbed(SupplementaryInformationsection12).Homininremainslargeenoughtoallowdirectradiocarbondatesmayeven-tuallybediscoveredinthecave,butareasonablehypothesisisthatthephalanxandmolarbelongtotheolderoccupation.DiscussionThemolecularpreservationoftheDenisovaphalanxisexceptionalinthatthefractionofendogenousrelativetomicrobialDNAisabout70%.Bycontrast,inallNeanderthalremainsstudiedsofartherelativeabundanceofendogenousDNAisbelow5%,andtypicallybelow1%.Furthermore,theaveragelengthofhomininDNAfragmentsintheDenisovaphalanxis58basepairs(bp)(SL3003)and74bp(SL3004)inspiteoftheenzymatictreatmentthatremovesuracilresiduesanddecreasestheaveragefragmentsize,whereasinmostwell-preservedNeanderthalsamplesitis50bporsmallerwithoutthistreatment.Thus,althoughmanyNeanderthalsarepreservedunderconditionsapparentlysimilartothoseinDenisovaCave,theDenisovaphalanxisoneoffewbonesfoundintemperateconditionsthatareaswellpre-servedasmanypermafrostremains37,38.Itisnotclearwhythisis.ItisnotduetosomeconditionthataffectsallhomininremainsinDenisovaCavebecausethefractionofendogenousDNAinthetoothis0.17%;thatis,typicalofotherLatePleistocenehomininremains.Itispossiblethatarapiddesiccationofthetissueafterdeath,whichwouldlimitdegradationoftheDNAbyendogenousenzymesaswellasmicrobialgrowth,hasallowedthisexceptionalpreservation.TheDenisovaindividualandthepopulationtowhichitbelongedcarrysomeexceptionallyarchaicmolecular(mtDNA)aswellasmor-phological(dental)features.Nevertheless,thepicturethatemergesfromanalysisofthenucleargenomeisonewheretheDenisovapopu-lationisasistergrouptoNeanderthals.ThreepossibilitiescouldaccountforhowsucharchaicfeatureshavecometobepresentinDenisovans.OnepossibilityisthatthesefeatureswereretainedinDenisovansbutbecamelostinmodernhumansandNeanderthals.Asecond,notmutuallyexclusive,possibilityisthattheyenteredtheDenisovapopulationthroughgeneflowfromsomeevenmoredivergedhominin.AlthoughsuchgeneflowcannotbedetectedwiththecurrentmtDNAandnuclearDNAdata,furthersequencingofotherhomininremainsmayinthefutureallowtestingforit.Athirdpossibilitythatcouldaccountfortheapparentlyarchaicdentalmor-phology,butnotthemtDNA,isareversaltoancestraltraits.Aftertheydivergedfromoneanother,DenisovansandNeanderthalshadlargelyseparatepopulationhistoriesasshownbyanumberofobservations.First,patternsofallelesharingindicatethatDenisovanancestorsdidnotcontributegenesatadetectableleveltopresent-daypeoplealloverEurasiawhereasNeanderthalsdid8.Thus,Neanderthalsatsomepointinteractedwithancestorsofpresent-dayEurasiansinde-pendentlyofDenisovans.Second,thegeneticdiversityofNeanderthalsacrosstheirgeographicalrangeinthelastthirtyorfortythousandyearsoftheirhistorywasextremelylow,indicatingthattheyexperiencedoneormorestronggeneticbottlenecksindependentlyoftheDenisovans.Third,ourresultsindicatethatDenisovansbutnotNeanderthalscon-tributedgenestoancestorsofpresent-dayMelanesians.Fourth,thedentalmorphologyshowsnoevidenceofanyderivedfeaturesseeninNeanderthals.Infact,dentalremainsfromtheSimadelosHuesosofAtapuerca,forwhichagesbetween350,000and600,000yearshavebeenproposed39,40,alreadycarryNeanderthal-likemorphologicalfeaturesthatarenotseenintheDenisovamolar.AninterestingquestionishowwidespreadDenisovanswere.ApossibilityisthattheylivedinlargepartsofEastAsiaatthetimewhenNeanderthalswerepresentinEuropeandwesternAsia.OneobservationcompatiblewiththispossibilityisthatDenisovanrela-tivesseemtohavecontributedgenestopresent-dayMelanesiansbutnottopresent-daypopulationswhichcurrentlylivemuchclosertotheAltairegionsuchasHanChineseorMongolians(Table1).Thus,theyhaveatleastatsomepointbeenpresentinanareawheretheyinteractedwiththeancestorsofMelanesiansandthiswaspresumablynotinsouthernSiberia.Furtherstudiesofbothmolecularandmor-phologicalfeaturesofhomininremainsacrossAsiashouldclarifyhowwidespreadDenisovanswereandhowtheywererelatedtoarchaichomininsotherthanNeanderthals.TheDenisovaindividualbelongstoahominingroupthatsharesacommonancestorwithNeanderthalsbuthasadistinctpopulationhistory.WedefinethisgroupbasedongenomicevidenceandcallitDenisovans,butrefrainfromanyformalLinnaeantaxonomicdesig-nationsthatwouldindicatespeciesorsubspeciesstatusforeitherNeanderthalsorDenisovans.Inourview,theseresultsshowthatontheEurasianmainlandthereexistedatleasttwoformsofarchaichomininsintheUpperPleistocene:awesternEurasianformwithmorphologicalfeaturesthatarecommonlyusedtodefinethemasNeanderthals,andaneasternformtowhichtheDenisovaindividualsbelong.Inthefuture,whenmorecompletegenomesfromtheseandotherarchaichomininswillbesequencedfromremainsthatallowmoremorphologicalfeaturestoassessed,theirrelationshipswillbecomeevenbetterunderstood.ThiswillbeanimportantendeavourastheemergingpictureofUpperPleistocenehomininevolutionisoneinwhichgeneflowamongdifferenthominingroupswascommon. ARTICLE RESEARCH23/30DECEMBER2010|VOL468|NATURE|1059Macmillan Publishers Limited. All rights reserved©2010

PAGE 8

METHODSSUMMARYThethirteensectionsoftheSupplementaryInformationprovideafulldescriptionofthemethods.Received15August;accepted30November2010.1.Hublin,J.J.TheoriginofNeandertals.Proc.NatlAcad.Sci.USA106,16022…16027(2009).2.Pope,G.G.CraniofacialevidencefortheoriginofmodernhumansinChina.Am.J.Phys.Anthropol.35,(Suppl.15),243…298(1992).3.Rightmire,G.P.BrainsizeandencephalizationinearlytoMid-PleistoceneHomo.Am.J.Phys.Anthropol.124,109…123(2004).4.Wu,X.&Poirier,F.E.HumanEvolutioninAsia(OxfordUniv.Press,1995).5.Brown,P.etal.Anewsmall-bodiedhomininfromtheLatePleistoceneofFlores,Indonesia.Nature431,1055…1061(2004).6.Morwood,M.J.etal.ArchaeologyandageofanewhomininfromFloresineasternIndonesia.Nature431,1087…1091(2004).7.Morwood,M.J.etal.Preface:researchatLiangBua,Flores,Indonesia.J.Hum.Evol.57,437…449(2009).8.Green,R.E.etal.AdraftsequenceoftheNeandertalgenome.Science328,710…722(2010).9.Beauval,C.etal.AlateNeandertalfemurfromLesRochers-de-Villeneuve,France.Proc.NatlAcad.Sci.USA102,7085…7090(2005).10.Briggs,A.W.etal.TargetedretrievalandanalysisoffiveNeandertalmtDNAgenomes.Science325,318…321(2009).11.Caramelli,D.etal.AhighlydivergentmtDNAsequenceinaNeandertalindividualfromItaly.Curr.Biol.16,R630…R632(2006).12.Green,R.E.etal.AcompleteNeandertalmitochondrialgenomesequencedeterminedbyhigh-throughputsequencing.Cell134,416…426(2008).13.Krause,J.etal.NeanderthalsincentralAsiaandSiberia.Nature449,902…904(2007).14.Krings,M.etal.AviewofNeandertalgeneticdiversity.NatureGenet.26,144…146(2000).15.Lalueza-Fox,C.etal.MitochondrialDNAofanIberianNeandertalsuggestsapopulationaffinitywithotherEuropeanNeandertals.Curr.Biol.16,R629…R630(2006).16.Orlando,L.etal.RevisitingNeandertaldiversitywitha100,000yearoldmtDNAsequence.Curr.Biol.16,R400…R402(2006).17.Serre,D.etal.NoevidenceofNeandertalmtDNAcontributiontoearlymodernhumans.PLoSBiol.2,e57(2004).18.Endicott,P.,Ho,S.Y.&Stringer,C.UsinggeneticevidencetoevaluatefourpalaeoanthropologicalhypothesesforthetimingofNeanderthalandmodernhumanorigins.J.Hum.Evol.59,87…95(2010).19.Krause,J.etal.ThecompletemitochondrialDNAgenomeofanunknownhomininfromsouthernSiberia.Nature464,894…897(2010).20.Derevianko,A.P.etal.PaleoenvironmentandPaleolithicHumanOccupationofGornyAltai[inRussian](InstituteofArcheologyandEthnographySBRASPress,2003).21.Derevianko,A.ThePaleolithicofSiberia:NewDiscoveriesandInterpretations(Univ.IllinoisPress,1998).22.Derevianko,A.P.etal.Archaeology,Geology,andthePleistoceneandHolocenePalaeogeographyoftheMountainousAltai[inRussian](Nauka,1998).23.Derevianko,A.P.,Shunkov,M.V.&Volkov,P.V.APaleolithicbraceletfromDenisovacave.Archaeol.Ethnol.Anthropol.Eurasia34,13…25(2008).24.Green,R.E.etal.TheNeandertalgenomeandancientDNAauthenticity.EMBOJ.28,2494…2502(2009).25.Rohland,N.&Hofreiter,M.ComparisonandoptimizationofancientDNAextraction.Biotechniques42,343…352(2007).26.Lindahl,T.etal.DNAN-glycosidases:propertiesofuracil-DNAglycosidasefromEscherichiacoli.J.Biol.Chem.252,3286…3294(1977).27.Briggs,A.W.etal.RemovalofdeaminatedcytosinesanddetectionofinvivomethylationinancientDNA.NucleicAcidsRes.38,e87(2010).28.Meyer,M.&Kircher,M.Illuminasequencinglibrarypreparationforhighlymultiplexedtargetcaptureandsequencing.ColdSpringHarb.Protoc.2010,doi:10.1101/pdb.prot5448(2010).29.Briggs,A.W.etal.PatternsofdamageingenomicDNAsequencesfromaNeandertal.Proc.NatlAcad.Sci.USA104,14616…14621(2007).30.Kircher,M.,Stenzel,U.&Kelso,J.ImprovedbasecallingfortheIlluminaGenomeAnalyzerusingmachinelearningstrategies.GenomeBiol.10,R83(2009).31.Li,H.&Durbin,R.FastandaccurateshortreadalignmentwithBurrows-Wheelertransform.Bioinformatics25,1754…1760(2009).32.Paten,B.etal.EnredoandPecan:genome-widemammalianconsistency-basedmultiplealignmentwithparalogs.GenomeRes.18,1814…1828(2008).33.Reich,D.etal.ReducedneutrophilcountinpeopleofAfricandescentisduetoaregulatoryvariantintheDuffyantigenreceptorforchemokinesgene.PLoSGenet.5,e1000360(2009).34.Keinan,A.,Mullikin,J.C.,Patterson,N.&Reich,D.MeasurementofthehumanallelefrequencyspectrumdemonstratesgreatergeneticdriftinEastAsiansthaninEuropeans.NatureGenet.39,1251…1255(2007).35.Li,J.Z.etal.Worldwidehumanrelationshipsinferredfromgenome-widepatternsofvariation.Science319,1100…1104(2008).36.Maricic,T.,Whitten,M.&Pa¨a¨bo,S.MultiplexedDNASequenceCaptureofMitochondrialGenomesUsingPCRProducts.PLoSONE5,e14004(2010).37.Poinar,H.N.etal.Metagenomicstopaleogenomics:large-scalesequencingofmammothDNA.Science311,392…394(2006).38.Rasmussen,M.etal.AncienthumangenomesequenceofanextinctPalaeo-Eskimo.Nature463,757…762(2010).39.Bischoff,J.L.etal.High-resolutionU-seriesDatesfromtheSimadelosHuesosHominidsYields6001266kyrs:ImplicationsfortheEvolutionoftheEarlyNeanderthalLineage.J.Archaeol.Sci.34,763…770(2007).40.Bischoff,J.L.etal.TheSimadelosHuesosHominidsDatetoBeyondU/ThEquilibrium(.350kyr)andPerhapsto400…500kyr:NewRadiometricDates.J.Archaeol.Sci.30,275…280(2003).SupplementaryInformationislinkedtotheonlineversionofthepaperatwww.nature.com/nature.AcknowledgementsWethankC.Bustamante,A.I.Krivoshapkin,M.Lachmann,R.Nielsen,K.Pruefer,A.Tsybankov,L.VigilantandW.Zhaiforcomments;K.Finstermeierforgraphicswork;theMPI-EVAsequencinggroup,R.SchultzandS.Weihnachtsmannfortechnicalsupport;andP.Fujita,A.HinrichsandK.LearnedfordesigningtheUCSCgenomebrowserportaltotheDenisovadata.ThePresidentialInnovationFundoftheMaxPlanckSocietyandtheKrekelerFoundationprovidedfinancialsupport.M.S.wassupportedbyaUSNationalInstitutesofHealthgrant(R01-GM40282).TheNationalScienceFoundationprovidedanInternationalPostdoctoralFellowship(OISE-0754461)toJ.M.G.,aFellowshipinBiologicalInformaticstoP.L.F.J.andaHOMINIDgrant(1032255)toD.R.AuthorContributionsJ.Kr.,T.M.,Q.F.andM.M.performedtheexperiments;D.R.,R.E.G.,M.K.,J.Kr.,N.P.,E.Y.D.,A.W.B.,U.S.,P.L.F.J.,T.M.,J.M.G.,T.B.-M.,C.A.,S.M.,H.L.,E.E.E.,M.St.,J.Ke.,M.Sl.andS.P.analysedgeneticdata;B.V.,M.R.,S.T.,M.V.S.,A.P.D.andJ.-J.H.analysedarchaeologicalandanthropologicaldata;D.R.andS.P.wroteandeditedthemanuscript.AuthorInformationTherawsequencedatafromthetwoDenisovafossils,thesevenpresent-dayhumans,andthetoothmtDNAhavebeendepositedintheEuropeanNucleotideArchiveatEMBL-EBIunderaccessionnumbersERP000318,ERP000121andFR695060,respectively.ThealignmentsofDenisovasequencereadstothehumanandchimpanzeegenomesareaccessibleforbrowsinganddownloadfromhttp://genome.ucsc.edu/Denisova.Reprintsandpermissionsinformationisavailableatwww.nature.com/reprints.Theauthorsdeclarenocompetingfinancialinterests.Readersarewelcometocommentontheonlineversionofthisarticleatwww.nature.com/nature.CorrespondenceandrequestsformaterialsshouldbeaddressedtoD.R.(reich@genetics.med.harvard.edu),B.V.(bence.viola@eva.mpg.de)orS.P.(paabo@eva.mpg.de). RESEARCH ARTICLE 1060|NATURE|VOL468|23/30DECEMBER2010Macmillan Publishers Limited. All rights reserved©2010


printinsert_linkshareget_appmore_horiz

Download Options

close


  • info Info

    There are only PDFs associated with this resource.

  • link PDF(s)



Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.