Giant lungfish genome elucidates the conquest of land by ...

Lungfishes belong to lobe-fined fish (Sarcopterygii) that, ... The vast size of this genome, which is about 14× larger than that of humans, ... Skiptomaincontent Thankyouforvisitingnature.com.YouareusingabrowserversionwithlimitedsupportforCSS.Toobtain thebestexperience,werecommendyouuseamoreuptodatebrowser(orturnoffcompatibilitymodein InternetExplorer).Inthemeantime,toensurecontinuedsupport,wearedisplayingthesitewithoutstyles andJavaScript. Advertisement nature articles article Giantlungfishgenomeelucidatestheconquestoflandbyvertebrates DownloadPDF Subjects EvolutionarybiologyEvolutionarygeneticsGenomeevolutionPhylogenetics AbstractLungfishesbelongtolobe-finedfish(Sarcopterygii)that,intheDevonianperiod,‘conquered’thelandandultimatelygaverisetoalllandvertebrates,includinghumans1,2,3.Herewedeterminethechromosome-qualitygenomeoftheAustralianlungfish(Neoceratodusforsteri),whichisknowntohavethelargestgenomeofanyanimal.Thevastsizeofthisgenome,whichisabout14×largerthanthatofhumans,isattributablemostlytohugeintergenicregionsandintronswithhighrepeatcontent(around90%),thecomponentsofwhichresemblethoseoftetrapods(comprisingmainlylonginterspersednuclearelements)morethantheydothoseofray-finnedfish.Thelungfishgenomecontinuestoexpandindependently(itstransposableelementsarestillactive),throughmechanismsdifferenttothoseoftheenormousgenomesofsalamanders.The17 fullyassembledlungfishmacrochromosomesmaintainsyntenytoothervertebratechromosomes,andallmicrochromosomesmaintainconservedancienthomologywiththeancestralvertebratekaryotype.Ourphylogenomicanalysesconfirmpreviousreportsthatlungfishoccupyakeyevolutionarypositionastheclosestlivingrelativestotetrapods4,5,underscoringtheimportanceoflungfishforunderstandinginnovationsassociatedwithterrestrialization.Lungfishpreadaptationstolivingonlandincludethegainoflimb-likeexpressionindevelopmentalgenessuchashoxc13andsall1intheirlobedfins.Increasedratesofevolutionandtheduplicationofgenesassociatedwithobligateair-breathing,suchaslungsurfactantsandtheexpansionofodorantreceptorgenefamilies(whichencodeproteinsinvolvedindetectingairborneodours),contributetothetetrapod-likebiologyoflungfishes.Thesefindingsadvanceourunderstandingofthismajortransitionduringvertebrateevolution. DownloadPDF MainLungfish(Dipnoi)sharewithland-dwellingvertebratestheabilitytobreatheairthoughlungs,whicharehomologoustoourown.Sincetheirdiscoveryinthenineteenthcentury,lungfishhaveattractedscientificinterestandwereinitiallythoughttobeamphibians6,7.Wenowknowthattheyaremorecloselyrelatedtotetrapodsthantoray-finnedfish.Oftheextantlungfishspecies(ofwhichthereareonlysix),fourliveinAfrica,oneinSouthAmericaandone(N. forsteri)inAustralia.LungfishappearedinthefossilrecordintheDevonianperiod,around400 millionyearsago(Ma)1.Somescholarshiphasdiscussedlungfishas‘livingfossils’,becausetheirmorphologybarelychangedovermillionsofyears:forexample,>100-million-year-oldfossilsfromAustraliastronglyresemblethesurvivingspecies(whichrepresentsoneoftheoldestknownanimalgenera,discoveredexactly150 yearsago)2.Owingtotheancestralcharacters(suchasbodyshape,largescalesandpaddle-shapedfins)ofN. forsteri,itresembles‘archetypal’extinctlungfishmuchmorethanthetwootherlineagesofextantlungfish.TheSouthAmericanand, inparticular,theAfricanlungfishhavealmostcompletely losttheirscalessecondarilyandhavesimplifiedtheirfinmorphologyintothinfilaments,althoughtheydoshowthealternatinggaitsthataretypicalofterrestriallocomotion.Togetherwiththecoelacanthsandtetrapods,lungfisharemembersoftheSarcopterygii(lobe-finnedfish);however,owingtotheshortbranchthatseparatesthesethreeancientlineagesithasremaineddifficulttoresolvetheirrelationships.DevelopmentsofpowerfulDNAsequencingandcomputationalmethodsenableustonowrevisitlong-standingevolutionaryquestionsregardingtheserelationshipsusingwhole-genome-deriveddatasetswithmorerobustorthologyinferencesthanhavehithertobeenpossible.Previousanalysesusinglargetranscriptomicdatasetshavetendedtosupportthehypothesisthatlungfisharetheclosestlivingrelativesoftetrapods4,5.Lungfisharethereforecrucialforunderstandingtheevolutionandpreadaptationsthataccompaniedthetransitionofvertebratelifefromwatertoland.Thismajorevolutionaryeventrequiredanumberofevolutionaryinnovations,includinginrespiration,limbs,posture,thepreventionofdesiccation,nitrogenexcretion,reproductionandolfaction.Lungfishareknowntohavethelargestanimalgenome(http://www.genomesize.com/search.php),butthemechanismsthatledtoandmaintainedtheirgenomesizesarepoorlyunderstood.Therefore,theAustralianlungfishmightprovideinsightsbothintotetrapodinnovationsandevolution,andthestructureofgiantgenomes.Genomesequencing,assemblyandannotationThelargestanimalgenomesequencedsofaristhe32-Gb8genomeoftheaxolotlsalamander(Ambystomamexicanum).Toovercomethechallengesofsequencingandassemblingtheeven-largergenomesoflungfish,weusedlong-andultra-long-readNanoporetechnologytogenerate1.2 Tbin3 batches:601 GbwithanN50read-lengthof9 kb;532 GbwithanN50of27 kb;and1.5 GbwithanN50of46 kb,allfromajuvenileAustralianlungfish.WeassembledthesethreebatchesintocontigsusingtheMARVELassembler8(ExtendedDataFig.1a,Methods).Thisyieldeda37-GbassemblywithanN50contigsize1.86 Mb(SupplementaryTable1).Tocorrectforinsertionsand/ordeletions,gaps,single-nucleotidepolymorphismsandsmalllocalmisalignmentsintheprimaryassembly,weused1.4-TbDNAand499.8-GbRNAIlluminareads.Thegenome-correctionDNAdata—sequencedatmorethan30×coverage—wereusedtoestimategenomesizethroughfrequenciesofk-mers(ExtendedDataFig.2).Weascertainedthehighcompletenessofthe37-Gbassemblybyobservingthat88.2%oftheDNAand84%oftheRNAsequencing(RNA-seq)readsalignedtothegenome,whichgivesanestimatedtotalgenomesizeof43 Gb(about30%largerthantheaxolotl8).Thismatchesthek-mervaluebutissmallerthanthatpredictedbyflowcytometry(52 Gb9)andFeulgenphotometry(75 Gb10).Next,wescaffoldedthecontigsusing271-Gbchromosomeconformationcapture(Hi-C) IlluminaPE250readstoachromosome-scaleassemblywithanN50of1.75 Gb(ExtendedDataFig.1d,Methods).WealsousedHi-Cdatatodetectmisjoins,bybinningHi-Ccontactsalongthediagonalandidentifyingpointsthatweredepletedofcontacts(ExtendedDataFig.1e).Thelargestscaffoldscorrespondtothe17 macrochromosomesarmsofthekaryotypeofN. forsteri.Wealsoassembledalltenmicrochromosomesintosinglescaffolds(SupplementaryInformation).Weconstructedacomprehensivemulti-tissuedenovotranscriptomeassembly(BUSCOscoreofover98%corevertebrategenes)usingRNAextractedfromthesameindividuallungfish.Forannotationofprotein-codinggenes,wecombinedevidencefromtranscriptalignmentsandhomology-basedgeneprediction.Thisresultedin31,120 high-fidelitygenemodels.WeassessedthecompletenessofthegenomeassemblyusingthepredictedgenesetandtheBUSCOpipeline,detecting91.4%ofcorevertebrategenes(233 genes)and90.9%ofvertebrateconservedgenes(2,586 genes)(SupplementaryTable2).Wepredicted17,095 noncodingRNAs(ncRNAs),including1,042transferRNAs(tRNAs),1,771 ribosomalRNAs(rRNAs)and3,974 microRNAs(SupplementaryTable3,SupplementaryInformation).Phylogenyoflungfish,coelacanthandtetrapodsPhylogeneticrelationshipsamongcoelacanths,lungfishesandtetrapodshavebeendebated4,5,11.WeusedBayesianphylogenomics(Fig.1)with697 one-to-oneorthologuesfor10 vertebrates,withacomplexmixturemodelthatcanovercomelong-branchattractionartefacts4andalsousednoncodingconservedgenomicelements(96,601 alignedsites)(ExtendedDataFig.3a).Bothdatasetsunequivocallysupportlungfish4,5astheclosestlivingrelativesoflandvertebrates,withwhichtheysharedalastcommonancestoraround420 Ma(ExtendedDataFig.3b).Fig.1:Bayesianphylogenybasedon697 one-to-one orthologues.ThisanalysisusedtheCAT-GTRmodelinPhyloBayesMPI.Allbranchesweresupportedbyposteriorprobabilitiesof1.Theproteinandanoncodingconservedgenomicelementdatasets(ExtendedDataFig.3a)recoveredidenticalandhighlysupportedvertebraterelationships(posteriorprobability = 1.0and100%bootstrapforallbranches).Scalebarisexpectedaminoacidreplacementspersite.FullsizeimageSyntenyconservedofmacro-andmicrochromosomesLineage-specificpolyploidyeventsareimportantevolutionaryforces12thatcanalsoleadtogenomeexpansionsinlungfish9,13.Despitethemassivegenomeexpansioninlungfishrelativetootheranimals,thelungfishchromosomalscaffoldsstronglyresembletheancestralchordatekaryotype(Fig.2a,ExtendedDataFig.4a,b).Onthebasisof17 chordatelinkagegroups(CLGs)14,15and6,337 markersmappedontothelungfishgenome,weuncoveredconservedsynteniccorrespondencebetweenlungfishchromosomesandCLGs(Fig.2a).Theancestorofvertebratesunderwenttworoundsofwhole-genomeduplication.LungfishalsoretainedmoreancientCLGchromosomalfusionsthroughthesetworoundsofvertebrateduplication15.Inlungfish,CLGfusionsfrombeforethesecondroundofwhole-genomeduplicationsarepreservedintactbutsubstantiallyexpanded(Fig.2b).AlmostalladditionalCLGfusionshappenedrecently,asindicatedbysharpsyntenicboundaries(Fig.2b).This,alongwiththe‘vertebrate-typical’genenumberofN. forsteri,confirmsthediploidyofthegenome.Fig.2:Conservedsyntenyandchromosomalexpansioninlungfish.a,MappingofCLGsontolungfishchromosomes.Orthologousgenefamilynumbersareshown.Eachdotrepresentsanorthologousgenefamily,CLGsareaspreviouslydefined15.Scaffolds01–17representlungfishmacrochromosomes,andscaffolds18–27representmicrochromosomes.SignificantlyenrichedCLGsonlungfishchromosomesindicatedbyrectangles(forrawdata,seeExtendedDataFig.4f).b,Expansionofhomologouschromosomesinlungfish(left),comparedtospottedgar(right)(hereonlyLG8isshown;theotherchromosomesareinExtendedDataFig.4a).ChromosomesarepartitionedintobinsandCLGcontentisprofiled;chromosomalpositionisplottednexttoeachchromosome.LG8ingarhasaprominentjawed-vertebrate-specificfusionoftheCLGsEandO,whichisretainedthroughoutthewholechromosomeinlungfish(despitethelatterbeing>30-foldlarger).ThesmallboxinthemiddleistheunexpandedLG8ofspottedgar.c,Preservationofmicrochromosomes.Chickenmicrochromosomesareplotted(forgar,seeExtendedDataFig.4d)alongwiththeirlungfishhomologueswith>50 orthologues.Scaffolds01–17representlungfishmacrochromosomes,andscaffolds18–27representmicrochromosomes.Forchicken,onlymicrochromosomesareshown.Significantlyenrichedchickenmicrochromosomesonlungfishchromosomesindicatedbyrectangles(forrawdata,seeFig.4e).Mostchickenmicrochromosomesareinone-to-onecorrespondencewithlungfish,butsomelungfishmicrochromosomeshaverecentlybeenincorporatedintomacrochromosomes.Theselungfishmacrochromosomes(forexample,scaffold 01orscaffold 02)havesignificantassociationwithbothchickenmacro-andmicrochromosomes.However,thosefusionsarerecentinlungfish,becausethepositionsofchickenorthologuesarerestrictedtospecificareasofthelungfishchromosomes, asisevidentfromthesharpsyntenicboundaries(indicatedbypinkarrowsonscaffold01,scaffold02andscaffold06).Silhouettesarefromapreviouspublication36.SignificancesweredeterminedbyFischer’sexacttest,P value ≤ 0.01.FullsizeimageAlltenlungfishmicrochromosomes(inferredfromkaryotype9andourassembly(ExtendedDataFig.4))couldbehomologizedtothemicrochromosomesofchickenandgar(Fig.2c,ExtendedDataFig.4c,d)—andeventheymostlyretainedtheirco-linearity.This,alongwiththeconservationofsomemicrochromosomesingar,chickenandgreenanole15,16,suggeststhatmicrochromosomesmaydatebacktotheearliestvertebrates.Thecompleteretentionofmicrochromosomesinthemassivelyexpandedlungfishgenomesuggeststhatstabilizingselectionmaintainstheseancestralunits.Insupportofthis,lungfishmicrochromosomesshow—onaverage—highergenedensitiesandalowerdensityoflonginterspersednuclearelements(LINEs),whicharethemajorcontributorstogenomesize(ExtendedDataFig.4b);thisalsosuggestsdifferentexpansiondynamicsofvertebratemicro-andmacrochromosomes.HallmarksofthegiantlungfishgenomeAmaximumlikelihoodreconstructionoftheancestralgenomesizesofvertebratesshows2 majorindependentgenome-expansioneventsinlungfishandsalamanderlineages(ExtendedDataFig.3c),initiallyatsimilarratesinbothlineages(161–165 Mbpermillionyears)butsubsequentlyatslowerratesintheAustralianlungfish(about39 Mbpermillionyears),butpossiblynotintheotherlineagesofextantlungfishes.Thegenomeexpansionhappenedinearlylungfishes(around400–200Ma),andslowedduringthebreakupofGondwana(fromaround200 Matopresent)(ExtendedDataFig.3c).Independently,genomesizeincreasedinsalamandersintwoindependentwavesofDNA-repeatexpansion(Fig.3b,ExtendedDataFigs.3c,5).LINEsmakeupmuchoftherecentgenomegrowthofthelungfish(<15%divergence,around9%(4 Gb),alsoinanearlierburstinlungfishbutnotaxolotl)(ExtendedDataFig.5a).Becausemobilizedtransposableelementscaninterruptgenefunction,onemightspeculatethatsuchburstsofactivityoftransposableelementsmighthavecausednovelgenefunctions.Fig.3:Compositionofrepetitiveelementsinthelungfishgenome.a,Thepiechartsshowoverallcompositionofrepetitiveelementsfromunmaskedassembly(firsttransposableelementannotation)(left),togetherwiththeannotationfromthehardmaskedgenome(secondtransposableelementannotation)(right).Thebarchartshowsthelandscapeofmajorclassesoftransposableelements.Kimurasubstitutionlevel(%)foreachcopyagainstitsconsensussequenceusedasproxyforexpansionhistoryofthetransposableelements.Oldercopies(oldexpansion)accumulatedmoremutationsandshowhigherdivergencefromtheconsensussequences.RC, rolling-circletransposons;SINE,shortinterspersednuclearelement;TE,transposableelement.b,Principalcomponent(PC)analysisofcompositionofrepetitiveelements(LTR,LINE,SINE,DNAandunknown,filteredby80/80rule)ofvertebrates.FullsizeimageAlthoughsyntenicallyhighlyconserved,thelungfishgenomehasundergoneextremeexpansionthroughtheaccumulationoftransposableelements.Weperformedstandardrepeat-maskingproceduresonthe37-Gbgenomeassembly,whichidentified67.3%(24.65 Gb)asrepetitive(Fig.3a,SupplementaryTable4).Toourknowledge,thisisthehighestrepetitiveDNAcontentinagenomefoundin theanimalkingdom.Wetestedwhethertheremaining13 Gbofthegenomehavesignaturesofrepetitivenessthatareobscuredbygenomesizebyapplyingasecondroundofrepeatannotationonthehard-maskedgenome.Thisrevealedanadditional23.92%ofrepetitiveDNA(Fig.3a),whichwasmostlyclassifiedas‘unknown’(adding11%totheunknownportionofrepetitiveDNA)or‘LINE’(8.5%)(SupplementaryTables5,6).Intotal,around90%ofthelungfishgenomeisrepetitive,anditexpandedintwowaves(Fig.3a,ExtendedDataFig.5).Toinvestigatewhethertransposableelementsarestillactive,weanalysedpoly(A)-RNA-derivedRNA-seqdatathatprobablyrelatestoproteinsrelevantfortranspositionactivity.Allmajorcategoriesoftransposableelements(1,106outof1,821(60.7%))wereexpressed(ExtendedDataFig.6a).Transposableelementfamilieswithhighercopynumberswerealsohighlyexpressedinallthreetissueswetested.This,andthefindingofsimilarcopiesformanytransposableelementfamilies,suggeststhatseveraltypesoftransposableelementremainactiveandcontributetotheongoingexpansionofthelungfishgenome.Identificationofinsertionpolymorphismsbetweentwo,ideallyrelativelycloselyrelatedlungfishspecies(suchasProtopterusfromAfrica)arenecessarytoconfirmtransposableelementactivity.Apparently,thetransposonsilencingmachinerydidnotadapttoreduceoverabundanttransposableelementsbycopynumberexpansionorstructuralchanges(SupplementaryTable7).Therepeatlandscape(proportionsofmajorclassesoftransposableelement)oflungfishresemblestetrapods(includingaxolotl),whereasthethirdextantsarcopterygianlineage(thecoelacanths)ismore‘fish’-like(Fig.3b).Thetwolargestanimalgenomesyetsequencedexpandedthroughdifferenttemporaldynamics.Whereaslongterminalrepeat(LTR)elementsarethemostabundantclassoftransposableelement(59%)inaxolotl8,LINEs(25.7%;mostlyCR1andL2elements)dominateinlungfish(ExtendedDataFigs.5,6).Thesetworetrotransposonclassesbelongtothesamecopy-and-paste(andnotcut-and-paste)category,butpropagateviadifferentmechanisms17.Althoughglobalrepeatcompositionsdifferbetweenlungfishandaxolotl,thesameLTRclassaffectstheirgenicregions(ExtendedDataFig.6,SupplementaryInformation).Tofurtherunderstandgenomegrowthinlungfish,wecomparedthegenomestructureofN. forsteriwiththatofothergenomes(ExtendedDataFigs.6c,d,7).Althoughcompactgenomeshavesmallintrons,intragenicnoncodingregionsusuallyincreasewithgenomesize18.Thelargestintronofthelungfishis5.8 Mb(inthedmbt1gene)andaverageintronsizeis50 kbasinaxolotl,comparedto1 kbinfuguand6 kbinhuman.IntronsintheN. forsterigenomecompriseabout8 Gb(21%ofgenome)—asimilarproportiontothatinhuman(21%),buthalfthatoffugu(40%).Thissuggeststhatsimilarmechanismsaffectthegenicandintergeniccompartments,followingexpectationsforgenomesizeevolution19.Inmostgenes,thefirstintrontypicallyisthelargest.Thebiologicalrelevanceofthisremainsunclear.Thefirstintronsinlungfishandaxolotlarealsomuchlargerthandownstreamintrons(ExtendedDataFig.7),whichindicatesthattherelativelylargerfirstintronsinsmallergenomesareprobablynotduetothespacerequirementsofregulatoryorstructuralmotifs20.Ithaspreviouslybeensuggestedthatthesizeofintragenicnoncodingsequencesandtheextentofintronexpansionareassociatedwithorganismalfeatures(suchasmetabolicrate18)orfunctionalcategoriesofgene8(forexample,developmentalornondevelopmentalgenes).Similartoaxolotl8,theintronsindevelopmentalgenesinlungfisharesmallerthaninnondevelopmentalgenes(P = 2.166 × 10−8,Mann–WhitneyUtest)(SupplementaryTable8).Genomicpreadaptationsinfish–tetrapodtransitionPositiveselectionanalysisuncovered259 genes,manyofwhicharerelatedtooestrogenandcategoriesrelatedtofemalereproduction(SupplementaryInformation,SupplementaryTable9).Wecomparedtheseratedynamics(16,471 genefamilies)(SupplementaryTables10,11),andfoundthatinthelungfishlineage24 familieshavecontractedand107 familieshaveexpanded—possiblyrelatedtoevolutionaryinnovations.AirbreathingandtheevolutionoflungsAllland-livingvertebratesandadultlungfishareairbreathers.Thepulmonarysurfactantprotein Bfamilyofgeneshasexpandedconsiderablyinthelungfishgenome.Surfactantsarenecessarycomponentsofthelipoproteinmixturethatcoversthelungsurfaceandensuresproperpulmonaryfunction.Inlungfish,thenumberofsurfactantgenesincreasedtoanumbertypicalfortetrapods(2–3×morethanincartilaginousandbonyfish)(SupplementaryTable12).Thismayindicateanadaptationtoairbreathinginlungfish.Wefurtherinvestigatedtheexpressionofshh,whichencodesanimportantregulatoroflungdevelopment21,duringlungfishembryogenesis(ExtendedDataFig.8a).shhisstronglyexpressedinthedevelopinglungs(embryosatstages43–48),visualizingthedevelopmentoftheright-sidedlung(Neoceratodushasaunilaterallung).Thislungdevelopsinamannernotablysimilartothoseofamphibians22.Altogether,thishighlightsmolecularsignaturesoflungsthatwerenecessaryfortheconquestoflandbysarcopterygians.OlfactionandevolutionofthevomeronasalorganWealsonotedexpansionsofgenesinvolvedinolfaction.Thegenecomplementofreceptorsforairborneodorants(whichislargeandcomplexintetrapodsandsmallinfish)isconsiderablyexpandedinlungfish,whereasseveralreceptorclassesforwaterborneodourshaveshrunk—inparticular,zetaandetareceptors,whichaboundinteleostfishes(SupplementaryTable13).Thevomeronasalorgan(VNO)ispresentinmosttetrapods23,24,beinglinkedtopheromonereceptionandexpressingalargerepertoireofvomeronasalreceptorgenes(particularlyinamphibians).InN. forsteri,thevomeronasalreceptorgenefamily—knownfromfishandevenlampreys,althoughitsfunctioninthesespeciesisunknown—hasexpandedconsiderably.Lungfishpossessa‘VNOprimordium’25.Thenotableexpansionofthevomeronasalreceptorgenefamily(especiallyV2Rgenes)inN. forsteri(SupplementaryTable14)showsthattheVNOisatetrapodinnovation,whichemergedinthewater-to-landtransition.LobedfinsandevolutionofterrestriallocomotionSarcopterygianshaveelaboratedendochondralskeletons:lobedfinsthataredistallybranched,formingdigitsthataresuitableforsubstrate-basedlocomotion.Ouranalysisindicatessarcopterygianoriginsfor31 conservedtetrapodlimb-enhancerelements26(Fig.4a,ExtendedDataFig.8b).Thehs72(refs.27,28)enhancer(relatedtosall1)drivesautopodalexpression(Fig.4b).Wefoundsall1stronglyexpressedinlungfishembryos,inexpressionpatternssimilartothosereportedfortetrapods29(Fig.4b)butabsentduringzebrafishfindevelopment30.Similarfunctionsofsall1duringmouselimbdevelopment29suggestthatthisgenecontributedtotheacquisitionofsarcopterygianlobedfinsalreadyinlungfish.Fig.4:Regulatorypreadaptationoflobedfinandhoxdgeneregulation.a,Analysisof330 validatedmouseandhumanlimbenhancersshowsdeepevolutionaryoriginofthelimbregulatoryprogram;31enhancersareassociatedwiththeemergenceofthelobedfin.b,Thehs72enhancerlocatedneartheSall127,28genedrivesstrongLacZinmouseautopods(n = 3outof3 embryos,LacZ-stainedembryoscourtesyofVISTAenhancer26)(top).sall1isexpressedinasimilarautopodial-likedomaininlungfishpectoralfins(n = 2outof2 fins)(bottom).dpf,dayspost-fertilization.c,Left,hoxc13isexpressedinadistallungfishareathatoverlapswiththecentralmetapterygialaxis(sox9)andfinfold(and1)(arrowheads)(n = 2outof2 fins).Right,similarexpressionpresentinaxolotllimbs(arrowhead)(n = 4outof4 limbs),indicatingadeepsarcopterygianoriginforthisexpressiondomain.d,Duringlungfishfindevelopment,hoxd11andhoxd13areexpressedinmostlynonoverlappingproximalandposterior–distalfindomains(n = 4outof4 finseach).e,ThelungfishhoxdclusterhasincreasedinsizecomparedtomouseandXenopus,butmaybesmallerthantheaxolotlhoxdcluster.Inlungfishandaxolotlexpansionhasoccurredinthe3′and5′regionsofthecluster,whereasthecentralhoxd8,hoxd9,hoxd10andhoxd11region(lilacbox)remainedstableatapproximately25 kb,formingaseparate‘minicluster’.Thehoxdclusterisregulatedby3′andlong-rangeenhancers.hoxd9,hoxd10andhoxd11(lilac),andhoxd13(green),aresubjecttoenhancersharing33andco-expressedinthedistallimbinmouseandXenopus33,37,whereastheincreasedgenomicdistancebetweenhoxd13andhoxd9,hoxd10andhoxd11hasdisruptedtheirco-expressioninthedistalappendagesoflungfishandaxolotl.Thepreservedclusteringofhoxd8,hoxd9,hoxd10andhoxd11canbeexplainedbyenhancersharing3′ofthecluster33,whichprobablyplacesconstraintsontheirintergenicdistances.AxolotlandXenopushoxd11 andhoxd13afterref.37;lungfishhoxd11 andhoxd13domainsafterref.36andd(SupplementaryTable16listsprimersforprobes).Scalebars,0.2 mm.Silhouettesarefromref.36.FullsizeimageHoxclustersandtefin-to-limbtransitionThe4 clustersofhoxgenesinNeoceratodus(hoxa,hoxb,hoxcandhoxd)comprise43 genes(ExtendedDataFig.9);thepresenceofhoxb10andhoxa14inlungfishconfirmstheirlossatthefish-to-tetrapodtransition11.OurRNA-seqanalysisoftheexpressionofhoxgenesinthefinsoflarvalNeoceratodus(ExtendedDataFig.8c)showedanunexpectedexpressionofhoxcgenes.Theexpressionofhoxcgenesinpairedfinsorlimbshaspreviouslybeenreportedonlyformammals31,relatedtothenailbed.Weobservedhoxc13expressioninaxolotllimbs(Fig.4c),butitwasabsentinthepectoralfinsofray-finnedfish(ExtendedDataFig.8d).TranscriptlocalizationinNeoceratodusembryosshowedexpressionofhoxc13inthedistalfin(Fig.4c).Thisindicatesanearlygainofhoxc13expressioninsarcopterygians,suggestingco-optionofthisdomainintetrapodstopatterndermallimbelements(suchasnails,hoovesandclaws).Togetherwithsall1,thisdemonstratesanearlysarcopterygianoriginoflimb-likegeneexpressionthatwasreadyfortetrapodco-option,facilitatingthefin-to-limbtransitionandcolonizationoftheland.HoxclusterexpansionversusregulationConsistentwiththeoverallgenomeexpansion,thehoxclustersofNeoceratodusarelargerthaninmouse,chickenandXenopus,buthaveanunevenpatternofexpansion(ExtendedDataFig.9).Theclusteringofhoxdgenesresultsintheircoregulationbyenhancers3′and5′ofthecluster,leadingtoco-expressionofhoxd9,hoxd10,hoxd11,hoxd12andhoxd13inthedistalappendages32,33,34,35.DuringfindevelopmentinNeoceratodus,expressionofhoxd11isnearlyabsentfromthehoxd13territory36(Fig.4d)whereasinaxolotlhoxd9,hoxd10andhoxd11areexcludedfromthehoxd13digitdomain37(ExtendedDataFig.8e).Suchapparentlossofcoregulationbetweenhoxd13andhoxd9,hoxd10andhoxd11issimilartothatcausedbyexperimentallyincreaseddistancesinthehoxdcluster32,andsuggestsadisruptionofenhancersharingcausedbytheexpansionoftheintergenicregionsbetweenhoxd11 andhoxd13(Fig.4e).Weperformedadditionalanalysesinmouse,Xenopus,lungfishandaxolotl,whichshowedthat—despite5–10×differencesinthesizeofthehoxdcluster—theregioncomprisinghoxd8,hoxd9,hoxd10andhoxd11remainedfixedataround25 kb(Fig.4e).Thisapparentconstraintisprobablyduetosharingofenhancerslocatedatthe3′endofthecluster33.Altogether,thisindicatesthathoxdexpansionhaspartiallydisruptedlong-rangeenhancersharing,butthat—conversely—suchmechanismshavelocallyalsoconstrainedintergenicdistances.Wehavesequencedandassembledatthechromosomelevel(SupplementaryTable15)thelargestanimalgenome,andhavesubstantiatedthehypothesisthatlungfisharetheclosestlivingrelativesoftetrapods.Despitetheuniquegenomeexpansionhistoryoflungfish,genicorganizationandchromosomalhomologyismaintainedevenatthelevelofmicrochromosomes.Genomicpreadaptationsinlungfishforthewater-to-landtransitionofvertebratesincludealargercomplementoflung-expressedsurfactantgenes,whichmighthavefacilitatedtheevolutionofair-breathingthroughalung.Inaddition,thenumberofVNOolfactoryreceptors(aswellasotherreceptorgenefamiliesthatpermitdetectionofairborneodours)increasedinthelineagethatledtoair-breathinglungfish.Theunevenexpansionofhoxclustersdemonstratestheregulatoryconsequencesof,andconstraintson,genomeexpansion.Theevolutionarytrajectoryoflimbenhancersshowsanearly-fishoriginofthelimbregulatoryprogram,withimportantchangestowardspreadaptationsforterrestrializationprecedingthefin-to-limbtransition.Geneexpressiondomainsthatcharacterizethetetrapodlimb,butwhichwerepreviouslypresumedtobeabsentfromfins(suchasthoseofsall1andhoxc13),appearedinthelobe-finnedlineage.Suchnoveltiesmighthavepredisposedthesarcopterygianstoconquertheland,demonstratinghowthelungfishgenomecancontributetoabetterunderstandingofthismajortransitioninvertebrateevolution.MethodsNostatisticalmethodswereusedtopredeterminesamplesize.Theexperimentswerenotrandomizedandinvestigatorswerenotblindedtoallocationduringexperimentsandoutcomeassessment.BiologicalmaterialsBiopsymaterialforDNAandRNAisolationwasobtainedfromajuvenileAustralianlungfish(N. fosteri)importedfromAustralia(CITESpermitno.:PWS2017-AU-000242).Owingtotheimmaturestatusofthegonad,thesexcouldnotbedetermined.Thesamespecimenwasusedforgenomesequencing(muscle),constructionoftheHi-Clibrary(spleen)andtranscriptomesequencingofbrain,gonadandliver.Thesecondsetofreadswasgeneratedfromlungfishembryos(embryonicstage52,GenBankaccessionnumbersSRR6297462–6297470)36.EmbryoswerebredandcollectedunderpermitARA2009.039atMacquarieUniversity.DNAextraction,genomesequencingandassemblyHighmolecularweight(HMW)andultra-HMWDNAwaspreparedbyFutureGenomicsandNextomics,andsequencedusingNanoporetechnology(forstatistics,seeSupplementaryTable1).gDNAforgenomecorrectionfromsnap-frozenlungfishmuscletissue(0.3g)wasisolatedbyastandardgDNAisolationprotocol.LibrarypreparationwasperformedusingtheWestburgNGSDNAlibrarykit.ThefinallibrarywasexcisedbyPippinprepwith400-bpDNAsizeandsequenced(IlluminaNova-seqS2;PE150)atViennaBioCenterNGSfacility.Hi‐Clibrarywasgeneratedaspreviouslydescribed38,39,withmodificationsdetailedinSupplementaryMethods.FinalHi‐Clibrariesweresequenced(IlluminaNova-seqSP;PE150)atViennaBioCenterNGSfacility.GenomeassemblyNinety-sixmillionreadscomprising1.2TbwereassembledusingtheMARVELgenomeassembler8.Wefirstaligned1%ofthereadsagainstallotherreads.Fromthese1%-against-allalignments,wederivedinformationontherepetitiveelementspresentinthereadsandusedtransitivetransfertorepeat-annotateallreadsusedintheassembly.Regionsweredeemedrepetitivewhenthedepthofthealignmentsforagivenreadexceededtheexpecteddepthfourfold.Giventhealignmentofthe1%againsteveryotherreadintheassembly,wethentransferredtherepeatannotationofthe1%usingthealignmentstotherespectivepositioninthealignedreads.Here,theassumptionisthatwhenregion(a,b)inreadAalignsto(c,d)inreadBandfora≤ rb≤ re≤ b(inwhichrbandrearerepetitiveelements);thisthancanbemappedusingthealignmenttoacorrespondingregioninB,whichthencanbetaggedasrepetitiveaswell.Thefinalrepeat-maskingtrackcovered28.7%ofthe1.2Tb.Wethenprocessedwithanall-against-allalignmentwithrepeatmaskinginplace,yieldingfivebillionalignments.Onthebasisofthesealignments,wederivedreadqualitiesat100-bpresolution,highlightinglowsequencingqualityregionsinthereads.Usingthealignmentsandthereadqualitiesstructuralweaknesses(chimericbreaks,high-noiseregionsandothersequencingartefacts)inthereadswererepaired(SupplementaryMethods,ExtendedDataFig.10).Repairedreadswerethenusedforanewroundofalignments,againwithrepeatmasking,inplace.Afteralignment,thedefaultMARVELassemblypipelineproceededasshownintheincludedexamplesofthesourcedistribution(ExtendedDataFig.1).ForthecurrentMARVELsourcecoderepository,seehttps://github.com/schloi/MARVEL.Forsampleexecutionscripts,seehttps://github.com/schloi/MARVEL/tree/master/examples.ScaffoldingWeusedanagglomerativehierarchical-clustering-basedscaffoldingapproachusingvariousnormalizations(ExtendedDataFig.1).Fordetails,seeSupplementaryMethods.Wecreatedinitialclustersbyselectingthelargestcontigswiththefewestcontactsbetweenthem,eachcontigservingasasinglecluster.Wethenaddedcontigsonthebasisofuniqueassignabilitytoclusters.Thiswasfollowedbyscaffoldingtheclusterseparately,visualinspectionofanapproximatecontactmapderivedduringthescaffoldingprocessandreturnofwronglyassignedcontigstothesetofunassignedcontigs.Wecreatedcontactmapsforallclustersandmergedorsplitclustersonthebasisofthesignalwithinthose.Theprocessofassigningcontigs,scaffolding,mergingandsplittingclusterswasrepeateduntilnomoreusefulchangescouldbemadetotheclusters(SupplementaryTable15forcomparisonofchromosomeandscaffoldDNAcontent).Forthepublicsourcecoderepository,seehttps://github.com/schloi/MARVEL/.TheMARVELassemblerandscaffolderhaspreviouslybeenusedtoobtainachromosome-scaleaxolotlgenomeassembly,whichhasbeenvalidatedincomparisontothepreviouslypublishedchromosome-scalemeioticscaffolding40andisavailableaspreviouslydescribed41.GenomeassemblycorrectionForcorrectionoferrors(insertionsand/ordeletions(indels),basesubstitutionsandsmallgaps)remainingafterthegenomeassembly,weappliedatwo-stepprocedureusingDNA-sequencingandRNA-seqreadsseparately.Inbrief,wesequencedthesamegenomicDNAsampleandgenerated4,693,324,032high-qualityreadpairs(2 × 150bp)(30×coverage).Additionally,weusedtheRNA-seqreadsfromthedenovotranscriptomeassemblytocorrectindels,butnotbasesubstitutions,intranscribedregions(SupplementaryMethods,SupplementaryResults,ExtendedDataFig.10).TranscriptomeassemblyRNAwasisolatedfrombrain,spinalcord,eyes,gut,gonad,liver,jaw,gills,pectoralfin,caudalfin,trunkmusclesandlarvalfin.LibrarieswereconstructedusingNEBNextUltraIIDirectionalRNAlibrarypreparationkit(NewEnglandBiolabs),IlluminaTruSeqRNAsamplepreparationkit(Illumina)orLexogenTotalRNA-seqLibraryPrepKitV2(Lexogen).Paired-endsequencing,performedwithIlluminaplatforms,yieldedapproximately1,150millionrawreads.Rawreads,filteredandcorrectedusingTrimmomaticv.0.3642andRCorrectorv.1.0.243,wereassembledusingdenovoandreference-guidedapproaches.Fordenovoassembly,onlyreadsderivedfrompoly(A)-selectedRNAwereprocessedusingtheOysterRiverProtocol(ORP)v.2.2.844.Inbrief,readswereassembledusingTrinityv.2.8.4(k-mer = 25),SPAdesv.3.13.345(k-mer = 55),SPAdes(k-mer = 75)andTrans-Abyssv.2.0.146(k-mer = 32).ThefourdifferentassemblieswerethenmergedusingtheOrthoFusermodule47,48implementedinORP.Completenessofthedenovo-assembledtranscriptomewasassessedwithBUSCOv.349usingcorevertebrategenesandVertebratagenes(vertebrata_odb9database)inthegVolantewebserver50.Forreference-guidedassembly,allreadswerealignedtotheN.forsterigenome(eachsampleindependently)usingtheprogramHISAT2v.2.1.051(maximumintronlengthsetto3Mb).TheresultingmappingfileswereparsedbyStringTiev.1.3.652andtranscriptsreconstructedfromeachalignedsampleweremergedinasingleconsensus.gtffile.RepeatsandtransposableelementsannotationNeoceratodusforsterirepeatsequenceswerepredictedusingRepeatMasker(v.4.0.7)withdefaulttransposableelementDfamdatabaseandadenovorepeatlibraryconstructedusingRepeatModeler(v.1.0.10),includingtheRECON(v.1.0.8),RepeatScout(v.1.0.5)andrmblast(v.2.6.0),withdefaultparameters.TransposableelementsnotclassifiedbyRepeatModelerwereanalysedusingPASTEC(https://urgi.versailles.inra.fr/Tools/)andDeepTE53.RepeatsequencesofA. mexicanum(AmexG_v3.0.0,https://www.axolotl-omics.org/)werepredictedusingthesameapproach.RepetitivesequencesofAnoliscarolinensis(GenBankaccessionGCA_000090745.2),Xenopustropicalis(GCA_000004195.4),Rhinatremabivittatum(GCA_901001135.1),Latimeriachalumnae(GCA_000325985.2),Lepisosteusoculatus(GCA_000242695.1),Daniorerio(GCA_000002035.4)andAmblyrajaradiata(GCF_010909815.1))wereidentifiedusingDfamTEToolsContainer(https://github.com/Dfam-consortium/TETools)includingRepeatModeler(v.2.0.1)andRepeatMasker(v.4.1.0).Tofurtherexaminetheremainingintergenicsequences,wepredictedrepetitivesequencesagainusingthesameworkflowonthegenomehard-maskedwithrepeatsalreadypredictedbyRepeatMasker.Kimuradistance-baseddistributionanalysisandtransposable-element-compositionprincipalcomponentanalysisKimurasubstitutionlevelsbetweentherepeatconsensustoitscopieswerecalculatedusingautilityscriptcalcDivergenceFromAlign.plbundledinRepeatMasker.RepeatlandscapeplotswereproducedwiththeRscriptnf_all_age_plot.Randnf_am_rb_age_plots.R,usingthedivsumoutputfromcalcDivergenceFromAlign.pl.PrincipalcomponentanalysisonrepetitiveelementcompositionwasperformedinR(v.3.6)usingfactoextrapackage(v.1.0.6).Repetitiveelementcompositions(SINE,LINE,DNA,LTRandunknown)werecalculatedfromthepredictedlibraries.Repetitiveelementcopieswerefilteredbythe80/80rule(equalorlongerthan80bp,equalormorethan80percentidentitycomparedwiththeconsensussequence).Repetitiveelementcompositionofothervertebrateswasobtainedfromref.54.TransposableelementcompositionbygenelengthandLTRfamilyanalysisRepetitivesequencecompositionwithingenes(groupedbylength)wasexaminedbycalculatingthecoverage(inbp)ofeachclassofrepetitiveelement,normalizedbygenelength.WeexaminedLTRfamilyenrichmentingenicregions.Allcalculationsandvisualizationsaresummarizedinthejupyternotebookfilete_general_analysis.ipynb.AllpythonscriptsranonPython≥3.7andusedthepackagegffutils(v.0.10.1)(https://github.com/daler/gffutils)tooperatelargegeneandrepetitiveelementannotationfilesfromlargegenomes.PlotsweregeneratedusingPlotlyPythonAPI(https://plot.ly).TransposableelementcontentingenicregionsIntronpositionwascalculatedbyGenomeTools(v.1.5.9).Thesumofthecoverageoftherepetitiveelement(forexample,LINECR1)wasnormalizedbythelengthofthegenicfeatureconsidered(SupplementaryTable17)(forexample,intron8)usingpythonscriptte_cnt_class.py.TransposableelementexpressionTransposableelementexpressionwasassessedwithTEtools55ongonad,brainandliverpoly(A)-RNAdata.Becauseofthelargesizeoflungfishgenome,arandomsubsetof10%ofalltransposableelementcopieswasused.Transposable-element-familycountswerenormalizedbytransposable-element-familyconsensuslength(count × 106/consensuslength)andlibrarysize.Normalizedcountswereplottedagainsttransposable-element-familycopynumbers.Annotationofprotein-codinggenesProtein-codinggeneswerepredictedbycombiningtranscriptandhomology-basedevidence.Fortranscriptevidence,assembledtranscripts(asdescribedin‘Transcriptomeassembly’)weremappedtotheassemblyusingGmaplv.2019-05-1256andthegenestructurewasinferredusingthePASApipelinev.2.2.357.ExpressionofeachtranscriptwasmeasuredusingthewholeRNA-seqdataset(asdescribedin‘Transcriptomeassembly’)andthepseudoalignmentalgorithmimplementedinKallistov.0.46.158.Forhomologyevidence,wecollectedmanuallycuratedproteinsfromUniProtKB/SWISSPROTdatabase(UniProtKB/Swiss-Prot2020_03)59andproteinsequencesofCallorhinchusmilii,L.chalumnae,L. oculatusandX.tropicalisfromEnsembl(http://www.ensembl.org)andNCBI(https://www.ncbi.nlm.nih.gov/genome),andalignedthemtotherepeat-maskedassemblyusingExoneratev.2.260.Transcriptandhomology-basedevidencewerethencombinedbyprioritizingtheformer(homology-basedpredictedgeneswereremovedwhenintersectingagenepredictedusingthereconstructedtranscripts).Thecombinedgenesetwasthenprocessedbytworoundsof‘PASAcompare’toadduntranslatedregion(UTR)annotationsandmodelsforalternativelysplicedisoforms.Low-qualitygenemodelswereremovedbyapplyingthreefurtherquality-filteringstepsinaniterativefashion:(1)single-exongeneswereretainedonlywhennosimilaritywithexonsofmulti-exonicgeneswasfound(similaritywasidentifiedwiththeglsearch36moduleimplementedintheFASTAv.36.3.8gpackage61withe-valuecut-offsof1 × 10−10andidentitycu-toffsof80);(2)genesintersectingrepeatelementswereremovedwhen>50%(single-exonicgenes)and>90%(multi-exonicgenes)werecoveredbyrepeats;and(3)geneswithinternalstopcodon(s)wereremoved.Thecompletenessofthepredictedprotein-codinggenesetwasassessedwithBUSCOusingthecorevertebrategenesandtheVertebratagenes(vertebrata_odb9database)inthegVolantewebserver.Toannotatethelungfishhoxclusters,hoxgeneswerefirstidentifiedusingBLASTwithvertebrateorthologuesasquery(SupplementaryMethods).AnnotationofncRNAgenesncRNAgeneswereannotatedusingtRNAscan-s.e.v.2.0.362andInfernalv.1.1.263.Thesameprocedurewasappliedtothegenomesofthenineotherfocalspecies.Foreachofthetenspecies,thecorrespondingmicroRNAsets(obtainedfrommiRBasev.2264database)wereusedtopredictmicroRNAtargetsiteson3′UTRsofcanonicalmRNAsusingmiRandav.3.365.FurtherdetailsareprovidedinSupplementaryInformation.AnnotationofconservednoncodingelementsWhole-genomealignmentsThemaskedversionsofthegenomeassembliesofthetenspeciesusedforthephylogenetictree(Fig.1)wereusedtobuildawhole-genomealignmentwiththehumangenomeasreference(ten-waywhole-genomealignment).Inbrief,eachpairwisealignmentwasconstructedusingLastzv.1.03.7366andfurtherprocessedusingUCSCGenomeBrowsertools67.Multiplealignmentsweregeneratedusingasinputtheninepairwisealignmentsin.mafformatwiththeprogramsMultizv.11.2andRoast.v.3.068.DetectionofconservedelementsThephylogenetichiddenMarkovmodel(phylo-HMM)implementedinphastCons69(runinrho-estimationmode)wasusedtopredictaconsistentsetofconservedgenomicelementsintheten-specieswholegenomealignment.AneutralmodelofsubstitutionswascalculatedusingphyloFit69withthegeneralreversiblesubstitutionmodelfromfourfolddegeneratesites.Rawconservednoncodingelements(CNEs)detectedbyphastConsweremergedwhentheirdistancewas<10bp,andsubsequentlyCNEs<50bpwereremoved.Protein-codingCNEsandthoseintersectingncRNAgenes,pseudogenes,retrotransposedelementsandantisensegenes(annotatedinthehumangenome)wereremoved.ExpansionofthegenomeinintergenicregionsThefinalfilteredsetofCNEswasusedtoinvestigateexpansionofintergenicspaces.Wecomparedthedistanceofnonexonicelementsthatareconservedinlungfishandthreetetrapods(human,chickenandaxolotl).ToobtaininformativeCNEpairs,weselectedthoseCNEsthat:(1)werepresentinallfourgenomes;(2)werelocatedinintergenicspace;(3)werelocatedinthesamecontigorchromosomeineachspecies;and(4)didnothaveageneinbetweenthem.Theremainingsetof223CNEpairswereusedtocalculateintergenicdistanceandregion-specificexpansionofthelungfishgenome(SupplementaryTable18).Lineage-specificaccelerationofCNEsTheprogramphyloPwasusedtotesteachCNEforlineage-specificacceleratedevolution69,70inthelungfishbranch.AlikelihoodratiotesttocomputethePvalueofaccelerationwithrespecttoaneutralmodelofevolutionforeachoftheconservedelementsinthealignmentwasused.CNEsshowingfalse-discovery-rate(FDR)-adjustedP values 80%)trimmedwithBMGE75.Orthologywasensuredbymanualinspectionofmaximumlikelihoodgenetrees(IQ-TREE)andalignments(MAFFTginsi)forlocishowinghighbranch-lengthdisparity,andfiveindividualsequenceswereremoved.Lociwereconcatenatedintoafinalmatrixcontaining10taxaand697loci,totalling383,894alignedaminoacidpositions,ofwhich208,588(54%)werevariable.PhylogenywasinferredusingPhyloBayesMPIv.1.776underthesite-heterogeneousCAT-GTRmodel,showntoavoidphylogeneticartefactswhenreconstructingbasalsarcopterygianrelationships4.TwoindependentMarkovchainMonteCarlochainswererununtilconvergence(>4,000cycles),assessedaposterioriusingPhyloBayes’built-infunctions(maxdiff = 0,meandiff = 0,ESS>100forallparameters,afterdiscardingthefirst25%cyclesasburn-in).Post-burn-intreesweresummarizedintoafullyresolvedconsensustreewithposteriorprobabilitiesof1forallbipartitions.Whole-genome-alignment-basedphylogenyTheten-specieswholegenomealignmentwasprocessedbyMafFilterv.1.3.077tokeeponlyalignmentblocks>300bpthatwerepresentinallspecies.Filterednoncodingblockswerethenconcatenatedandexportedin.phylipformat.PoorlyalignedregionswereremovedusingtrimAlv.1.2withoption‘-automated1’.Thefinaldataset(99,601alignednucleotides)wasusedtoreconstructthephylogenywithRAxMLv.8.2.4undertheGTRGAMMAmodeland1,000bootstrapreplicates.GenomesizeevolutionGenomesizeevolutionwasmodelledbymaximumlikelihoodusingthe‘fastAnc’functioninthephytoolsRpackage78.Weusedatime-calibratedtreerepresentingallmajorjawedvertebratelineagesobtainedfromthephylotranscriptomictreeofref.5;agesareagenome-wideestimatesacross100time-calibratedtreesinferredfrom100independentgenejackknifereplicatesinferredinPhyloBayesv.4.179underalog-normalautocorrelatedclockmodelwith16cross-validatedfossilsasuniformcalibrationswithsoftbounds,theCAT-GTRsubstitutionmodelandabirth–deathtreeprior.Genomesizedata(haploidDNAcontentorc-value)wereobtainedfromref.80.Genomesizeestimateswereaveragedperspecies(ifseveralwereavailable)and,insixspecies,genomesizewasapproximatedastheaverageofcloselyrelatedspecieswithinthesamegenera.ForNeoceratodus,thek-mer-basedestimationwasused(43Gb;c-value = 43.97pg).Ancestralgenomesizeswereusedtocalculatetheratesofgenomeevolutionforselectedbranches.MolecularclockanalysesDivergencetimeswereinferredwitharelaxedmolecularclockwithautocorrelatedrates,asimplementedinMCMCTreewithinthePAMLpackagev.4.9h81.Atotalofsixfossilcalibrationswereusedasuniformpriors82.Forfurtherdetails,seeSupplementaryMethods.DynamicsofgenefamilysizeCAFE83wasusedtoinfergenebirthanddeathrates(lambda)andretrievegenefamiliesundersignificantdynamics.Asinput,wetookthespeciestreewithdivergencetimefromtheoutputofMCMCTreeandtheresultsofgeneclustersfromHcluster_sg.Eachgeneclusterwasdeemedtobeagenefamily.WeranCAFEunderamodelinwhichagloballambdawassetacrossthewholetree.Tosymbolizeeachgenefamily,wetookthelongestmemberasrepresentativeandBLAST-searchedwithdiamond84againstSWISSPROTandNRdatabases.Thebesthitfrombothwasretained.Tocomparetherepertoireofolfactoryreceptors,tastereceptorsandpulmonarysurfactantproteinsacrossallstudiedspecies,wefollowedthesameprocedureforeachspecies.First,wecollectedsequencesofolfactoryreceptors,tastereceptorsandpulmonarysurfactantproteinsfromSwiss-ProtandNRdatabaseasquery.ForsequencesfromNRdatabase,weonlykeptthosewithidentifiersstartingwith‘NP_’,whicharesupportedbytheRefSeqeukaryoticcurationgroup.Second,wemappedthequerysettoeachgenomeusingExonerateinservermodel(maxintronsettosixmillionforlungfishandaxolotl).Thealignmentwasextendedtostartandstopcodonwhenpossible.Third,weBLAST-searchedallretrievedsequencestoNRdatabaseandremovedthosewithabesthitthatwasnotanolfactoryreceptor,tastereceptororpulmonarysurfactant.Thefinalresultsequenceshadalignmentcoveragerangingfrom32%to100%(firstquartile95%),andpercentageofidentityfrom17%to100(firstquartile62%)toitsquery.Followingapreviousstudy85,weseparatedthefinalsequencesintothreecategoriesonthebasisoftheiralignmenttotheirquery:(1)pseudogene,sequenceswithprematurestopcodonorframeshift;(2)truncatedgene,sequenceswithoutprematurestopcodonandframeshiftbutbrokenopenreadingframe(ORF)(startorstopcodonmissing);and(3)intactgene,sequenceswithintactORF.PositiveselectionanalysisTwomodelswerecalculated.Model1wasusedtofindgenespositivelyselectedinlungfishandmodel2wasusedforgenescommonlypositivelyselectedintetrapodsandlungfish.GenomesincludedwereN. forsteriandA. mexicanumfromthisstudy,andtheEnsemblgenomesD.rerio(Danio_rerio.GRCz11),A.carolinensis(Anolis_carolinensis.AnoCar2.0),L.oculatus(Lepisosteus_oculatus.LepOcu1),L.chalumnae(Latimeria_chalumnae.LatCha1),C.milii(Callorhinchus_milii.Callorhinchus_milii-6.1.3),X.tropicalis(GCF_001663975.1_Xenopus_laevis_v2),G.gallus(Gallus_gallus.GRCg6a)andH.sapiens(Homo_sapiens.GRCh38).TheX.tropicalisgenome(GCF_001663975.1_Xenopus_laevis_v2)wasdownloadedfromNCBI.ProteinandcDNAfilesfromallspeciesweredownloaded.Toidentifyorthologousproteins,allproteinsequenceswerecomparedtolungfishusingInparanoid86(defaultsettings).TomatchproteinandcDNA,sequencesweresearchedbyTBLASTNandonly100%hitswerekept.CodonalignmentsfortheproteinandcDNAsequencepairswereconstructedusingpal2nalv.1487.ResultingsequenceswerealignedbyMUSCLE88(option:-fastaout)andpoorlyalignedpositionsanddivergentregionsofcDNAwereeliminatedbyGblocksv.0.91b89(options:-b410-b5n --b35 --t = c).Anin-housescriptwasusedtoconverttheGblocksoutputtoPAMLformat.Asaphylogenetictree,wetookthespeciestreewithdivergencetimesfromMCMCTreeasinputfordetectionofpositiveselectionwithC. miliiasoutgroup.Forthephylogeneticanalysesbymaximumlikelihood,the‘EnvironmentforTreeExploration’(ETE3)toolkit90—whichautomatesCodeMLandSlranalysesbyusingpreconfiguredevolutionarymodels—wasused.Fordetectionofgenesunderpositiveselectioninlungfish,wecomparedthebranch-specificmodelbsA1(neutral)withmodelbsA(positiveselection)usingalikelihoodratiotest(FDR≤0.05).Todetectsitesunderpositiveselection,naiveempiricalBayesprobabilitiesforallfourclasseswerecalculatedforeachsite.Siteswithaprobability>0.95foreithersiteclass2a(positiveselectioninmarkedbranchandconservedinrest)or2b(positiveselectioninmarkedbranchandrelaxedinrest)wereconsidered.Twomodelswerecalculated.Inmodel1,onlythebranchforlungfishwasmarked;inmodel2,alltetrapodsandlungfishweremarkedforpositiveselection.FunctionalclusteringwasdonewithIPA(Qiagen,www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis)andDAVID(https://david.ncifcrf.gov/home.jsp)usinghumanhomologueswithdefaultsettings.InsituhybridizationInsituhybridizationwasperformedaspreviouslydescribed36,91,withmodifications(SupplementaryMethods).hoxgeneRNA-seqanalysishoxgeneRNA-seqanalysiswasperformedonastage-52lungfishlarvaRNA-seqdataset(SRR6297462–SRR6297470)39(SupplementaryMethods).LimbenhanceranalysisThreehundredandthirtynonredundantVISTAenhancerelements26,92weresearchedbyBLASTNagainstX.laevis,X.tropicalis,Nanoranaparkeri,axolotl,reedfish,sterlet,gar,elephantshark,coelacanth(LatCha1)andNeoceratodusgenomestodetermineconservation(SupplementaryMethods).ReportingsummaryFurtherinformationonresearchdesignisavailableinthe NatureResearchReportingSummarylinkedtothispaper. Dataavailability DataareavailablefromNCBIBioprojectunderaccessioncodePRJNA644903.Allotherrelevantdataareavailablefromthecorrespondingauthorsuponreasonablerequest. Codeavailability Customcodehasbeendepositedathttps://github.com/labtanaka/meyer_lungfish.ForthecurrentMARVELsourcecoderepository,seehttps://github.com/schloi/MARVEL.Forsampleexecutionscripts,seehttps://github.com/schloi/MARVEL/tree/master/examples. References1.Clack,J.,Sharp,E.&Long,J.inTheBiologyofLungfishes(edsJorgensen,J.M.&Joss,J.)1–42(CRC,2011).2.Kemp,A.ThebiologyoftheAustralianlungfish,Neoceratodusforsteri(Krefft1870).J.Morphol.190,181–198(1986). GoogleScholar  3.Carroll,R.L.VertebratePaleontologyandEvolution(W.H.Freeman,1988).4.Irisarri,I.&Meyer,A.Theidentificationoftheclosestlivingrelative(s)oftetrapods:phylogenomiclessonsforresolvingshortancientinternodes.Syst.Biol.65,1057–1075(2016).PubMed  GoogleScholar  5.Irisarri,I.etal.Phylotranscriptomicconsolidationofthejawedvertebratetimetree.Nat.Ecol.Evol.1,1370–1378(2017).PubMed  PubMedCentral  GoogleScholar  6.Krefft,G.DescriptionofagiantamphibianalliedtothegenusLepidosirenfromtheWideBaydistrict,Queensland.Proc.Zool.Soc.Lond.1870,221–224(1870). GoogleScholar  7.Gunther,A.XIX.DescriptionofCeratodus,agenusofganoidfishes,recentlydiscoveredinriversofQueensland,Australia.Phil.Trans.R.Soc.B161,511–571(1871).ADS  GoogleScholar  8.Nowoshilow,S.etal.Theaxolotlgenomeandtheevolutionofkeytissueformationregulators.Nature554,50–55(2018).CAS  PubMed  ADS  GoogleScholar  9.Rock,J.,Eldridge,M.,Champion,A.,Johnston,P.&Joss,J.KaryotypeandnuclearDNAcontentoftheAustralianlungfish,Neoceratodusforsteri(Ceratodidae:Dipnoi).Cytogenet.CellGenet.73,187–189(1996).CAS  PubMed  GoogleScholar  10.Pedersen,R.A.DNAcontent,ribosomalgenemultiplicity,andcellsizeinfish.J.Exp.Zool.177,65–78(1971).CAS  PubMed  GoogleScholar  11.Amemiya,C.T.etal.TheAfricancoelacanthgenomeprovidesinsightsintotetrapodevolution.Nature496,311–316(2013).CAS  PubMed  PubMedCentral  ADS  GoogleScholar  12.Fox,D.T.,Soltis,D.E.,Soltis,P.S.,Ashman,T.-L.&VandePeer,Y.polyploidy:abiologicalforcefromcellstoecosystems.TrendsCellBiol.30,688–694(2020).CAS  PubMed  GoogleScholar  13.Vervoort,A.TetraploidyinProtopterus(Dipnoi).Experientia36,294–296(1980). GoogleScholar  14.Putnam,N.H.etal.Theamphioxusgenomeandtheevolutionofthechordatekaryotype.Nature453,1064–1071(2008).CAS  PubMed  ADS  GoogleScholar  15.Simakov,O.etal.Deeplyconservedsyntenyresolvesearlyeventsinvertebrateevolution.Nat.Ecol.Evol.4,820–830(2020).PubMed  PubMedCentral  GoogleScholar  16.Braasch,I.etal.Thespottedgargenomeilluminatesvertebrateevolutionandfacilitateshuman–teleostcomparisons.Nat.Genet.48,427–437(2016).CAS  PubMed  PubMedCentral  GoogleScholar  17.Jurka,J.,Kapitonov,V.V.,Kohany,O.&Jurka,M.V.Repetitivesequencesincomplexgenomes:structureandevolution.Annu.Rev.GenomicsHum.Genet.8,241–259(2007).CAS  PubMed  GoogleScholar  18.Zhang,Q.&Edwards,S.V.Theevolutionofintronsizeinamniotes:aroleforpoweredflight?GenomeBiol.Evol.4,1033–1043(2012).PubMed  PubMedCentral  GoogleScholar  19.Lynch,M.&Conery,J.S.Theoriginsofgenomecomplexity.Science302,1401–1404(2003).CAS  PubMed  ADS  GoogleScholar  20.Bradnam,K.R.&Korf,I.Longerfirstintronsareageneralpropertyofeukaryoticgenestructure.PLoSONE3,e3093(2008).PubMed  PubMedCentral  ADS  GoogleScholar  21.Kugler,M.C.,Joyner,A.L.,Loomis,C.A.&Munger,J.S.Sonichedgehogsignalinginthelung.Fromdevelopmenttodisease.Am.J.Respir.CellMol.Biol.52,1–13(2015).PubMed  PubMedCentral  GoogleScholar  22.Rankin,S.A.etal.AmolecularatlasofXenopusrespiratorysystemdevelopment.Dev.Dyn.244,69–85(2015).CAS  PubMed  GoogleScholar  23.Døving,K.B.&Trotier,D.Structureandfunctionofthevomeronasalorgan.J.Exp.Biol.201,2913–2925(1998).PubMed  GoogleScholar  24.Syed,A.S.,Sansone,A.,Hassenklöver,T.,Manzini,I.&Korsching,S.I.CoordinatedshiftofolfactoryaminoacidresponsesandV2Rexpressiontoanamphibianwaternoseduringmetamorphosis.Cell.Mol.LifeSci.74,1711–1719(2017).CAS  PubMed  GoogleScholar  25.Nakamuta,S.,Nakamuta,N.,Taniguchi,K.&Taniguchi,K.Histologicalandultrastructuralcharacteristicsoftheprimordialvomeronasalorganinlungfish.Anat.Rec.(Hoboken)295,481–491(2012). GoogleScholar  26.Visel,A.,Minovitsky,S.,Dubchak,I.&Pennacchio,L.A.VISTAenhancerbrowser—adatabaseoftissue-specifichumanenhancers.NucleicAcidsRes.35,D88–D92(2007).CAS  PubMed  PubMedCentral  GoogleScholar  27.Pennacchio,L.A.etal.Invivoenhanceranalysisofhumanconservednon-codingsequences.Nature444,499–502(2006).CAS  PubMed  ADS  GoogleScholar  28.Dickel,D.E.,Visel,A.,&Pennacchio,L.A.Functionalanatomyofdistant-actingmammalianenhancers.Philos.Tras.R.Soc.Lond.B368,20120359(2013).CAS  GoogleScholar  29.Kawakami,Y.etal.Sallgenesregulateregion-specificmorphogenesisinthemouselimbbymodulatingHoxactivities.Development136,585–594(2009).CAS  PubMed  PubMedCentral  GoogleScholar  30.Camp,E.,Hope,R.,Kortschak,R.D.,Cox,T.C.&Lardelli,M.Expressionofthreespalt(sal)genehomologuesinzebrafishembryos.Dev.GenesEvol.213,35–43(2003).CAS  PubMed  GoogleScholar  31.Fernandez-Guerrero,M.etal.Mammalian-specificectodermalenhancerscontroltheexpressionofHoxcgenesindevelopingnailsandhairfollicles.Proc.NatlAcad.Sci.USA117,30509–30519(2020).CAS  PubMed  GoogleScholar  32.Spitz,F.,Herkenne,C.,Morris,M.A.&Duboule,D.Inversion-induceddisruptionoftheHoxdclusterleadstothepartitionofregulatorylandscapes.Nat.Genet.37,889–893(2005).CAS  PubMed  GoogleScholar  33.Andrey,G.etal.AswitchbetweentopologicaldomainsunderliesHoxDgenescollinearityinmouselimbs.Science340,1234167(2013).PubMed  GoogleScholar  34.Montavon,T.&Duboule,D.ChromatinorganizationandglobalregulationofHoxgeneclusters.Philos.Trans.R.Soc.Lond.B.368,20120367(2013). GoogleScholar  35.Woltering,J.M.,Noordermeer,D.,Leleu,M.&Duboule,D.ConservationanddivergenceofregulatorystrategiesatHoxlociandtheoriginoftetrapoddigits.PLoSBiol.12,e1001773(2014).PubMed  PubMedCentral  GoogleScholar  36.Woltering,J.M.etal.Sarcopterygianfinontogenyelucidatestheoriginofhandswithdigits.Sci.Adv.6,eabc3510(2020).CAS  PubMed  PubMedCentral  ADS  GoogleScholar  37.Woltering,J.M.,Holzem,M.&Meyer,A.Lissamphibianlimbsandtheoriginsoftetrapodhoxdomains.Dev.Biol.456,138–144(2019).CAS  PubMed  GoogleScholar  38.Nagano,T.etal.ComparisonofHi-Cresultsusingin-solutionversusin-nucleusligation.GenomeBiol.16,175(2015).PubMed  PubMedCentral  GoogleScholar  39.Wutz,G.etal.TopologicallyassociatingdomainsandchromatinloopsdependoncohesinandareregulatedbyCTCF,WAPL,andPDS5proteins.EMBOJ.36,3573–3599(2017).CAS  PubMed  PubMedCentral  GoogleScholar  40.Smith,J.J.etal.Achromosome-scaleassemblyoftheaxolotlgenome.GenomeRes.29,317–324(2019).CAS  PubMed  PubMedCentral  GoogleScholar  41.Nowoshilow,S.&Tanaka,E.M.Introducingwww.axolotl-omics.org–anintegrated-omicsdataportalfortheaxolotlresearchcommunity.Exp.CellRes.394,112143(2020).CAS  PubMed  GoogleScholar  42.Bolger,A.M.,Lohse,M.&Usadel,B.Trimmomatic:aflexibletrimmerforIlluminasequencedata.Bioinformatics30,2114–2120(2014).CAS  PubMed  PubMedCentral  GoogleScholar  43.Song,L.&Florea,L.Rcorrector:efficientandaccurateerrorcorrectionforIlluminaRNA-seqreads.Gigascience4,48(2015).PubMed  PubMedCentral  GoogleScholar  44.MacManes,M.D.TheOysterRiverProtocol:amulti-assemblerandkmerapproachfordenovotranscriptomeassembly.PeerJ6,e5428(2018).PubMed  PubMedCentral  GoogleScholar  45.Chikhi,R.&Medvedev,P.Informedandautomatedk-mersizeselectionforgenomeassembly.Bioinformatics30,31–37(2014).CAS  PubMed  GoogleScholar  46.Robertson,G.etal.DenovoassemblyandanalysisofRNA-seqdata.Nat.Methods7,909–912(2010).CAS  PubMed  GoogleScholar  47.Emms,D.M.&Kelly,S.OrthoFinder:solvingfundamentalbiasesinwholegenomecomparisonsdramaticallyimprovesorthogroupinferenceaccuracy.GenomeBiol.16,157(2015).PubMed  PubMedCentral  GoogleScholar  48.Li,W.,Jaroszewski,L.&Godzik,A.Clusteringofhighlyhomologoussequencestoreducethesizeoflargeproteindatabases.Bioinformatics17,282–283(2001).CAS  PubMed  GoogleScholar  49.Simão,F.A.,Waterhouse,R.M.,Ioannidis,P.,Kriventseva,E.V.&Zdobnov,E.M.BUSCO:assessinggenomeassemblyandannotationcompletenesswithsingle-copyorthologs.Bioinformatics31,3210–3212(2015).PubMed  GoogleScholar  50.Nishimura,O.,Hara,Y.&Kuraku,S.gVolanteforstandardizingcompletenessassessmentofgenomeandtranscriptomeassemblies.Bioinformatics33,3635–3637(2017).CAS  PubMed  PubMedCentral  GoogleScholar  51.Kim,D.,Langmead,B.&Salzberg,S.L.HISAT:afastsplicedalignerwithlowmemoryrequirements.Nat.Methods12,357–360(2015).CAS  PubMed  PubMedCentral  GoogleScholar  52.Pertea,M.,Kim,D.,Pertea,G.M.,Leek,J.T.&Salzberg,S.L.Transcript-levelexpressionanalysisofRNA-seqexperimentswithHISAT,StringTieandBallgown.Nat.Protocols11,1650–1667(2016).CAS  PubMed  GoogleScholar  53.Yan,H.,Bombarely,A.&Li,S.DeepTE:acomputationalmethodfordenovoclassificationoftransposonswithconvolutionalneuralnetwork.Bioinformatics36,4269–4275(2020).CAS  PubMed  GoogleScholar  54.Chalopin,D.&Volff,J.-N.Analysisofthespottedgargenomesuggestsabsenceofcausativelinkbetweenancestralgenomeduplicationandtransposableelementdiversificationinteleostfish.J.Exp.Zoolog.BMol.Dev.Evol.328,629–637(2017).CAS  GoogleScholar  55.Lerat,E.,Fablet,M.,Modolo,L.,Lopez-Maestre,H.&Vieira,C.TEtoolsfacilitatesbigdataexpressionanalysisoftransposableelementsandrevealsanantagonismbetweentheiractivityandthatofpiRNAgenes.NucleicAcidsRes.45,e17(2017).PubMed  GoogleScholar  56.Wu,T.D.&Watanabe,C.K.GMAP:agenomicmappingandalignmentprogramformRNAandESTsequences.Bioinformatics21,1859–1875(2005).CAS  PubMed  GoogleScholar  57.Haas,B.J.etal.ImprovingtheArabidopsisgenomeannotationusingmaximaltranscriptalignmentassemblies.NucleicAcidsRes.31,5654–5666(2003).CAS  PubMed  PubMedCentral  GoogleScholar  58.Bray,N.L.,Pimentel,H.,Melsted,P.&Pachter,L.Near-optimalprobabilisticRNA-seqquantification.Nat.Biotechnol.34,525–527(2016).CAS  PubMed  GoogleScholar  59.UniProtConsortium.UniProt:aworldwidehubofproteinknowledge.NucleicAcidsRes.47,D506–D515(2019). GoogleScholar  60.Slater,G.S.C.&Birney,E.Automatedgenerationofheuristicsforbiologicalsequencecomparison.BMCBioinformatics6,31(2005).PubMed  PubMedCentral  GoogleScholar  61.Pearson,W.R.FindingproteinandnucleotidesimilaritieswithFASTA.Curr.Protoc.Bioinformatics53,3.9.1–3.9.25(2016). GoogleScholar  62.Chan,P.P.&Lowe,T.M.tRNAscan-SE:searchingfortRNAgenesingenomicsequences.MethodsMol.Biol.1962,1–14(2019).CAS  PubMed  PubMedCentral  GoogleScholar  63.Nawrocki,E.P.&Eddy,S.R.Infernal1.1:100-foldfasterRNAhomologysearches.Bioinformatics29,2933-2935(2013).CAS  PubMed  PubMedCentral  GoogleScholar  64.Kozomara,A.,Birgaoanu,M.&Griffiths-Jones,S.miRBase:frommicroRNAsequencestofunction.NucleicAcidsRes.47,D155–D162(2019).CAS  PubMed  GoogleScholar  65.Enright,A.J.etal.MicroRNAtargetsinDrosophila.GenomeBiol.5,R1(2003).PubMed  PubMedCentral  GoogleScholar  66.Harris,R.ImprovedPairwiseAlignmentofGenomicDNA(PennsylvaniaStateUniv.,2007).67.Kent,W.J.etal.ThehumangenomebrowseratUCSC.GenomeRes.12,996–1006(2002).CAS  PubMed  PubMedCentral  GoogleScholar  68.Blanchette,M.etal.Aligningmultiplegenomicsequenceswiththethreadedblocksetaligner.GenomeRes.14,708–715(2004).CAS  PubMed  PubMedCentral  GoogleScholar  69.Siepel,A.etal.Evolutionarilyconservedelementsinvertebrate,insect,worm,andyeastgenomes.GenomeRes.15,1034–1050(2005).CAS  PubMed  PubMedCentral  GoogleScholar  70.Cooper,G.M.etal.Distributionandintensityofconstraintinmammaliangenomicsequence.GenomeRes.15,901–913(2005).CAS  PubMed  PubMedCentral  GoogleScholar  71.Cho,Y.S.etal.Thetigergenomeandcomparativeanalysiswithlionandsnowleopardgenomes.Nat.Commun.4,2433(2013).PubMed  PubMedCentral  ADS  GoogleScholar  72.Ruan,J.etal.TreeFam:2008update.NucleicAcidsRes.36,D735–D740(2008).CAS  PubMed  GoogleScholar  73.Whelan,S.,Irisarri,I.&Burki,F.PREQUAL:detectingnon-homologouscharactersinsetsofunalignedhomologoussequences.Bioinformatics34,3929–3930(2018).CAS  PubMed  GoogleScholar  74.Katoh,K.&Standley,D.M.MAFFTmultiplesequencealignmentsoftwareversion7:improvementsinperformanceandusability.Mol.Biol.Evol.30,772–780(2013).CAS  PubMed  PubMedCentral  GoogleScholar  75.Criscuolo,A.&Gribaldo,S.BMGE(BlockMappingandGatheringwithEntropy):anewsoftwareforselectionofphylogeneticinformativeregionsfrommultiplesequencealignments.BMCEvol.Biol.10,210(2010).PubMed  PubMedCentral  GoogleScholar  76.Lartillot,N.,Rodrigue,N.,Stubbs,D.&Richer,J.PhyloBayesMPI:phylogeneticreconstructionwithinfinitemixturesofprofilesinaparallelenvironment.Syst.Biol.62,611–615(2013).CAS  PubMed  GoogleScholar  77.Dutheil,J.Y.,Gaillard,S.&Stukenbrock,E.H.MafFilter:ahighlyflexibleandextensiblemultiplegenomealignmentfilesprocessor.BMCGenomics15,53(2014).PubMed  PubMedCentral  GoogleScholar  78.Revell,L.J.phytools:AnRpackageforphylogeneticcomparativebiology(andotherthings).MethodsEcol.Evol.3,217–223(2012). GoogleScholar  79.Lartillot,N.,Lepage,T.&Blanquart,S.PhyloBayes3:aBayesiansoftwarepackageforphylogeneticreconstructionandmoleculardating.Bioinformatics25,2286–2288(2009).CAS  PubMed  GoogleScholar  80.Gregory,T.R.Animalgenomesizedatabase,http://www.genomesize.com (2020).81.Yang,Z.PAML4:phylogeneticanalysisbymaximumlikelihood.Mol.Biol.Evol.24,1586–1591(2007).CAS  PubMed  GoogleScholar  82.Marjanović,D.ThemakingofcalibrationsausageexemplifiedbyrecalibratingthetranscriptomictimetreeofjawedvertebratesPreprintathttps://doi.org/10.1101/2019.12.19.882829(2019).83.DeBie,T.,Cristianini,N.,Demuth,J.P.&Hahn,M.W.CAFE:acomputationaltoolforthestudyofgenefamilyevolution.Bioinformatics22,1269–1271(2006).PubMed  GoogleScholar  84.Buchfink,B.,Xie,C.&Huson,D.H.FastandsensitiveproteinalignmentusingDIAMOND.Nat.Methods12,59–60(2015).CAS  PubMed  GoogleScholar  85.Niimura,Y.Olfactoryreceptormultigenefamilyinvertebrates:fromtheviewpointofevolutionarygenomics.Curr.Genomics13,103–114(2012).CAS  PubMed  PubMedCentral  GoogleScholar  86.O’Brien,K.P.,Remm,M.&Sonnhammer,E.L.L.Inparanoid:acomprehensivedatabaseofeukaryoticorthologs.NucleicAcidsRes.33,D476–D480(2005).PubMed  GoogleScholar  87.Suyama,M.,Torrents,D.&Bork,P.PAL2NAL:robustconversionofproteinsequencealignmentsintothecorrespondingcodonalignments.NucleicAcidsRes.34,W609–W612(2006).CAS  PubMed  PubMedCentral  GoogleScholar  88.Edgar,R.C.MUSCLE:multiplesequencealignmentwithhighaccuracyandhighthroughput.NucleicAcidsRes.32,1792–1797(2004).CAS  PubMed  PubMedCentral  GoogleScholar  89.Castresana,J.Selectionofconservedblocksfrommultiplealignmentsfortheiruseinphylogeneticanalysis.Mol.Biol.Evol.17,540–552(2000).CAS  PubMed  GoogleScholar  90.Huerta-Cepas,J.,Serra,F.&Bork,P.ETE3:reconstruction,analysis,andvisualizationofphylogenomicdata.Mol.Biol.Evol.33,1635–1638(2016).CAS  PubMed  PubMedCentral  GoogleScholar  91.Woltering,J.M.etal.Axialpatterninginsnakesandcaecilians:evidenceforanalternativeinterpretationoftheHoxcode.Dev.Biol.332,82–89(2009).CAS  PubMed  GoogleScholar  92.Monti,R.etal.Limb-EnhancerGenie:anaccessibleresourceofaccurateenhancerpredictionsinthedevelopinglimb.PLoSComput.Biol.13,e1005720(2017).PubMed  PubMedCentral  GoogleScholar  93.Osterwalder,M.etal.HAND2targetsdefineanetworkoftranscriptionalregulatorsthatcompartmentalizetheearlylimbbudmesenchyme.Dev.Cell31,345–357(2014).CAS  PubMed  PubMedCentral  GoogleScholar  94.Osterwalder,M.etal.Enhancerredundancyprovidesphenotypicrobustnessinmammaliandevelopment.Nature554,239–243(2018).CAS  PubMed  PubMedCentral  ADS  GoogleScholar  95.Bickelmann,C.etal.NoncanonicalHox,Etv4,andGli3geneactivitiesgiveinsightintouniquelimbpatterninginsalamanders.J.Exp.Zoolog.BMol.Dev.Evol.330,138–147(2018).CAS  GoogleScholar  96.Du,K.etal.Thesterletsturgeongenomesequenceandthemechanismsofsegmentalrediploidization.Nat.Eco.Evol.4,841–852(2020). GoogleScholar  DownloadreferencesAcknowledgementsWethankthelateJ.ClackandR.L.Carrollfortheircontributiontoourunderstandingofthewater–landtransitionofvertebrates.ThisworkwassupportedbytheGermanScienceFoundation(DFG)throughagranttoA.M.,T.B.andM.S.(Me1725/24-1,Bu956/23-1,Scha408/16-1)andtoJ.M.W.(Wo2165/2-1),andcorefundingfromtheIMPtoE.M.T.J.-N.V.andM.S.weresupportedbyajointgrantoftheFrenchResearchAgency(ANREvobooster)andDFG(SCHA408/13-1).I.I.wassupportedbytheSpanishMinistryofEconomyandCompetitiveness(MINECO)(JuandelaCierva-IncorporaciónfellowshipIJCI-2016-29566)andtheEuropeanResearchCouncil(grantagreementno.852725;ERC-StG‘TerreStriAL’toJ.deVries(UniversityofGöttingen)).W.Y.W.andO.S.weresupportedbytheAustrianScienceFundgrantsP3219andI4353.W.Y.W.issupportedbyCroucherScholarshipsforDoctoralStudy.A.K.wassupportedbyafellowshipfromtheJapaneseSocietyforthePromotionofScience(JSPS)postdoctoralfellowshipforOverseasResearchersProgram.WethankD.OcampoDaza(http://www.egosumdaniel.se/)forgenerouslysharinghisvertebrateillustrations,J.JossandP.Sordinoforthegiftoflungfishembryos,andL.PennacchioforVistaenhancerimages.AuthorinformationAuthornotesIkerIrisarriPresentaddress:DepartmentofAppliedBioinformatics,InstituteforMicrobiologyandGenetics,UniversityofGoettingen,Goettingen,GermanyTheseauthorscontributedequally:AxelMeyer,SiegfriedSchloissnig,PaoloFranchini,KangDu,Joost M.WolteringTheseauthorsjointlysupervisedthiswork:AxelMeyer,OlegSimakov,ThorstenBurmester,EllyM.Tanaka,ManfredSchartlAffiliationsDepartmentofBiology,UniversityofKonstanz,Konstanz,GermanyAxelMeyer, PaoloFranchini, JoostM.Woltering & PeiwenXiongResearchInstituteofMolecularPathology(IMP),Vienna,AustriaSiegfriedSchloissnig, SergejNowoshilow, AkaneKawaguchi & EllyM.TanakaDevelopmentalBiochemistry,Biocenter,UniversityofWürzburg,Würzburg,GermanyKangDu & ManfredSchartlTheXiphophorusGeneticStockCenter,TexasStateUniversity,SanMarcos,TX,USAKangDu & ManfredSchartlDepartmentofBiodiversityandEvolutionaryBiology,MuseoNacionaldeCienciasNaturales(MNCN-CSIC),Madrid,SpainIkerIrisarriDepartmentofNeuroscienceandDevelopmentalBiology,UniversityofVienna,Vienna,AustriaWaiYeeWong & OlegSimakovBiochemistryandCellBiology,Biocenter,UniversityofWürzburg,Würzburg,GermanySusanneKneitzInstitutfürZoologie,UniversitätHamburg,Hamburg,GermanyAndrejFabrizius & ThorstenBurmesterInstitutdeGénomiqueFonctionnelle,ÉcoleNormaleSuperieure,UniversitéClaudeBernard,Lyon,FranceCorentinDechaud & Jean-NicolasVolffFacultyofScience,UniversiteitLeiden,Leiden,TheNetherlandsHermanP.SpainkAuthorsAxelMeyerViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSiegfriedSchloissnigViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarPaoloFranchiniViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarKangDuViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJoostM.WolteringViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarIkerIrisarriViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarWaiYeeWongViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSergejNowoshilowViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSusanneKneitzViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarAkaneKawaguchiViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarAndrejFabriziusViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarPeiwenXiongViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarCorentinDechaudViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarHermanP.SpainkViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJean-NicolasVolffViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarOlegSimakovViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarThorstenBurmesterViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarEllyM.TanakaViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarManfredSchartlViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarContributionsA.M.,T.B.andM.S.conceivedthestudyandcoordinatedthework.A.M.andM.S.wrotethemanuscriptwithcontributionsfromallotherauthors.S.S.performedgenomeassemblyintocontigsandHi-Cscaffolding.P.F.undertooktranscriptomeanalysis,annotationandCNEanalyses.K.D.performedgenomeannotation,analysisofgenefamilydynamicsandgenomeexpansion.J.M.W.analysedandannotatedhoxclusters,andperformedembryonalRNA-seqandinsituhybridization.I.I.generatedphylogeneticanalyses,andmolecularclockandancestralcharacterstatereconstruction.W.Y.W.performedrepeatandsyntenicanalysis.S.N.undertookgenomecorrectionandinitialtranscriptalignment.S.K.performedpositiveselectionanalysis.A.K.undertookHi-Clibrarypreparationandlibrarypreparationforgenomecorrection.A.F.performedtranscriptomegeneration.P.X.annotatedncRNAs.C.D.andJ.-N.V.performedtransposonandrepeatanalysis.H.P.S.contributedresources.O.S.performedsyntenicanalyses.E.M.T.supervisedHi-Candgenomicsequencing,andanalyseddata.CorrespondingauthorsCorrespondenceto AxelMeyer,OlegSimakov,ThorstenBurmester,EllyM.TanakaorManfredSchartl.Ethicsdeclarations Competinginterests Theauthorsdeclarenocompetinginterests. AdditionalinformationPeerreviewinformationNaturethanksMarinaHaukness,RyanLorig-Roach,BenedictPaten,IgorSchneiderandtheother,anonymous,reviewer(s)fortheircontributiontothepeerreviewofthiswork.Peerreviewreportsareavailable.Publisher’snoteSpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.ExtendeddatafiguresandtablesExtendedDataFig.1Schematicoverviewofthescaffoldingprocedure.a,Scaffoldingconsistsconceptuallyoftwonestedloops.Theinnerloop,depictedontheright,takesalistofcontigs,theircontactinformationanditerativelyperformsaglobalagglomerativeclusteringuntilconvergenceoruntilnomorecontigscanbejoined.Thisloopisnestedinthemainprocedure,whichtakesasinputalistofseedcontigs,assignscontigstheseinitialclusters,scaffoldstheseandallowsforvisualinspectionandmergingorsplittingoftheclusters.b,N(x)plotoftheassembledcontigs.Onthey axisthecontiglengthisshown,forwhichthecollectionofallcontigsofthatlengthorlongercoversatleastxpercent(x axis)oftheassembly.c,N(x)plotofthescaffoldedgenome.Onthey axis,thecontiglengthisshownforwhichthecollectionofallscaffoldsofthatlengthorlongercoversatleastxpercent(x axis)oftheassembly.d,Hi-Ccontactheatmapofthescaffoldedportionofthelungfishgenomeassembly,orderedbyscaffoldlength.Blueboxesindicatethescaffoldboundaries.Thefourlargestscaffoldsrepresentbothchromosomearmsonasinglescaffold.Remainingscaffoldsaresplitintochromosomearmsorrepresentmicrochromosomes.e,Schemaillustratingthecontigmisjoindetectionprocess.Hi-Ccontactsarebinnedalongthediagonal.Pointsthatarenotcrossedbyasufficientnumberofcontactsaredeemedpotentialmisjoinsandarethusseparated(dottedline).ExtendedDataFig.2k-merfrequencyanalysisandtranscriptcoveragebygenomicsequences.a,TheIlluminadatasetwasusedtogeneratethespectraofk-merabundancesusingsevenk-mersizes.b–e,Transcriptcoveragebygenomicsequences.b,Histogramoftheproportionofalltranscriptlengthscoveredbythealignmenttocontigs.c,Histogramoftheproportionofalltranscriptlengthscoveredbythealignmenttoscaffolds.d,e,Histogramoftheproportionofthetranscriptlengthscoveredbythealignmenttocontigs(d)ortoscaffolds(e)ofthosetranscriptswithalignmentsthatwereimprovedafterscaffolding.ExtendedDataFig.3CNE-basedphylogeny,divergencetimesandratesofgenomeevolution.a,Maximumlikelihoodphylogenyfromnoncodingconservedalignmentblockstotalling99,601 informativesites(usingRAxML;GTRGAMMAmodel).Allbranchesweresupportedby100%bootstrapvalue;scalebarisinexpectednucleotidereplacementspersite.BranchlengthsofthetreesobtainedbytheCNEmethodorfromtheproteinsequencesshowahighcorrelation(R2 = 0.84,P 3 s.d.fromthe mean.Overall,theseregionsrepresent0.09%ofthegenome.b,Representativeregionshowingreadpile-upwithcoverageinexcessof3 s.d.fromthemean.TheentireregioniscontainedwithinaregionannotatedasrepetitivebyRepeatMasker(redinterval).Supplementaryinformation SupplementaryInformationThisfilecontainsSupplementaryMethods,SupplementaryResults,SupplementaryTablelegends,andSupplementaryReferences.ReportingSummaryPeerReviewFileSupplementaryTable1Basicstatisticsforthelungfishgenomelong-readsequencingandfinalassembly.SupplementaryTable2Assessmentofthecompletenessofthegenomeassemblyafterannotation.TheorthologysearchpipelineBUSCOwasusedwiththeCoreVertebrateGenes(CVG)andVertebrataconservedgenes(vertebrata_odb9)genesets.SupplementaryTable3Comparisonofnumbersandstructuralfeaturesofdifferentnon-codingRNAclassesandregionsinlungfishandothervertebrates.Thespreadsheethasthefollowingsections:(1)NumberofdifferenttypesofncRNAsintenfocalgenomes.Thelungfishgenomecontains17,095ncRNAgenes,including1,042tRNAgenes,1,771rRNAgenes,and3,974microRNAgenes.Lengthof5’UTR,3’UTR,CDS,andintronsofcanonicalmRNAs.(2)PredictedmiRNAtargetsitesineightgenomes.Comparedtootherspecies,lungfishdoesnotshowsignificantdifferenceinmiRNAtargetdensity,suggestingthepotentiallyneutralevolutionof3’UTRinlungfish.(3)Lengthof5’UTR,3’UTR,CDS,andintronineightfocalgenomes.Lungfishhaslongernon-codingregionsinthegenesthanotherspecies.SupplementaryTable4Lungfishrepetitiveelementstatisticsafterthefirstroundofmasking.Thetablereportstherepetitiveelement,numberofelements,length(bp)occupiedinthewholegenome,percentageofsequence(%),average_copy_length(bp).SupplementaryTable5TEstatisticsafterdoublemasking,withmergedwithresultsfromthefirstroundofmasking.Thetablereportstherepetitiveelement,numberofelements,length(bp)occupiedinthewholegenome,percentageofsequence(%),average_copy_length(bp).SupplementaryTable6ClassificationofconsensussequencesfromRepeatModelerbyDeepTE,PASTECandblast.Thetableshowsthefurtherclassificationresultofeachrepetitiveelementconsensussequencefromotherannotators."NA"referstonomatchingresultfromthetool.Thecolumn "merge_strategy"suggeststhebestwaytomergeannotationsfromdifferenttools.SupplementaryTable7Repertoireofthesmallnon-codingRNAprocessingmachinerygenesinvertebrates.Presenceorabsenceofgenesweretakenfromref.31anddataofAustralianlungfishandaxolotladded.Presence(green)orabsence(red)isindicated.SupplementaryTable8Comparisonofintronsizes(inbp)betweendevelopmentalandnon-developmentalgenesinthelungfishgenome.SupplementaryTable9Listofgenesunderpositiveselectioninmodel1andmodel2andfunctionalclustering.Thespreadsheethasthefollowingsections:(1)Positivelyselectedgenesinthelungfishgenome(model1)andinthecommonlineageoflungfishandtetrapods(model2).(2)FunctionalclusteringbytheDAVIDandIngenuitiysoftwareforgenesidentifiedformodel1.(3)FunctionalclusteringbytheDAVIDandIngenuitiysoftwareforgenesidentifiedformodel2.SupplementaryTable10Numbersofgenefamiliesthataresignificantlyexpandedorcontractedinlungfishandothervertebrates.ResultsarefromanalysesusingtheCAFEprogram,version4.NumbersaregivenforeachbranchofthephylogenydepictedinFigure1.SupplementaryTable11Genefamilydynamicsinlungfishandothervertebrates.Thespreadsheethasthefollowingsections:(1)Genefamiliesthatweresignificantly(p<0.01)expandedorcontractedinthelungfishbranch.(2)Genefamiliesthatweresignificantly(p<0.01)expandedorcontractedonancestralandterminalbranchesin10vertebratespecies.SupplementaryTable12Numberoffunctionalpulmonarysurfactantgenescomparedamongspecies.Thegenenumberisthesumofintactandtruncatedpredictions.SupplementaryTable13Repertoireofolfactoryandtastereceptors.Thenumberoffunctionalolfactoryreceptorsandtastereceptorgenesaregivenforlungfishandninerepresentativeaquatic,amphibianorterrestrialspecies.Numbersarethesumofintactandtruncatedpredictions.Odorantreceptorareassignedtogroupsrelatingtotheoriginoftherespectiveodorsaccordingtoref.32.SupplementaryTable14NumberandlengthoftheregionsinthegenomethatarenotannotatedasrepetitivebyRepeatMaskerbuthavingacoverageinexcessof3standarddeviants.SupplementaryTable15RankorderlistofestimatedchromosomeDNAcontentandofDNAcontentinscaffolds.Leftcolumn:ListofestimatedchromosomeDNAcontent.ChromosomalDNAcontentwascalculatedbymeasuringchromosomeareafromRocketal1996anddeterminingthefractionofthetotalwithagenomesizeof43Gb.Rightcolumn:ListofDNAcontentinscaffolds.Thislistisorderedbysizeanddoesnotimplyanyrelationshiptothechromosomeslistedontheleft.SupplementaryTable16ListofOligonucleotides.Thistablelisttheoligonucleotidesequencesusedforin-situhybridizationprobesynthesisforaxolotlhoxd9,hoxc13,hoxd9,lungfishsal1,hoxc13,sox9,andcichlidhoxc13.SupplementaryTable17Countsofrepetitiveelementsingenicregions.Thetablereportsthegenomicfeatures(i.e.intronandexon),subfeatures(i.e.UTR,intron/exonnumber),repetitiveelementclasses,numberofelements(bp),lengthoffeatures(bp),percentageoffeatureoccupiedbyrepetitiveelement(%)andnumericalorderofsubfeatures(usedtogeneratetheplot).SupplementaryTable18DistancebetweenCNEpairsinhuman,chicken,axolotlandlungfish.Thetablereportsthe223pairsofnon-exonicconservedelements(CNE)thatwereidentifiedinlungfishandthreetetrapods(human,chickenandaxolotl),andusedtocalculatetheintergenicdistanceandtheregion-specificexpansionofthelungfishgenome.TheselectedinformativeCNEpairs1)werepresentinthefourspeciesgenomes,2)werelocatedinintergenicspaceofthesamecontig/chromosomeineachspeciesand3)didnothaveageneinbetweenthem.Meanandmedianexpansionincomparisontoaxolotlandlungfish(thelineagethathaveundergonedrasticgenomeexpansion)areshown.SupplementaryTable19CNEshowingacceleratedevolutioninlungfish.TheprogramphyloPwasusedtotestthenon-codingconservedelements(CNE)forlineage-specificacceleratedevolutioninthelungfishlineage,usingascomplementarytreetheotherninelineagesinourmultispeciesalignment.Thep-valueforeachCNEwascomputedusingalikelihoodratiotestusingthe“ACC”modeimplementedinphyloPandcorrectedwiththeBenjamini–Hochbergfalsediscoveryrate(FDR)multipletestcorrectionprocedure.CNEID,sizeandlocationinthehumangenomechromosomeareshown.ThelastcolumnindicateswhetherthefocalCNEislocatedinintergenicoringenicspace(UTRorintron).Rightsandpermissions OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproductioninanymediumorformat,aslongasyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.Theimagesorotherthirdpartymaterialinthisarticleareincludedinthearticle’sCreativeCommonslicense,unlessindicatedotherwiseinacreditlinetothematerial.Ifmaterialisnotincludedinthearticle’sCreativeCommonslicenseandyourintendeduseisnotpermittedbystatutoryregulationorexceedsthepermitteduse,youwillneedtoobtainpermissiondirectlyfromthecopyrightholder.Toviewacopyofthislicense,visithttp://creativecommons.org/licenses/by/4.0/. ReprintsandPermissionsAboutthisarticleCitethisarticleMeyer,A.,Schloissnig,S.,Franchini,P.etal.Giantlungfishgenomeelucidatestheconquestoflandbyvertebrates. Nature590,284–289(2021).https://doi.org/10.1038/s41586-021-03198-8DownloadcitationReceived:13July2020Accepted:06January2021Published:18January2021IssueDate:11February2021DOI:https://doi.org/10.1038/s41586-021-03198-8SharethisarticleAnyoneyousharethefollowinglinkwithwillbeabletoreadthiscontent:GetshareablelinkSorry,ashareablelinkisnotcurrentlyavailableforthisarticle.Copytoclipboard ProvidedbytheSpringerNatureSharedItcontent-sharinginitiative Furtherreading Comparativeanalysisrevealswithin-populationgenomesizevariationinarotiferisdrivenbylargegenomicelementswithhighlyabundantsatelliteDNArepeatelements C.P.Stelzer J.Blommaert D.B.MarkWelch BMCBiology(2021) InvestigationoftheactivityoftransposableelementsandgenesinvolvedintheirsilencinginthenewtCynopsorientalis,aspecieswithagiantgenome FedericaCarducci ElisaCarotti MariaAssuntaBiscotti ScientificReports(2021) Giantgenomesoflungfish GrantOtto NatureReviewsGenetics(2021) EarliestmigratorycephalicNCcellsarepotenttodifferentiateintodentalectomesenchymeofthetwolungfishdentitions:tetrapodomorphancestralconditionofunconstrainedcapabilityofmesencephalicNCcellstoformoralteeth MartinKundrát TheScienceofNature(2021) Fishgenomicsanditsimpactonfundamentalandappliedresearchofvertebratebiology SyedFarhanAhmad MaryamJehangir CesarMartins ReviewsinFishBiologyandFisheries(2021) CommentsBysubmittingacommentyouagreetoabidebyourTermsandCommunityGuidelines.Ifyoufindsomethingabusiveorthatdoesnotcomplywithourtermsorguidelinespleaseflagitasinappropriate. DownloadPDF AssociatedContent Milestone GenomicSequencing Advertisement Explorecontent Researcharticles News Opinion ResearchAnalysis Careers Books&Culture Podcasts Videos Currentissue Browseissues Collections Subjects FollowusonFacebook FollowusonTwitter Signupforalerts RSSfeed Aboutthejournal JournalStaff AbouttheEditors JournalInformation Ourpublishingmodels EditorialValuesStatement Awards JournalImpact Contact Editorialpolicies HistoryofNature Sendanewstip Publishwithus ForAuthors ForReferees Submitmanuscript Search Searcharticlesbysubject,keywordorauthor Showresultsfrom Alljournals Thisjournal Search Advancedsearch Quicklinks Explorearticlesbysubject Findajob Guidetoauthors Editorialpolicies