Discord between morphological and phylogenetic species boundaries: incomplete lineage sorting and recombination results ...

MISSING IMAGE

Material Information

Title:
Discord between morphological and phylogenetic species boundaries: incomplete lineage sorting and recombination results in fuzzy species boundaries in an asexual fungal pathogen
Physical Description:
Mixed Material
Language:
English
Creator:
Stewart, Jane E.
Timmer, Lavern W.
Lawrence, Christopher
Pryor, Barry M.
Peever, Tobin L.
Publisher:
BioMed Central (BMC Evolutionary Biology)
Publication Date:

Notes

Abstract:
Background: Traditional morphological and biological species concepts are difficult to apply to closely related, asexual taxa because of the lack of an active sexual phase and paucity of morphological characters. Phylogenetic species concepts such as genealogical concordance phylogenetic species recognition (GCPSR) have been extensively used; however, methods that incorporate gene tree uncertainty into species recognition may more accurately and objectively delineate species. Using a worldwide sample of Alternaria alternata sensu lato, causal agent of citrus brown spot, the evolutionary histories of four nuclear loci including an endo-polygalacturonase gene, two anonymous loci, and one microsatellite flanking region were estimated using the coalescent. Species boundaries were estimated using several approaches including those that incorporate uncertainty in gene genealogies when lineage sorting and non-reciprocal monophyly of gene trees is common. Results: Coalescent analyses revealed three phylogenetic lineages strongly influenced by incomplete lineage sorting and recombination. Divergence of the citrus 2 lineage from the citrus 1 and citrus 3 lineages was supported at most loci. A consensus of species tree estimation methods supported two species of Alternaria causing citrus brown spot worldwide. Based on substitution rates at the endo-polygalacturonase locus, divergence of the citrus 2 and the 1 and 3 lineages was estimated to have occurred at least 5, 400 years before present, predating the human-mediated movement of citrus and associated pathogens out of SE Asia. Conclusions: The number of Alternaria species identified as causing brown spot of citrus worldwide using morphological criteria has been overestimated. Little support was found for most of these morphospecies using quantitative species recognition approaches. Correct species delimitation of plant-pathogenic fungi is critical for understanding the evolution of pathogenicity, introductions of pathogens to new areas, and for regulating the movement of pathogens to enforce quarantines. This research shows that multilocus phylogenetic methods that allow for recombination and incomplete lineage sorting can be useful for the quantitative delimitation of asexual species that are morphologically indistinguishable. Two phylogenetic species of Alternaria were identified as causing citrus brown spot worldwide. Further research is needed to determine how these species were introduced worldwide, how they differ phenotypically and how these species are maintained. Keywords: Coalescent, Species delimitation, Species tree, Gene tree
General Note:
Stewart et al. BMC Evolutionary Biology 2014, 14:38 http://www.biomedcentral.com/1471-2148/14/38; Pages 1-14
General Note:
doi:10.1186/1471-2148-14-38 Cite this article as: Stewart et al.: Discord between morphological and phylogenetic species boundaries: incomplete lineage sorting and recombination results in fuzzy species boundaries in an asexual fungal pathogen. BMC Evolutionary Biology 2014 14:38

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
© 2014 Stewart et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Resource Identifier:
System ID:
AA00021511:00001

Full Text

PAGE 1

RESEARCHARTICLEOpenAccessDiscordbetweenmorphologicalandphylogenetic speciesboundaries:incompletelineagesorting andrecombinationresultsinfuzzyspecies boundariesinanasexualfungalpathogenJaneEStewart1,5*,LavernWTimmer2,ChristopherBLawrence3,BarryMPryor4andTobinLPeever1AbstractBackground: Traditionalmorphologicalandbiologicalspeciesconceptsaredifficulttoapplytocloselyrelated, asexualtaxabecauseofthelackofanactivesexualphaseandpaucityofmorphologicalcharacters.Phylogenetic speciesconceptssuchasgenealogicalconcordancephylogeneticspeciesrecognition(GCPSR)havebeen extensivelyused;however,methodsthatincorporategenetreeuncertaintyintospeciesrecognitionmaymore accuratelyandobjectivelydelineatespecies.Usingaworldwidesampleof Alternariaalternata sensulato,causal agentofcitrusbrownspot,theevolutionaryhistoriesoffournuclearlociincludinganendo-polygalacturonase gene,twoanonymousloci,andonemicrosatelliteflankingregionwereestimatedusingthecoalescent.Species boundarieswereestimatedusingseveralapproachesincludingthosethatincorporateuncertaintyingene genealogieswhenlineagesortingandnon-reciprocalmonophylyofgenetreesiscommon. Results: Coalescentanalysesrevealedthreephylogeneticlineagesstronglyinfluencedbyincompletelineage sortingandrecombination.Divergenceofthecitrus2lineagefromthecitrus1andcitrus3lineageswassupported atmostloci.Aconsensusofspeciestreeestimationmethodssupportedtwospeciesof Alternaria causingcitrus brownspotworldwide.Basedonsubstitutionratesattheendo-polygalacturonaselocus,divergenceofthecitrus2 andthe1and3lineageswasestimatedtohaveoccurredatleast5,400yearsbeforepresent,predatingthe human-mediatedmovementofcitrusandassociatedpathogensoutofSEAsia. Conclusions: Thenumberof Alternaria speciesidentifiedascausingbrownspotofcitrusworldwideusing morphologicalcriteriahasbeenoverestimated.Littlesupportwasfoundformostofthesemorphospeciesusing quantitativespeciesrecognitionapproaches.Correctspeciesdelimitationofplant-pathogenicfungiiscriticalfor understandingtheevolutionofpathogenicity,introductionsofpathogenstonewareas,andforregulatingthe movementofpathogenstoenforcequarantines.Thisresearchshowsthatmultilocusphylogeneticmethodsthat allowforrecombinationandincompletelineagesortingcanbeusefulforthequantitativedelimitationofasexual speciesthataremorphologicallyindistinguishable.Twophylogeneticspeciesof Alternaria wereidentifiedascausing citrusbrownspotworldwide.Furtherresearchisneededtodeterminehowthesespecieswereintroduced worldwide,howtheydifferphenotypicallyandhowthesespeciesaremaintained. Keywords: Coalescent,Speciesdelimitation,Speciestree,Genetree *Correspondence: janestew@uga.edu1DepartmentofPlantPathology,WashingtonStateUniversity,Pullman,WA, USA5Currentaddress:DepartmentofPlantPathology,UniversityofGeorgia, Athens,Georgia Fulllistofauthorinformationisavailableattheendofthearticle 2014Stewartetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginalworkisproperlycredited.TheCreativeCommonsPublicDomain Dedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle, unlessotherwisestated.Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 http://www.biomedcentral.com/1471-2148/14/38

PAGE 2

BackgroundThedelimitationofspeciesandevolutionaryrelationships amongthemisfundamentaltobiology.However,the applicationofspeciesconceptstoputativelyasexual taxacanbedifficultandcontroversial[1,2].Notallspeciesconceptscanbeappliedtoasexualspecies,and someresearchershaveevensuggestedthatasexuallineagesdonotrepresentspeciesatall[3].Forexample,the morphologicalspeciesconcept(MSC)orbiologicalspeciesconcept(BSC)maynotbeadequatefordelineating asexualfungi[2,4,5]especiallywheremorphologicaldifferencesarenotobserved.However,anynewalleleconferringanadaptiveadvantagetoanasexualorganismin aparticularecologicalnichemaybeselected.Selective pressureonthatgeneisexpectedtoaffecttheentire genomethroughgenetichitchhikingthushavingthe potentialtorapidlyformanewcrypticspecies[4].These discreteentitiesmayberecognizedasspeciesratherthan aspartofcontinuousdistributionofphenotypes.Asexual taxaarealsoexpectedtodivergeintodiscretelineages underprocessessuchasdivergentselectionand/or geographicisolation[4]. Todate,thesystematicsofasexualfungihasrelied heavilyonphylogeneticapproa chestostudycrypticspeciationamongcloselyrelatedta xa[5-7].Inasexualfungi, phylogeneticspeciesconcepts canidentifyphylogenetically distinctlineageswiththeimplicationthatnewspecieshave formedthatarenotyetmorphologicallydistinct[8].These methodsmostofteninvolvethe concatenationofsequence alignments,usingmethodssuchasthegenealogical concordancephylogeneticspeciesrecognition(GCPSR) whichisanoperationalcriterionforspeciesrecognition [2,9].TheGCPSRfocusesonspeciesidentificationthrough multi-genegenealogiesandreciprocalmonophylytoidentifyfungalspecies[5].Thismethodhasapplicabilityfor bothasexualandsexuallineagesandspeciesboundaries areestimatedbyconcordantcladesofmulti-genegenealogies.Theabsenceofmonophylyandconflictamongthe multiplegenetreesidentifiesspecieslimitsfortaxa[5]. However,speciesboundariesofcloselyrelatedtaxa,inthe initialstagesofdivergence,canbedifficulttoascertain usingmultilocusphylogeneticmethodsbecausegenetrees ofrecentlydivergedtaxacandiffersubstantiallyintheir evolutionaryhistories[10]. Processessuchasincompletelineagesorting,recombination,andhorizontaltransfercancausediscordamong geneandspeciestrees,maskingtrueevolutionaryrelationshipsamongcloselyrelatedtaxa[11].Incongruence,in itself,cansignalpossiblerecombination,reticulation,and incompletelineagesorting.Individualgenetreesmayhave differentevolutionaryhistories[12,13]whichlimitsthe accuracyofspeciestreeestimationusingconcatenation ofloci[14].Incompletelineagesortingiscausedwhenancestralpolymorphismspersistthroughspeciationevents andeachancestralpolymorphismcanleadtodifferentallelescarriedamongdescendants[14,15].Coalescent-based methods,whichstochasticallyjoinsampledgenelineages astheyarefollowedbackintime,havebeendevelopedto incorporatelineagesortingandthepresenceofincongruentgenomicregionsintophylogeneticestimationprocedures[16-18],eveninthepresenceoflineagesortingand lackofreciprocalmonophylyatanysinglelocus[10]. Coalescentmethodshaverecentlybeenusedtoassess speciestreesforarangeoftaxaincludingsexuallyreproducingspeciessuchastheTennesseecavesalamanders ( Gyrinophilus ;[19]),tropicallowlandbirds( Manacus ;[20]), grasshoppers( Melanoplus ;[10]),andrice[21].Fewresearchers,however,haveexaminedtheutilityofthese methodsforcloselyrelatedasexualtaxathataremorphologicallyindistinguishable[22].Phylogeneticanalysesof closely-relatedtaxa,suchasrice[21], Drosophila [23],and crypticfungalspecies,suchas Penicillium [22,24],areat theintersectionofpopulationgeneticsandphylogenetics wheretheeffectsofcoalescentstochasticityresultsinhigh levelsofgenetreeincongruence[17,25-27].Estimating speciestreesforthesetaxacanbeproblematic;these methodscouldprovetobeusefulforclosely-related asexualfungaltaxa. Theputativelyasexua lcitruspathogen, Alternaria alternata ,providesanidealcasestudyfortheapplication ofquantitativespeciesrecognitionusingspeciestree estimationmethodsthatincorporateuncertaintyingene trees.Andrewetal.[28]developedaspeciesphylogeny forsmall-spored Alternaria usingfourgenomicregions includingaproteincodinggeneandthreeanonymous, non-codingregions.Significantincongruencewasfound amonggenegenealogiesandseveralputativerecombinationeventswereidentifiedwithintwoofthenon-coding regionsindicatingdivergentevolutionaryhistoriesamong theloci[28].Hypothesestoexplainthisincongruenceincludedrecombinationandincompletelineagesorting.A recentstudyofthematingsystemof A.alternata causing brownspotinFloridafoundsignaturesofrecombination [29]butstudiesofthelargerworldwidepopulationofthe pathogenarelacking. Currently,thereisalargediscordbetweenthenumber ofmorphologicalandphylogeneticspeciesthatare thoughttocausecitrusbrownspot.Thefungusinfects tangerinesandmandarins( Citrusreticulata Blanco)and tangerinexgrapefruit( C.reticulata x C.paradisiMacfad.) hybridsworldwide.WhenfirstreportedinAustralia[30], thepathogenwasidentifiedas A.citri Ellis&N.Pierce duetoitsmorphologicalsimilaritytothecausalagent ofapostharvestdisease,citrusblackrot.Sincethis time,thepathogenhasbeenreferredtoas A.alternata ‘ tangerinepathotype ’ basedonmorphologicalsimilarityto A.alternata [31-33].Further,molecularcomparisonswere performedusingrestrictionfragmentlengthpolymorphismsStewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page2of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 3

(RFLPs)ofthenuclearribosomalDNA(internaltranscribed spacerregion,ITS)among11closelyrelated,small-spored Alternaria taxa.Resultsfromthisstudyshowedthat morphologicallysimilar Alternaria speciescollected fromdifferenthostsandthatproducehostspecific toxinssharedcommonRFLPfingerprintsandidentical ITSsequences,leadingtheauthorstoconcludethat Alternaria fungiknowntoproducehost-specifictoxins areintraspecificvariantsof A.alternata specializedin hostpathogenicity[33]. Tennew Alternaria specieshavebeendescribedfrom citrushosts[34]andphylogeneticstudieshaveattempted tomapthesemorphospeciesontophylogeniesestimated frommoleculardata.Usingaworldwidesampleofisolates, Peeveretal.[35,36]andAndrewetal.[28]evaluatedthese morphospeciesusingphylogeneticcriteriaandfoundthree distinctlineages(labeledClades1,2and3in[35]).Two oftheselineageswerefoundinFlorida(Clades1and2), whereasthethirdlineageoccurredonlyinTurkey,Israel, AustraliaandSouthAfrica(Clade3).Thesethreeworldwidelineagescorrespondedtoseveralmorphospecies including A.citriarbusti (Clade1), A.tangelonis and A.colombiana (Clade2),and A.dumosa A.turkisafria A.perangusta and A.interrupta (Clade3)[34].Peeveretal. [36]furthertestedtheconcor dancebetweenthe10citrusassociatedmorphospecies[34]oncitrususingabroader rangeofisolates.Peeveretal.[36]foundeightdistinct Alternaria cladesfromcitrushoststhatcouldbeinterpretedasphylogeneticspec iesundertheGCPSRconcept. Thisincongruencebetweenthenumberofspeciesdefined usingmorphologicalandphyloge neticcriteriaraisessignificantquestionsaboutthenumberof Alternaria taxathat causebrownspotdisease. Coalescentanalysesofgenegeneaolgies,whichdescribes descendent/ancestorrelationshipswherethegeneofinterestundergoescoalescencetoacommonancestor,canbe usedtoexaminetheevolutionaryhistoryofagenebackwardsintimeandcanbeusedtoincorporateincomplete lineagesortingintophylogeneticanalyses[37].Currently, thereisalackofagreementamongresearchersabouthow many Alternaria speciescausecitrusbrownspot,ranging fromasmanyas10speciestoaslittleas1species.The mainobjectiveofthisstudywastoquantitativelyestimate thenumberofspeciesof Alternaria causingcitrusbrown spotonaworldwidescaleutilizingnewlydeveloped methodsthatincorporatethecoalescentandaccount forrecombinationorincompletelineagesorting.The evolutionaryhistoriesandrecombinationofknowngeneticallydistinctlineages,citr us1,citrus2andcitrus3were evaluatedfromalargerworldwidesampleof A.alternata tangerinepathotypeandmorelociinordertoestablishspeciesboundariesofthesepreviouslyobservedphylogenetic lineagesusinggenesequenceconcatenationmethods andseveralapproachesthat incorporateuncertaintyin genegenealogieswhenlineagesortingandnon-reciprocal monophylyofgenetreesiscommon.MethodsIsolatesOnehundredandfortytwoisolateswerecollectedfrom brownspotlesionsoncultivarsoftangerines,mandarins ( Citrusreticulata Blanco)andtangerinexgrapefruit hybrids[ C.reticulata x C.paradise (Macfad.)]in12countries,includingArgentina(AR),Australia(AU),Brazil(BZ), Colombia(CB),Greece(GR),Iran(IR),Israel(IS),Italy(IT), Peru(PE),Spain(SP),Turkey(TU),andUSA(FL).Sixty-five oftheseisolatesfromAustralia,Colombia,Israel,South Africa,andUnitedStatesoverlappedwithaprevious study([35];Figure1A,Additionalfile1:TableS1).Isolates fromArgentina,Australia,Brazil,Colombia,Greece,Iran, Israel,Italy,Peru,Spain,Turkey,andUSAweresampled fromtangerinehybridsinseveralgeographicallyseparated citrusgroveswithineachcountry.Australianisolateswere collectedfromtangerinehybridsintheNararaarboretum (Narara,NewSouthWales),andisolatesfromFlorida, USAwerecollectedfromasmallarea(2,500m2)ina singlegroveofMinneol atangelo[35,36,38].DNAextractionFungiwerecultivatedinpotatodextrosebroth(FisherSci, Pittsburg,PA)for5 – 7daysatroomtemperatureonan orbitalshakerat150rpm.GenomicDNAwasextracted frompowdered,lyophiliz edmyceliumfollowingthe methodsofPeeveretal.[38],usingeitheraQiagen DNeasyKitoraPhenol-Chloroformprocedure.Extracted DNAwasquantifiedusingaNanodrop1000(NanoDrop products,Wilmington,DE,USA),andatotalof30ngwas usedastemplateforPCR.Isolatesweremaintainedin long-termstorageonsterilizedfilterpaperat 20Cas previouslydescribed[38].Endo-polygalacturonasesequencingEachisolatewassequencedatthe endo-polygalacturonase locus[39]usingprimersandconditionssimilartothose usedbyPeeveretal.[35,36,40]andAndrewetal.[28]. Thisregionhasbeenextensivelyusedbecauseother commonlyusedgenomicregions,suchasribosomal regions,mitochondriallargeandsmallsubunits,andthe beta-tubulingene,showlittlevariationamong Alternaria isolatescollectedfromcitrus[41].Amplifiedproducts werevisualizedin2%ethidiumbromide-stainedagarose gels.AmplifiedDNAfragmentsweresequenceddirectly onbothstrandsfollowingtreatmentwithEXOSAP-IT (USB,Cleveland,OH)usingtheBigDyeterminatorkit (AppliedBiosystems,FosterCity,CA).Sequencereads wereperformedoneitheraPEBiosystemsmodel3700 automatedDNASequencerbytheLaboratoryforBiotechnologyandBioinformaticsatWashingtonStateStewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page3of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 4

Figure1 A.Geographicoriginof Alternariaalternata sampledfromcitrusin12countries(smallblackcircles). Thenumbersofisolates sampledfromeachcountryareinparentheses.Thepiechartforeachcountryrepresentsthenumberof endoPG haplotypeswitheachcolor denotingadifferent endoPG haplotype[(Hap1(Fuschia),Hap2(orange),Hap3(blue),Hap4(darkblue),Hap5(red),Hap6(maroon),Hap7(yellow), Hap8(green),Hap9(pink),Hap10(lightpink),Hap11(brown),Hap12(purple),andHap13(lightyellow)]. B .An endo-polygalacturonase ( endoPG ) phylogenyestimatedamongaworldwidesampleofcitrusbrownspotisolatesusingBayesianinferencewith A.tomato asanoutgroup.Three phylogeneticlineagesidentifiedcorrespondtoClades1,2,and3ofPeeveretal.[35].Nodesupportisgivenasposteriorprobabilitiesand bootstrapvaluesbasedonBayesianandlikelihoodanalyses. Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page4of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 5

University,Pullman,WAoratElimBiopharmaceuticals, Inc,Wayward,CA.OPA1-3,OPA2-1, and Flank-F3sequencingAsubsetofisolateswithunique endoPG haplotypes representingeachsampledlocation( n =34)wasselected foradditionalsequencingandanalyses.Additionalloci includedtwoanonymous,non-codingSCARmarkers OPA1-3andOPA2-1[28],andonenon-codingmicrosatelliteflankingregionFlank-F3.LociOPA1-3and OPA2-1werepreviouslyused forphylogeneticstudies ofsmall-spored A.alternata [36,42].Amplificationand sequencingconditionswereaspreviouslyreported [28,36,40,42].Clonesequen cescontainingmicrosatellites[43]weredownloadedfromGenBank[accession: DQ272483toDQ272487],andprimersweredesigned toamplifythemicrosatelli teflankingregionsusing Primer3[44].PrimersequencesforAA-Flank-F3were (Flank3F-5 -AGCCAAAACACGTTGATACC-3 /Flank3R5 ATCCGCAGCGAAAAGAACT3).Twentymicroliter PCRreactionmixturescontained20nggenomicDNA, 1PCRbuffer(NewEnglandBiolabs(NEB),Ipswich,MA), 4nmolofeachdNTP(NEB),50pmolprimer,and1U ofTaqpolymerase(NEB).Cyc lingconditionsconsisted ofdenaturationat94Cfor4min;44cyclesof94Cfor1 min,55Cfor30sec,and72Cfor2min;finalextension wasat72Cfor7mindependingontheoptimalconditions foreachprimerset.PhylogeneticanalysesandcongruenceamonglociEachlocuswasanalyzedindependently.Maximum likelihoodandBayesianphylogeneticanalyseswereperformedforeachlocususingPhyML[45]andMrBayes3.0 [46].DT-ModSel[47]wasusedtoestimatethenucleotidesubstitutionmodelsbestrepresentingeachdataset. AnKimuraK80modelwasselectedforthe endoPG (basefrequencies=equal;transversion=5.29;proportionof variablesites=0),andOPA2-1(basefrequencies=equal; transversionratio=5.52;proportionofvariablesites=0). TheKimuraK80modelwithproportionofinvariable siteswasselectedforOPA1-3(basefrequencies=equal; tratio=1.47;proportionofvariablesites=0).TheJukes Cantor69modelwasselectedforFlank-F3(basefrequencies=equal;proportionofvariablesites=0).Maximumlikelihoodanalyseswereperformedunderthe heuristicsearchwithTBRbranch-swapping,andbootstrap supportwasestimatedusing1000pseudoreplicates. ForBayesiananalyses,Metropolis-coupledMarkovchain MonteCarlosearchesincluded2runswithfourchains eachrunfor3,000,000generationsandensuringthat theaveragesplitfrequenciesbetweentherunswasless than1%.Treesweresampledevery200generations.Each rungenerated60,001treesofwhichthefirst18,000trees (30%ofthetotalnumberofgenerations)werediscarded as  burnin Ž ,asvisuallydeterminedbyevaluatinglogfiles inTRACERversion1.5[48]. Inordertotesttopologycongruenceamongphylogenies fromdifferentloci,theShimod aira-Hasegawa(SH)testof topologicalcongruence[49]wa sconductedontheBayesian phylogeniesasimplementedinPAUP*10_4b[50]with1000 RELLresamplingreplicates.CoalescentanalysesAncestralhistoriesofthecitrusbrownspotlineages wereestimatedusingthecoalescent[37,45].Sequence datawerealignedandeditedmanuallybyeyeandusing clustalWimplementedinBioEditv7.0.53forWindows [51].IsolateswereassignedtohaplotypesusingDnaSPv 5.1[52,53].Toverifythesuitabilityofeachlocusforcoalescentanalyses,theneutralityofeachlocuswasestimated usingFuandLi ’ sDandTajima ’ s D andpotentialrecombinationwithineachlocuswasexaminedusingRmin[54]as implementedinDnaSP.Incompatibilitymatrices[55] wereestimatedinSNAPCladeandSNAPMatrixasimplementedinSNAPworkbench[56]tovisualizeincompatiblenucleotidesites,suchasthosearisingfrom recombinationorrecurrentmutation.Sequenceswere collapsedintouniquehaplotypesusingSNAPmap[57] andSITESversion1.1[58]byremovingindelsandincompatiblesites. Evolutionaryhistoriesweresimulatedusingcoalescent analysesforeachlocus.TwolociFlank-3andOPA1-3 showedevidenceforrecombinationandthereforecoalescentanalysesforallgenomicregionswereimplemented usingrecomversion5.8(withinSNAPworkbench),which allowsforcoalescentanalyseswithrecombination,assumingtheinfinite-sitesmodel,neutralevolution,panmixiaand constantpopulationsize[59].Usingahaploidcoalescent model,recom5.8estimatespopulationrecombinationrate, ,(2Ner)andmutationrate, (2Ne ).Theseestimateswere thenusedtoobtainestimates ofthenumberofrecombinationeventsandthetimetothemostrecentcommonancestor(TMRCA).Forallanalyses,theancestralstateof eachsegregatingsitewasestimatedbycomparingeachsite tothatof A.tomato isolateBMP2031[60].Siteswhichhad ambiguousancestralstateswereremovedfromthedatasets. OPA1-3hadfivesites(205,230,301,382,and436)with threestatesandonesite(349)withfourstates,andthe endoPG hadonesite(346)withthreestates.Coalescent analyseswereperformedinSNAPworkbench[56],with fiveindependentruns(1millionsimulationseach)per genomicregiontoensureconvergenceforeachparameter estimate.Conditionswereswitchedto+bforgenomic regionOPA1-3,whichabortslowprobabilitypathsand returnstozero.Genegeneaologiesandminimalrecombinationgraphs(ARGs)wereconstructedtographically representtheevolutionaryhistoryofthecitruslineages estimatedbythecoalescent.NoputativerecombinationStewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page5of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 6

eventsweredetectedwithin endoPG andOPA2-1.Genetree [61],whichassumescoalescentanalyseswithoutrecombination,wasusedtoestimatethecoalescentgenegenealogies fortheseregionsandcomparedtotherecom5.8results. Fiveindependentsimulationswith1millionrunseachwere conductedonhaplotypesofeachregiontoestimate theagesofmutationsandtheTMRCA,asinferredby an A.tomato rootedtree,toassessconvergence.Agraphof thetreewasgeneratedwithcoalescentunittimesusing Treepic[62].Anancestralre combinationgraph(ARG)can beusedtovisualizearecombiningcoalescenthistorythat cannotbedisplayedusingabifurcatingtree[63].The ancestralhistoryoftherecombiningOPA1-3andFlank-F3 haplotypeswerereconstructe dusingaparsimonyapproach thataccountsforbothmutationandrecombination backwardsintime.Beagle[63]wasusedtoproducean ARGforFlank-F3,whereaskwargwasusedforOPA1-3 [63].Beaglecomputesminimumrecombinationhistories withanexhaustiveapproach.Kwarg,ontheotherhand, implementsaheuristicsearchforplausiblehistoriesand doesnotguaranteetheminimalrecombinationhistory. Beaglewasrunforbothgenomicregions,howeverdueto theincreasedcomplexityandputativenumberofrecombinationeventswithinOPA1-3,eachtrialruncrashed.SpeciestreeestimationFourmethodswereusedtoestimatespeciestrees amongthefourgenetrees.Theseincludedconcatenation,genealogicalconcor dancephylogeneticspecies recognition(GCPSR;[5]),minimizingdeepcoalescence (MDC;[14,17,64]),andamixtureofcoalescentandtheYule process(*BEAST)[48].ForMDCand*BEAST,taxaarerequiredtobeassignedtospecies apriori .Therefore,taxa wereassignedtospeciesbasedonidentified endoPG clades (citrus1,citrus2,andcitrus3,Additionalfile1:TableS1). Aphylogenyoftheconcatenateddatasetwasimplemented inMrBayes.LociwerepartitionedandthepreviouslyestimatedevolutionarymodelswereusedforBayesiananalyses. Metropolis-coupledMarkovchainMonteCarlosearchesincluded2runswithfourchainseachrunfor3,000,000generationsandensuringthattheaveragesplitfrequencies betweentherunswaslessthan1%.Eachrungenerated 60,001treesofwhichthefirst18,000trees(30%ofthetotal numberofgenerations)werediscardedas  burnin Ž GCPSRidentifiesspeciesboundariesbycomparing multiplegenetreesamongthesamesetoftaxa[5,65]. Putativespeciesareidentifiedwhenrepresentativesof aspeciesformedwell-supportedclades(95posterior probability/70bootstrap)inallgenetrees[66].Bayesian andlikelihoodtreesearcheswereusedtoestimatespecieswithinthecitrusbrownspotworldwidepopulation accordingtoGCPSRcriteria. TheMDCapproachassumesthatdiscordanceofgene treesistheresultofincompletelineagesorting[14,17]. Thedeepcoalescencemeasureisacountofthenumber ofextragenelineagesthatresultfromfittingagene treeintoaspeciestree,therebysummingtheextra genelineagesasameasureofdiscordance.MDCthen searchesforaspeciestreebyminimizingthenumber ofdeepcoalescencesacrossloci.MDCanalysiswas implementedinMesquitev2.73[67].Genetreeuncertaintywasaccommodatedintospeciestreeinference byresampling(500times)theposteriorprobabilityof treetopologiesobtainedfromtheBayesianphylogeneticanalysesusingtheMesquitesoftwaremodule AUGIST[64].Tofitthegenetreesintoaspeciestree thatminimizedthenumberofdeepcoalescenceacross loci,genetreeswereconsideredrooted( A.tomato as theroot)andaheuristicsearchutilizingsubtreepruning andre-graftingwasused.Allequallyparsimoniousspecies treeswereretained,anda50%majority-ruleconsensus treewasgenerated.Biparitionfrequenciesfornodeswere usedasmeasuresofspeciestreeuncertainty. *BEASTgeneratesposteriorsamplesbysimultaneously estimatinggeneandspeciestreesunderahierarchical coalescentmodelwhileallowingforindependentevolutionaryprocessesineachgenomicregion.BEAUTi version1.7.5[48]wasusedtocreateXML-formatted inputfilesfor*BEASTv1.7.5.Substitutionmodelswere chosenaspreviouslydescribedandwereunlinked acrossgeneswithparametersestimatedseparatelyfor eachgene.Asaneededpriorin*BEAST,isolateswere assignedtospeciesgroupsundertheTraitstabbased onthethreecladesidentifiedinthe endoPG phylogeny. EvolutionaryrateswereestimatedunderaYuleprocess [68].AYulemodelwaschosenasthespeciestreeprior, whichassumesaconstantlineagebirthrateforeach branchinthetree.Thistreepriorismostsuitablefor treesdescribingtherelatio nshipsbetweenindividuals fromdifferentspeciesandisoftenthoughtofasdescribingthenetrateofspeciation[69].Speciestreeestimationswerecarriedoutbasedonstrictmolecular clockassumption,followingthemethodsofHeledand Drummond[48].Datasetswererunfor50million generationsinBEAST,samplingevery5,000generations.Analyseswereperformedtwice.Postburnin treeswerecombinedwithth eprogramLogCombiner (BEASTv1.6.0),andchainswereassumedtoconvergewhentheaveragestandarddeviationofsplit frequencieswasfoundtobe<0.011.Themaximum cladecredibilitytreewith posteriorprobabilityofeach nodewascomputedwiththeprogramTreeAnnotator (BEASTv1.6.0).LogfileswereevaluatedinTRACER version1.5[48].Thespeciestreewascalculatedusing TreeAnnotatorversion1.7.2withaburn-inof5000 trees.FigTreeversion1.3.1[70]wasusedtovisualize theconsensustreenodeages,branchlengthsandposteriorprobabilities.Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page6of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 7

ResultsWorldwidephylogeneticlineagesAlternariaalternata sampledfromtangerineandtangerine hybridsonaworldwidescale( n =142)revealedatotal of13 endoPG haplotypes(Additionalfile1:TableS1, Figure1A).Bayesianandmaximumlikelihoodanalyses of endoPG resultedintwophylogeneticlineagesusing a95posteriorprobabilityand70bootstrapvaluecriterion (Figure1B)[66].Thesetwolineagesaresubsequently referredtoas ‘ citrus2 ’ and ‘ citrus3 ’ .Allotherhaplotypes (Hap8,Hap11,andHap12)wereplacedintoanothergroup calledcitrus1basedonpreviouslypublishedresultsthat showedtheexistenceof threelineages[35].CoalesenceanalysesThirty-fiverepresentativeisolateswereselectedbasedon unique endoPG haplotypeandgeographicallocationto resolvetheancestralhistoriesofthelineagesusing coalescent-basedapproaches.Neutralityandintra-locus recombinationrateswereestimatedforallthegenomic regionsusingD(Tajima ’ sandFuandLi ’ s)statisticsand Rmin.Wewereunabletorejectthenullhypothesisof neutralityfortwoloci(Table1). Coalescent-derivedgenegenealogieswereestimated forthetwonon-recombiningloci, endoPG andOPA2-1 (Figures2A-B).Inthe endoPG genealogy,isolatesof thecitrus1lineagewerepositionedancestraltoall otherhaplotypesbyseveralmutations(Figure2A).In locusOPA2 – 1signalsofincompletelineagesorting, whereancestralpolymorphi smshavenotbeensorted, wereevident(Figure2B).Forexample,someisolatesfrom citrus1andcitrus3wereinterspersedamongidentified clades,suggestiveofnon-rec iprocalmonophyly(Figure2B). Further,littlevariationwasrevealedintheOPA2-1genealogy,butelevenmutations ,respectively,separated the A.alternata lineagesfromthe A.tomato outgroup. Isolatesfromalllineagessharedhaplotypes,suchthat isolatesfromlineagescitrus1and3formedaparaphyletic group.ThiswasalsoobservedintheOPA2-1,wherehaplotypesDandGincludedisolatesfromlineagescitrus1and 3.Twoclusterswereformed,oneincludingisolatesfrom citrus1and3,whiletheotherincludedapolytomyofhaplotypesfromthecitrus1andcitrus2lineages(Figure2B). LociOPA1-3andFlank-F3hadevolutionaryhistories ofrecombination.ThreerecombinationeventswereobservedinFlank-F3,whichinallcasesinvolvedisolatesfrom lineagecitrus1and2(Figure3A).OPA1 – 3hadthemost complexevolutionaryhistoryinvolvingeightrecombination events(Figure3B).Theseeventsincludedisolatesfromall threelineagessuggestingpossiblyasexualpastassuggested byBerbeeetal.[71].PhylogeneticanalysesAnalyseswereconductedforeachlocusindependently for35isolateswithrepresentative endoPG haplotypes selectedfromtheoriginalworldwidesample.Analysis of endoPG yieldedaphylogenywithsimilartopologyto thatestimatedamongthetotalworldwidesample,supportingthreeclades(Figur e4A).Eighthaplotypeswere identifiedinlocusOPA2-1,howevernomonophyly wasobservedinthelineages(Figure4B).Atotalof7 haplotypeswereidentified locusFlank-3(Figure4C). Onlyonewell-supportedcladewasfound,includingall isolatesfromcitrus2.LocusOPA1-3,with13haplotypes, hadthemostphylogeneticresolution,resultinginsix well-supportedclades(Figure4D).Among503sites,75 sites(14.7%)werepolymorphic.Ofthese,55(10.9%) wereparsimonyinformative(Table1).Oftheseclades, onecorrespondedtoaterminalcladeincludingallcitrus2 haplotypes,clearlyseparatedfrombutsistertoonecitrus 1haplotype,andpolyphyleticrelationshipswereobserved inlineages1and3(Figures4A-D). Table1SummarystatisticsofDNApolymorphismsinaworldwidesampleofbrownspotpathogenusingfourlociLocus(No.ofisolates) endoPG (142)WendoPG (35)OPA2-1(35)F3(35)OPA1-3(35)Combined(35) No.ofsites(A)427(422)427(421)538(535)222(220)503(493)1672(1686) No.ofpoly.sites(PI/PUIB)21(10/11)24(8/16)20(14/6)15(8/7)75(55/20)135(77/58) No.ofhaplotypesC1313771224 Haplotypediversity0.720.860.800.480.870.95 Nucleotidediversity0.0070.0090.0040.0090.0410.017 Tajima ’ sD 0.78NS 1.10NS 1.70NS 1.35NS0.06NS 0.62NSFuandLi ’ sD 4.05S 2.99S 2.96S 1.60NS 0.16NS 1.47NSReceventsRN/A0068N/AAExcludingindels.BNumberofpolymorphicsites(informative/uninformative).CExcludesoutgroupisolate,BMP2031( A.tomato ).WIncludesallisolatesintheworldwidedataset(ca.142).NSNon-significant,SSignificant; P >0.10.RNumberofrecombinationeventsestimatedbyRecomv.5.8[ 59 ].Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page7of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 8

DivergencetimeestimationfortheendoPGAtotalofsixsynonymousnucleotidechangeswereobserved whencomparinghaplotypesofthecitruslineages1and3 againsthaplotypesfromcitrus2.Thoughonlyaportionof theCDSregionwasanalyzed(427outof1137bp),using thepublishedratesforneutralgenesubstitutionrateof 0.910 9and16.710 9[72]resultedinadivergence dateofatleast5,400thousandyearsbeforepresent.Estimationofspeciestreesfromphylogenetically incongruentgenegenealogiesInthephylogeniesfromtheconcatenatedandindividual datasets,lineagescitrus1and3werepolyphyletic.Supportforthecitrus2lineagewasfoundinmostgenetrees whenGCPSRcriteriawasapplied(Figure4A-D).Thecitrus1and3lineageswereon lywell-supportedinthe endoPG phylogenyandthusconstitutedasingle,polyphyleticspecieswhentheotherlociwereconsidered. Congruenceoftreetopologywastestedandstatistically significantincongruencewasdetectedforallpairwise comparisonsofalllociwithOPA1-3( P =0.00)using Shimodaira-Hasegawatests(Table2).Allotherpairwise comparisonswerenon-significant. Thephylogenygeneratedfromtheconcatenated datasetyieldedfourwell-supportedcladeswithtopology similartotheendoPGphylogeny(Figure5).Thecitrus1 and3lineageswereagainpolyphyletic,withindividuals fromeachfallingintothreeclades.However,lineagecitrus2separatedintoonemonophyleticwell-supported (1.00posteriorprobability)lineage. Littlesupportformultiplespecieswasobtainedusing theMDCapproach.IntheMDC-estimatedspeciestree, alllineagesclusteredintoonecladewithhighbi-partition frequency(over98%)(Figure6A).Incontrast,*BEAST analysesrevealedtwowellsupportedspecies,citrus2 andasecondspeciescomprisinglineagescitrus1and 3(Figure6B).DiscussionCoalescent-basedapproachesandancestralrecombination graphswereusedtoelucidatetheevolutionaryhistoryof thecitrusbrownspotpathogenonaworldwidescaleand quantitativelyestimatethenumberofspeciescausingthis disease.Theevolutionaryhistoryoftheselineagesshowed patternsofincompletelineagesortingandrecombination ashasbeenobservedamongotherclosely-relatedtaxa [10,14]andwashintedatinpreviousphylogeneticstudiesof Alternaria [42].Lineagesorting,recombination, andhorizontaltransfer[17]makephylogeneticanalyses andspeciesdelimitationamongsmall-spored Alternaria challenging[28,36,40].Incontrasttopreviousstudies thatdescribedtenmorphospeciescausingAlternariabrown spotofcitrus[34]weidenti fiedoneortwospeciesamong threephylogeneticlineagesusingthe  minimizedeep coalescence Ž (MDC)andhierarchicalBayesianmodel approaches,respectively.Both methodsfailedtodifferentiatethecitrus1and3lineagesbutthehierarchicalBayesian modelmethoddifferentiatedthecitrus2lineagefromcitrus 1and3.Speciesboundariesbetweenthemorphospecies A.citriarbusti (citrus1)and A.dumosa A.turkisafria A.perangusta and A.interrupta (citrus3)werepoorlysupportedbybothmethods.However,theGCPSRand*BEAST analysesseparatedthemorphospecies A.tangelonis and A.colombiana (citrus2)fromcitrus1and3.Thehypothesisthatallhost-specifictoxinproducing,small-spored Alternaria taxainfectingcitrusaswellasotherhostssuch asstrawberry,Japanesepear,apple,andtomatorepresent sub-specificvariantsor  pathotypes Ž of A.alternata isnot Figure2 Coalescent-basedgenegenealogiesestimatedassumingnorecombinationforloci(A)EndoPGand(B)OPA2-1,withno conflictingsites. EachgenealogyisscaledtoTMRCAof1andnumbersrepresentmutations.Eachhaplotypeislabeledwithaletterandthe frequenciesofeachhaplotypeandrelationshiptoeachlineageisshown. Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page8of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 9

new[73].Oneofouranalysessupportedthishypothesis andallotheranalysesconclusivelydemonstratedthatthe numberoftaxacausingbrownspotdiseaseofcitrushas beenover-estimatedusingmorphologicalcriteria.Further studyisrequiredtodetermineifthecitrus2lineagerepresentsadistinctspecies,andifso,whatthesetwospecies shouldbenamed.EvolutionofthecitruspathogensAlternariaalternata isaninterestingmodelwithwhich tostudyspeciationinputativelyasexualtaxa.Asexual taxadonoteasilyfitwithinspeciesconceptsdeveloped forsexualtaxa[2,4,8,74,75]).Inadditiontobeingawellrecognizedsaprotroph, A.alternata infectsawiderange ofhosts,includingcitrus,pear,strawberry,andapple [73,76,77].Pathogenicformsof A.alternata arethought tohaveradiatedfromarecentcommonsaprophyticancestorthroughthehorizontalacquisitionofpathogenicityfactors[73,78].Isolatessampledforthisstudyare consideredrepresentativeofthe  tangerinepathotype Ž [38,79,80]andthesefungiproducehost-specificACTtoxinsthatarerequiredforpathogenicity[74,81].ACTtoxinsarestructurallysimilartothehost-specifictoxins producedbystrawberrypathotype(AFT-toxins)andJapanesepearpathotypeisolates(AKT-toxins)[81]andthe genescontrollingthebiosynthesisofthesetoxinsare homologous[81,82]Connectingthethreecitruslineages tothelineagesofstrawberryandJapanesepearpathotypeisolatesinfuturestudieswillallowpolarizationof thephylogenyofhost-specifictoxin-producing Alternaria andfacilitatestudiesoftoxingeneevolutionand theevolutionofpathogenicity.Assumingthatthemutationrateforthe endoPG geneissimilartootherprotein codinggenesatapproximately0.910 9to16.710 9mutationspersiteperyear[71,72,83]theestimatedtime ofdivergenceofthecitrus2lineageandthecitrus1and 3lineagesisatleast5,400yearsbeforepresent,and probablyoccurredmuchearlier.Thisdatesuggeststhat divergenceofcitrus2occurredlongbeforethemovementofcitrus,andpresumablyitspathogens,fromits putativecenteroforigininsoutheasternAsialessthan 600yearsago[84].Todate,onlycitrus2isolateshave beenisolatedfromNorthandSouthAmerica,whereas theotherlineagesarefoundonseveralOldandNew Worldcontinents.Thissuggeststhatindependentintroductionsofeachlineagemayhaveoccurredindifferent locations,presumablywiththehost. UsingGCPSRcriteria,manycrypticspecieshave beenidentified,includingseveralplantpathogenic fungi[85-87],human-infectingfungi[88-90],andan insectpathogen[11].Latinnameshavebeenassigned tosomeofthesecrypticspecies[91-93].Asfaraswe areaware,thisisthefirstreportthatcomparesthe GCPSRmethodtoothernewlydevelopedtreespecies estimationmethods.Ourresultssuggestthatinthe presenceofdivergentevolutionaryhistories,GCPSR willlikelyoverestimatethenumberofspecies.Although noreportscouldbefoundcomparingtheGCPSRmethod withothers,severalstudieshavecomparedspeciestree estimatesgeneratedwith*BEASTand/orMDCto concatenatedresults.Itisnowwell-documentedthat concatenateddatacanproduc ewell-supportedphylogeniesthatareinconsistentwiththetruespeciestree [13,94-96].Furthermore,Belfioreetal.[97]developed speciestreesusingconcatenationandBEST(Bayesian Figure3 Minimalancestralrecombinationgraphsforloci(A) Flank-F3and(B)OPA1-3,rootedwith A.tomato Recombination eventsareindicatedbyovals.Recombinantsitesconsistsofaprefix ‘ P ’ sequencethatisconcatenatedwithasuffix ‘ S ’ sequence.These designationsappearonedgesgoingintoarecombinationnode. Numberstotheright/leftofthepathsarethenumberofmutations inthatsegment.Thedirectionofthepathsisfrompasttopresent. Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page9of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 10

EstimationofSpeciesTree)forpocketgophersofthegenus Thomomys andfoundthattheconcatenatedtreeestimated fromsevenlociwasover-resolvedwhereasfewerspecies weresupportedinthephylogenyestimatedusingBEST [97].Thisresultissimilartowhatweobservedinthisstudy with A.alternata .Further,EckertandCarstens[11]tested theaccuracyofconcatenationandMDCgeneratedspecies treesusingsimulateddatasetsinthepresenceofgene flow.Asthelevelofgeneflowincreasedtheprobabilityof identifyingthetruespeciesusingconcatenationdropped tozero,althoughthisvalueonlydecreasedto74%using MDC[11].Wefoundsimilarresults,withthe*BEAST andconcatenatedtreeshavingsimilartopologies,but withnodesupportvaryingwidely.Ourresultssuggest thatspeciestreeestimationmethodsthataccountfor genetreeuncertaintyamonglociwithdivergedhistories,withsignalsoflineagesortingandrecombination mayresultinfewerwell-supportedspeciesthanconcatenation,especiallyamongcloselyrelatedfungaltaxa. Twoofthelociweemployedinthisstudy(Flank-F3 andOPA1-3)revealedstrongevidenceforahistoryof recombination.Asexual Alternaria speciesarethought tobederivedfromsexualancestors[71],ashasbeen suggestedforotherasexualascomycetes[98],andone species A.infectoria hasbeenconnectedtoa Lewia teleomorph[34].Althoughti mescalesforourancestral recombinationgraphs(ARG)arenotpossible,theARG forFlank-F3showedthreepossiblerecombinationevents wherecitrus2haplotypesarepossiblyderivedfromcitrus lineages1and3.Itisdifficulttodatetheseputativerecombinationeventsandtoknowiftheseeventsarethe resultofhistoricorcurrentsexualorasexualrecombination.Signaturesofrecombinationwerefoundinthe citrus1lineageinan Alternaria brownspotfungus Figure4 Bayesianphylogeniesestimatedforeachlocus,(A) endoPG ,(B)OPA2-1,(C)Flank-F3,and(D)OPA1-3. Boldbranches highlightnodesthatweresupportedby70%bootstrapvalues (maximumlikelihood,1000replications)and95posterior probabilities.Numbersinparenthesesrepresentthenumberof isolatesofeachhaplotype. EndoPG haplotypessampledfromthe citrus1,citrus2andcitrus3phylogeneticlineagesareindicatedby redcircles,bluesquares,andyellowstars,respectively. Table2PairwiseShimodaira-Hasegawatestsoftopological congruenceamongphylogeniesOPA2-1OPA1-3Flank-F3 endoPG 0.059a0.0000.063 OPA2-10.0000.128 OPA1-30.000aprobabilityofdifferenttopologiesamong2000bootstrappeddatasets generatedusingRELLsampling.Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page10of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 11

populationinFlorida[29]butthemechanismofthisrecombinationisnotknown. Alternaria maybeabletorecombinethroughparasexualand/orsexualmeans[29], whichconcurswiththefindingsthatmatingtypegenesin A.alternata areexpressed[5]andunderstrongpurifying selection[99].Furtherstudieswillbeneededtodetermine if Alternariaalternata sensulatoiscapableofforminga sexualstagethanhasheretoforebeenoverlooked.Ifso,resultsfromourARGmayrepresentcontemporaryrather thanhistoricalrecombination,althoughdatingrecombinationeventsintheARGisnotpossible.ConclusionsSpeciesdelimitationisimportantforthestudyofthe evolutionofpathogenicityan dtheemergenceofinfectious diseases.Further,thedelimitationofspeciesalsoplaysa criticalroleinglobalbiosecuritybyprovidingguidelinesfor restrictionsonthemovementofplantpathogensamong countrieswhichhasnationaland internationalsignificance [100,101].Thethreatofmovementofintroducedpathogensaroundtheworldhasresultedinthequarantineof manycropsortherejectionofexportedcrops.Incorrectly naminganewspeciesorwronglyidentifyingaspeciescan resultinsignificanteconomiclosses[102].In2001 – 2002, shipmentsofLiYapearimportedfromChinawererejected becauseofsignsof Alternaria spp.infection,whichalso occurredinAustralia,NewZealand,andUnitedKingdom. Tworelatedspecies, A.alternata and A.gaisen ,were includedinapestriskassessmentandfurtherresearch showedthatisolatessampledfromLiYapearweremorphologicallydistinctandgivenanewname, A.yaliinficiens [100].Ourresultsindicatethatcitrusbrownspotiscaused byamaximumoftwospeciesof Alternaria ,andthat taxonomicrevisionof Alternaria infectingcitrus,based oncongruentmorphological andgeneticanalyses,is needed.Oneofthesespecies(encompassinglineages citrus1andcitrus3)isfoundworldwidebutthesecond (lineagecitrus2)specieshasonlybeenfoundinthe Americas.Itisnotyetknownifphenotypicdifferences inaggressiveness,hostrange ,orgrowthratesexistbetween thesespecies,butifso,limitingthemovementofthecitrus 2speciesintonewcountriesmightbewarranted.This studyhighlightstheneedfortheuseofthesenewmethods toaccuratelyidentifycloselyrelated,morphologicallyindistinguishablespeciesthatareimportantinagricultureand potentiallyofregulatoryinterest.AvailabilityofsupportingdataThedatasetssupportingtheresultsofthisarticle areavailableinGenBankunderaccessionnumbers KF699389-KF699527. Figure5 Bayesianphylogenyderivedfromtheconcatenated datasetwithatotalof1672nucleotidesfromfourlociincluding the endo-polygalacturonase gene,SCARmarkersOPA2-1and OPA1-3,andonemicrosatelliteflankingregionAA-Flank-3. Nodesupportinposteriorprobabilityisindicatedforeachnode. Figure6 Speciestreeestimationsamongaworldwidesample ofcitrusbrownspotpathogenspeciesgroupsusing(A)the “ minimizedeepcoalescent ’ (MDC)approach,andahierarchical Bayesianmodelasimplementedin(B)*BEAST. Numbersabove branchesindicatenodesupportasposteriorprobabilities. Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page11of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 12

AdditionalfileAdditionalfile1:TableS1. Isolatesusedinthestudy[103-105]. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors ’ contributions JESandTLPconceivedofthestudy.JEScarriedoutmoleculargenetic studiesandbioinformatics,andanalyses.JES,TLP,LPT,CML,andBMPwrote andeditedthemanuscript.Allauthorsreadandapprovedthefinal manuscript. Acknowledgements WethankDr.MarionAndrew,Dr.FrancescaRotondoandKalynA.Thomas forsequencingassistance.WealsothankDr.ZaidAbdoforcriticallyreading andforcommentsthatvastlyimprovedthismanuscript,andDr.Hilal zkilinforfruitfuldiscussions.ThisresearchwasfundedbyNSFDEB0918668. Authordetails1DepartmentofPlantPathology,WashingtonStateUniversity,Pullman,WA, USA.2CitrusResearchandEducationCenter,UniversityofFlorida,LakeAlfred, FL,USA.3VirginiaBioinformaticsInstitute,Blacksburg,VA,USA.4Divisionof PlantPathologyandMicrobiology,SchoolofPlantSciences,Universityof Arizona,Tucson,AZ,USA.5Currentaddress:DepartmentofPlantPathology, UniversityofGeorgia,Athens,Georgia. Received:9October2013Accepted:21February2014 Published:3March2014 References1.PerkinsDD: InPraiseofDiversity. In MoreGeneManipulationsinFungi. EditedbyBennettJW,LasureLL.NewYork:AcademicPress;1991:3 – 26. 2.HarringtonTC,RizzoDM: Definingspeciesinthefungi. In Structureand DynamicsofFungalPopulations. EditedbyWorrallJJ.Dordrecht,The Netherlands:KluwerPress;1999:43 – 71. 3.MaynardSmithJ,SzathmryE: TheMajorTransitionsinEvolution. London: OxfordUniversityPress;1995. 4.KohnLM: Mechanismsoffungalspeciation. AnnuRevPhytopathol 2005, 43: 279 – 308. 5.TaylorJW,JacobsonDJ,KrokenS,KasugaT,GeiserDM,HibbettDS,Fisher MC: Phylogeneticspeciesrecognitionandspeciesconceptsinfungi. FungalGenetBiol 2000, 31: 21 – 32. 6.HibbettDS,TaylorJW: Fungalsystematics:isanewageofenlightenment athand? NatRevMicrobiol 2013, 11: 129 – 133. 7.HawsworthDL: IntegratingMorphologicalandMolecularDatainFungal Systematics .In SystematicsandEvolutionofFungi. EditedbyMisraJK, TewariJP,DeshmukhSK.NewYork,NY:CRCPress;2012:1 – 14. 8.KohnLM: Theclonaldynamicinwildandagriculturalplantpopulations. CanJBot 1995, 73: S1231 – S1240. 9.TaylorJW,JacobsonDJ,FisherMC: Theevolutionofasexualfungi: reproduction,speciation,andclassification. AnnuRevPhytopathol 1999, 37: 197 – 246. 10.CarstensBC,KnowlesLL: Estimatingspeciesphylogenyfromgene-tree probabilitiesdespiteincompletelineagesorting:Anexamplefrom Melanoplusgrasshoppers. SystBiol 2007, 56: 400 – 411. 11.EckertAJ,CarstensBC: Doesgeneflowdestroyphylogeneticsignal?The performanceofthreemethodsforestimatingspeciesphylogeneiesin thepresenceofgeneflow. MolPhylogenetEvol 2008, 49: 832 – 842. 12.KubatkoLS,CarstensBC,KnowlesLL: STEM:speciestreeestimationusing maximumlikelihoodforgenetreesundercoalescence. Bioinformatics 2009, 25: 9 – 973.13.KubatkoLS,DegnanJH: Inconsistencyofphylogeneticestimatesfrom concatenateddataundercoalescence. SystBiol 2007, 56: 17 – 24. 14.MaddisonWP: Genetreesinspeciestrees. SystBiol 1997, 46: 523 – 536. 15.HartlDL,ClarkAG: PrinciplesofPopulationGenetics. 4thedition.Sinauer Associates,Inc:Sunderland,MA;2007. 16.DegnanJH,SalterLM: Genetreedistributionsunderthecoalescent process. Evolution 2005, 9: 24 – 37. 17.MaddisonWP,KnowlesLL: Inferringphylogengydespiteincomplete lineagesorting. SystBiol 2006, 55: 410 – 431. 18.LuiL,PearlDK: Speciestreesfromgenetrees:ReconstructingBayesian posteriordistributionsofaspeciesphylogenyusingestimatedgenetree distributions. SystBiol 2007, 56: 504 – 514. 19.NiemillerML,FitzpatrickBM,MillerBT: Recentdivergencewithgeneflow inTennesseecavesalamanders(Plethodontidae:Gyrinophilus)inferred fromgenegenealogies. MolEcol 2008, 17: 2258 – 2275. 20.BrumfieldRT,LiuL,LumDE,EdwardsSV: Comparisonofspeciestree methodsforreconstructingthephylogenyofbeardedmanikins (Aves:Pipridae: Manacus )frommultilocussequencedata. SystBiol 2008, 57: 9 – 731. 21.CranstonKA,HurwitzB,WareD,SteinL,WingRA: Speciestreesfrom highlyincongruentgenetreesinrice. SystBiol 2009, 58: 489 – 500. 22.HenkDA,EagleCE,BrownK,VanDenBergMA,DyerPS,PetersonSW,Fisher MC: Speciationdespitegloballyoverlappingdistributionsin Penicillium chyrysogenum :thepopulationgeneticsofAlexanderFleming ’ slucky fungus. MolEcol 2011, 20: 4288 – 4301. 23.PollardDA,IyerVN,MosesAM,EisenMB: Widespreaddiscordanceofgene treeswithspeciestreeinDrosophila:evidenceforincompletelineage sorting. PLoSGenet 2006, 2: e173. 24.BrowneAGP,FisherMC,HenkDA:Species-specificPCRtodescribelocal-scale distributionoffourcrypticspeciesinthe Penicilliumchrysogenum complex. FungalEcol 2013, 6: 419 – 429. 25.PamiloP,NeiM: Relationshipsbetweengenetreesandspeciestrees. MolBiolEvol 1988, 5: 568 – 583. 26.LeacheAD: Speciestreediscordancetracestophylogeographicclade boundariesinNorthAmericanfencelizards( Sceloporus ). SystBiol 2008, 56: 547 – 559. 27.LinnenCR: Speciestreeestimationforcomplexdivergencehistories:a casestudyinNeodiprionsawflies .InEstimatingSpeciesTrees.Editedby KnowlesLL,KubatkoL. NewYork:Wiley-Blackwell;2010. 28.AndrewM,PeeverTL,PryorBM: Anexpandedmultilocusphylogenydoes notresolvemorphologicalspecieswithinthesmall-spored Alternaria speciescomplex. Mycologia 2009, 101: 95 – 109. 29.StewartJE,ThomasKA,LawrenceCB,DangH,PryorBM,PeeverTL: Signatures ofrecombinationinclonallineagesofthecitrusbrownspotpathogen, Alternariaalternata sensulato. Phytopathology 2013, 103: 741 – 749. 30.PierceNB: Blackrotoforanges. BotGazelle 1902, 33: 234 – 23. 31.NeergaardP: DanishSpeciesofAlternariaandStemphylium. London:Oxford UniversityPress;1945:560. 32.EllisMB: DematiaceousHyphomycetes. Kew:CommonwealthMycological Institute:19. 33.KusabaM,TsugeT: NuclearribosomalDNAvariationandpathogenic specializationin Alternaria fungiknowntoproducehost-specifictoxins. ApplEnvironMicrobiol 1994, 60: 3055 – 3062. 34.SimmonsEG: Alternariathemesandvariations(226 – 235):Classificationof citruspathogens. Mycotaxon 1999, 70: 263 – 323. 35.PeeverTL,IbaezA,AkimitsuK,TimmerLW: WorldwidePhylogeography ofthecitrusbrownspotpathogen, Alternariaalternata Phytopathology2002, 92: 794 – 802. 36.PeeverTL,SuG,Carpenter-BoggsL,TimmerLW: Molecularsystematicsof citrus-associated Alternaria spp. Mycologia 2004, 96: 119 – 134. 37.KingmanJFC: Thecoalescent. StochProcessAppl 1982, 13: 235 – 248. 38.PeeverTL,CanihosY,OlsenL,IbanezA,LiuY – C,TimmerLW: Population geneticstructureandhostspecificityofAlternariaspp.causingbrown spotonMinneolatangeloandroughlemoninFlorida. Phytopathology 1999, 92: 794 – 802. 39.IsshikiA,AkimitsuK,YamamotoM,YamamotoH: Endopolygalacturonaseis essentialforcitrusblackrotcausedby Alternariacitri butnotbrownspot causedby Alternariaalternata MolPlant-MicrobeInteract 2001, 14: 749 – 757. 40.PeeverTL,Carpenter-BoggsL,TimmerLW,CarrisLM,BhatiaA: Citrusblack rotiscausedbyphylogeneticallydistinctlineagesof Alternariaalternata Phytopathology 2005, 95: 512 – 518. 41.SuG,PeeverTL,TimmerLW: Molecularsystematicsofcitrus-associated Alternariasp Phytopathology 2001, 91: S190. 42.AndrewM: Molecularsystematicsofsmall-sporedAlternariaspecies. MS Thesis:WashingtonStateUniversity;2006. 43.Tran-DinhN,HockingA: Isolationandcharacterizationofpolymorphic microsatellitemarkersfor Alternariaalternata MolEcolResour 2006, 6: 405 – 407.Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page12of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 13

44.RozenS,SkaletskyH,KrawetzS,MisenerS: Primer3ontheWWWfor generalusersandforbiologistprogrammers InbioinformaticsMethods andProtocols. 2000:365 – 386. 45.GuindonS,GascuelO: Asimple,fast,andaccuratealgorithmtoestimate largephylogeniesbymaximumlikelihood. SystBiol 2003, 52: 696 – 707. 46.HuelsenbeckJP,RonquistF: MrBayes:Bayesianinferenceofphylogenetic trees. Bioinformatics 2001, 17: 754 – 755. 47.MininV,AbdoZ,JoyceP,SullivanJ: Performance-basedselectionof likelihoodmodelsforphylogenyestimation. SystBiol 2003, 52: 674 – 683. 48.HeledJ,DrummondAJ: Bayesianinferenceofspeciestreesfrom multilocusdata. MolBiolEvol 2010, 2010 (27) : 570 – 580. 49.ShimodairaH,HasegawaM: Multiplecomparisonsofloglikelihoods withapplicationstoph ylogeneticinference. MolBiolEvol 1999, 16: 1114 – 1116. 50.SwoffordDL: PAUP*.PhylogeneticAnalysisUsingParsimony(*andOtherMethods). Version4. Sunderland,Massachusetts :SinauerAssociates;2003. 51.HallTA: BioEdit:auser-friendlybiologicalsequencealignmenteditorand analysis. NucleicAcidsSympSer 1999, 41: 95 – 98. 52.LibradoP,RozasJ: DnaSPv5:Asoftwareforcomprehensiveanalysisof DNApolymorphism. Bioinformatics 2009, 25: 1451 – 1452. 53.RozasJ,Sanchez-DelJ,BarrioJC,MesseguerX,RozasR: DnaSP,DNA polymorphismanalysesbythecoalescentandothermethods. Bioinformatics 2003, 19: 2496 – 2497. 54.HudsonRR,KaplanNL: Statisticalpropertiesofthenumberof recombinationeventsinthehistoryofasampleofDNAsequences. Genetics 1985, 111: 147 – 164. 55.JakobsenIB,EastealS: Aprogramforcalculatinganddisplaying compatibilitymatricesasanaidindeterminingreticulateevolutionin molecularsequences. ComputApplBiosci 1996,12: 291 – 295. 56.PriceEW,CarboneI: SNAP:workbenchmanagementtoolforevolutionary populationgeneticsanalysis. Bioinformatics 2005, 21: 402 – 404. 57.AylorDL,PriceEW,CarboneI: SNAP:combineandmapmodulesfor multilocuspopulationgeneticsanalysis. Bioinformatics 2006, 22: 1399 – 1401. 58.HeyJ,WakelyJ: Acoalescentestimatorofthepopulationrecombination rate. Genetics 1997, 145: 833 – 846. 59.GriffithsRC,MarjoramP: AncestralinferencefromsamplesofDNA sequenceswithrecombination. JComputBiol 1996, 3: 479 – 502. 60.LawrenceDP,GannibalPB,PeeverTL,PryorBM: ThesectionsofAlternaria: formalizingspecies-groupconcepts. Mycologia 2013, 105: 530 – 546. 61.BahloM,GriffithsRC: Inferencefromgenetreesinasubdivided population. TheorPopulBiol 2000, 57: 79 – 95. 62.GriffithsRC,TavarS: Ancestralinferenceinpopulationgenetics. StatSci 1994, 9: 307 – 319. 63.LyngsoRB,SongYS,HeinJ: MinimumRecombinationHistoriesbyBranch andBound. In AlgorithmsinBioinformatics,Proceedings. Berlin:Springer-Verlag Berlin;2005:239 – 250. 64.OliverJC: AUGIST:inferringspeciestreeswhileaccommodatinggene treeuncertainty. Bioinformatics 2008, 24: 2932 – 2933. 65.AviseJC,BallRMJr: Principlesofgenealogicalconcordanceinspecies conceptsandbiologicaltaxonomy. OxfSurvEvolBiol 1990, 7: 45 – 67. 66.AlfaroME,ZollerS,LutzoniF: Bayesorbootstrap?Asimulationstudying comparingtheperformanceofBayseianmarkovchainMonteCarlo samplingandbootstrappinginassessingphylogeneticconfidence. Mol BiolEvol 2003, 20: 255– 266. 67.MaddisonWP,MaddisonDR: Mesquite:amodularsystemforevolutionary analysis.Version2.73. 2010.http://mesquiteproject.org. 68.YuleGU: Amathematicaltheoryofevolution,basedontheconclusions ofDr.J.C.Willis,F.R.S. PhilosTransRSocLondBiol 1925, 213: 21 – 87. 69.RannalaB,YangZ: Probabilitydistributionofmolecularevolutionary tree:anewmethodsforphy logeneticinference. JMolEvol 1996, 43: 304 – 311. 70.RambautA: FigTree,agraphicalviewerofphylogenetictrees. 2008. http://tree.bio.ed.ac.uk/software/figtree. 71.BerbeeML,PayneBP,ZhangG,RobertsRG,TurgeonBG: SharedITSDNA substitutionsinisolatesofoppositematingtyperevealarecombining historyforthreepresumedasexualspeciesinthefilamentous ascomycetegenus Alternaria MycolRes 2003, 107: 169 – 182. 72.KasugaT,WhiteTJ,TaylorJW: Estimationofnucleotidesubstitutionrates inEurotiomycetefungi. MolBiolEvol 2002, 19: 2318 – 2324. 73.NishimuraS,KohmotoK: Host-specifictoxinsandchemicalstructures from Alternaria species. AnnuRevPhytopathol 1983, 21: 87 – 116. 74.ChaverriP,CastleburyLA,SamuelsGJ,GeiserDM: Multilocusphogenetics structurewithinthe Trichodermaharzianum / Hypocrealixii complex. MolPhylogenetEvol 2003, 27: 302 – 313. 75.MillerAN,HuhndorfSM: Usingphylogeneticspeciesrecognitionto delimitspeciesboundariesandspeciesrelationshipswithin Lasiosphaeria Mycologia 2004, 96: 302 – 313. 76.KohmotoK,OtaniH,TsugeT: Alternariaalternata Pathogens. In PathogenesisandHostSpecificityinPlantDiss:Histopathological,Biochemical, GeneticandMolecularBases.Vol.II,Eukaryotes. EditedbyKohmotoK,Singh US,SinghRP.Oxford:PergamonPress;1995:51 – 63. 77.TimmerLW,PeeverTL,SolelZ,AkimitsuK: Alternaria diseasesofcitrus – Novel pathosystems. PhytopatholMediterrian 2003,42: 3 – 16. 78.NishimuraS,KohmotoK,OtaniH,RamachandranP,TamuraF: Pathological andepidemiologicalaspectsof Alternariaalternata infectiondepending onahost-specifictoxin. In Plantinfection:thePhysiologicalandBiochemical Basis. EditedbyAsadaY,BushnellWR,OuchiS,VanceCP.Tokyo;Berlin: JapanScientificSocietiesPress;Springer-Verlag;1982:199 – 21. 79.AkamatsuH,TagaM,KodamaM,JohnsonR,OtaniH,KohmotoK: MolecularkaryotypesforAlternariaplantpathogensknowntoproduce host-specifictoxin. CurrGenet 1999, 35: 647 – 656. 80.MasunakaA,OhtaniK,PeeverTL,TimmerLW,TsugeT,YamamotoM, AkimitsuK: AnisolateofAlternariaalternatathatispathogenictoboth tanginessandroughlemonproducestwohost-selectivetoxinsACTand ACRtoxins. Phytopathology 2005, 95: 241 – 247. 81.MasunakaA,TanakaA,TsugeT,PeeverTL,TimmerLW,YamamotoM, YamamotoH,AkimitsuK: DistributionandcharacterizationofAKT homologsinthetangerinepathotypeof Alternariaalternata Phytopathology 2000, 90: 762 – 768. 82.TsugeT,HarimotoY,AkimitsuK,OhtaniK,KodamaM,AkagiY,MayumiE, YamamotoM,OtaniH: Host-selectivetoxinsproducedbytheplant pathogenicfungus Alternariaalternata FEMSMicrobiolRev 2013, 37: 44 – 66. 83.LiW-H,TanimuraM,SharpPM: Anevaluationofthemolecularclock hypothesisusingmammalianDNAsequences. JMolEvol 1987, 25: 330 – 342. 84.ScoraRW: Onthehistoryandoriginofcitrus. TorreyBotClub 1975, 102: 369 – 375. 85.O ’ DonnellK: Molecularphylogenyofthe Nectriahaematococca — Fusarium solani speciescomplex. Mycologia 2000, 92: 919 – 938. 86.O ’ DonnellK,KistlerHC,TackeBK,CasperHH: Genegenealogiesreveal globalphylogeographicstructureandreproductiveisolationamong lineagesof Fusariumgraminearum ,thefunguscausingwheatscab. ProcNatlAcadSciUSA2000, 97: 7905 – 7910. 87.SteenkampET,WingfieldBD,DesjardinsAE: Crypticspeciationin Fusarium subglutinans Mycologia 2002, 94: 1032 – 1043. 88.KoufopanouV,BurtA,TaylorJW: Concordanceofgenegenealogies revealsreproductiveisolationinthepathogenicfungus Coccidioides immitis ProcNatlAcadSciUSA 1997, 94: 5478 – 5482. 89.KasugaT,TaylorJW,WhiteTJ: Phylogeneticrelationshipsofvarietiesand geographicalgroupsofthehumanpathogenicfungus, Histoplasma capsulatum Darling. JClinMicrobiol 1999, 37: 653 – 663. 90.CruseM,TelerantR,GallagherT,LeeT,TaylorJW: Crypticspeciesin Stachybotryschartarum Mycologia 2002, 94: 814 – 822. 91.FisherMC,KoenigG,WhiteTJ,TaylorJW: Molecularandphenotypic descriptionof Coccidioidesposadasii sp.nov.,previouslyrecognizedas thenon-Californiapopulationof Coccidioidesimmitis Mycologia 2002, 94: 73 – 84. 92.CouchBC,KohnLM: Amultilocusgenegenealogyconcordantwithhost preferenceindicatessegregationofanewspecies, Magnaportheoryzae fromM.grisea. Mycologia 2002, 94: 683 – 693. 93.O ’ DonnellK,WardTJ,GeiserDM,KistlerHC,AokiT: Genealogicalconcordance betweenthematingtypelocusandsevenothernucleargenessupports formalrecognitionofninephylogeneticallydistinctspecieswithinthe Fusariumgraminearumclade. FungalGenetBiol 2004, 41: 600 – 623. 94.DegnanJH,RosenbergNA: Discordanceofspeciestreeswiththeirmost likelygenetrees. PLoSGenet 2006, 2: 762 – 768. 95.RosenbergNA,TaoR: Discordanceofspeciestreeswiththeirmostlikely genetrees:thecaseoffivetaxa. SystBiol 2008,57: 131 – 140. 96.SenD,BrownCJ,TopEM,SullivanJ: Inferringtheevolutionaryhistoryof IncP-1plasmidsdespiteincongruenceamongbackbonegenetrees. MolBiolEvol 2013, 30: 154 – 166. 97.BelfioreNM,LiuL,MoritzC: Multilocusphylogeneticsofarapidradiation inthegenus Thomomys (Rodentia:Geomyidae). SystBiol 2008, 57: 294.Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page13of14 http://www.biomedcentral.com/1471-2148/14/38

PAGE 14

98.YunSH,BerbeeML,YoderOC,TurgeonBG: Evolutionofthefungalself-fertile reproductivelifestylefromself-sterileancestors. ProcNatlAcadSciUSA 1999, 96: 5592 – 5597. 99.StewartJE,KawabeM,AbdoZ,ArieT,PeeverTL: ContrastingCodonusage patternsandpurifyingselectionatthematinglocusinputatively asexual Alternaria fungalspecies. PLoSONE 2011, 6: e20083. 100.RobertsRG: Alternariayaliinficiens sp.nov.onYaLipearfruit:from interceptiontoidentification. PlantDis 2005, 89: 134 – 145. 101.BoykinLM,ArmstrongKF,KubatkoL,DeBarroP: Speciesdelimitationand globalbiosecurity. EvolBioinformatics 2012, 8: 1 – 37. 102.RossmanAY,Palm-HernndezME: Systematicsofplantpathogenetic fungi:whyitmatters? PlantDis 2008, 92: 1376 – 1386. 103.ElenaK: AlternariabrownspotofMinneolainGreece;evaluationofcitrus speciessusceptibility. PlantPatholEurJPlantPathol 2006, 115: 259 – 262. 104.GolmohammadiM,AndrewM,PeeverTL,PeresNA,TimmerLW: Firstreport ofbrownspotoftangerinehybridcultivarsMinneola,pageandfortune causedby Alternariaalternata inIran. PlantPathol 2006, 12: 2005 – 2098. 105.MarnJE,FernndezH,PeresNA,AndrewM,PeeverTL,TimmerLW: First reportofAlternariabrownspotofcitruscausedby Alternariaalternata inPeru. PlantDis 2006, 90: 686.doi:10.1186/1471-2148-14-38 Citethisarticleas: Stewart etal. : Discordbetweenmorphologicaland phylogeneticspeciesboundaries:incompletelineagesortingand recombinationresultsinfuzzyspeciesboundariesinanasexualfungal pathogen. BMCEvolutionaryBiology 2014 14 :38. Submit your next manuscript to BioMed Central and take full advantage of: € Convenient online submission € Thorough peer review € No space constraints or color “gure charges € Immediate publication on acceptance € Inclusion in PubMed, CAS, Scopus and Google Scholar € Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Stewart etal.BMCEvolutionaryBiology 2014, 14 :38 Page14of14 http://www.biomedcentral.com/1471-2148/14/38