alleledmcl

Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32OpenAccessRESEARCHARTICLEBioMedCentral2010Williamsetal;licenseeBioMedCentralLtd.
ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(http://creativecommons.
org/licenses/by/2.
0),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
ResearcharticleSNPidentification,verification,andutilityforpopulationgeneticsinanon-modelgenusLarissaMWilliams1,XinMa2,AdamRBoyko2,CarlosDBustamante2andMarjorieFOleksiak*3AbstractBackground:BytargetingSNPscontainedinbothcodingandnon-codingareasofthegenome,weareabletoidentifygeneticdifferencesandcharacterizegenome-widepatternsofvariationamongindividuals,populationsandspecies.
Weinvestigatedtheutilityof454sequencingandMassARRAYgenotypingforpopulationgeneticsinnaturalpopulationsoftheteleost,FundulusheteroclitusaswellascloselyrelatedFundulusspecies(F.
grandis,F.
majalisandF.
similis).
Results:Weused454pyrosequencingandMassARRAYgenotypingtechnologytoidentifyandtype458genome-wideSNPsanddeterminegeneticdifferentiationwithinandbetweenpopulationsandspeciesofFundulus.
Specifically,pyrosequencingidentified96putativeSNPsacrosscodingandnon-codingregionsoftheF.
heteroclitusgenome:88.
8%wereverifiedastrueSNPswithMassARRAY.
Additionally,putativeSNPsidentifiedinF.
heteroclitusESTsequenceswereverifiedinmost(86.
5%)F.
heteroclitusindividuals;fewerweregenotypedinF.
grandis(74.
4%),F.
majalis(72.
9%),andF.
similis(60.
7%)individuals.
SNPswerepolymorphicandshowedlatitudinalclinalvariationseparatingnorthernandsouthernpopulationsandestablishedisolationbydistanceinF.
heteroclituspopulations.
InF.
grandis,SNPswerelesspolymorphicbutstillestablishedisolationbydistance.
Markersdifferentiatedspeciesandpopulations.
Conclusions:Intotal,theseapproacheswereusedtoquicklydeterminedifferenceswithintheFundulusgenomeandprovidemarkersforpopulationgeneticstudies.
BackgroundHighthroughputsequencingandgenotypinghasbecomeincreasinglyfaster,lessexpensiveandmoreaccurate.
Inrecentyearsthishasleadtotheestablishmentofmyriaddatasetsrangingfromincreasedcoverageofvariationinthehumangenomeattheindividuallevel[1-5]tothesequencingofnon-modelprokaryoticandeukaryoticgenomesandtranscriptomes[6-11].
Formanyorganismssequencingofentiregenomesisstillunattained,butsmaller,moretargetedportionsofthegenomecanbeeasilysequencedandgenotyped.
Suchdatacanprovidegenome-widesequenceinformationwhichcanbeusedtocharacterizepopulationandselectionpressureparame-tersaswellasprovideevolutionaryinsightsthatarebroadlyapplicable[12].
Onenon-modelgenus,Fundulus,includescloselyrelatedspeciesthatrangeinphysiology,environmentalandhabitatpreference,andgeographiclocales;FundulusheteroclitusandFundulusmajalisinhabittheAtlanticcoast,andFundulusgrandisandFundulussimilisinhabittheGulfCoast.
ManyFundulusspeciesand/orpopula-tionshaveextensiveeuryhalinecapabilities,respondwelltovaryingrangesofhypoxia[13-15],livealongasteepthermocline,andhaveadaptedtoextremelypollutedareas[16].
Avarietyofstudieshaveinvestigatedtheunderlyinggeneticbasisofthisteleosts'phenotypicplas-ticity.
WhilesomeofthetranscriptomeisknownforF.
heteroclitus[17-27]muchofthegenome-widevariationwithinandbetweenpopulationsandspeciesforthisgenusisrelativelyunknown.
Establishingasetofgeneticmarkers,whichcanbeusedtoassessregionsofthegenomeinvolvedinlocaladapta-tionandinspeciationisimportanttounderstandfunda-mentalsimilaritiesanddifferencesbetweenpopulationsandspeciesofFundulus.
Oncemarkersareestablishedtheycanbefurtherstudiedtolookforsignaturesofselec-tiontoanynumberofevolutionaryforces(e.
g.
,pollution,hypoxia,salinity,temperature).
Afewstudieshaveestab-*Correspondence:moleksiak@rsmas.
miami.
edu3RosenstielSchoolofMarineandAtmosphericSciences,UniversityofMiami,4600RickenbackerCauseway,Miami,FL33149,USAFulllistofauthorinformationisavailableattheendofthearticleWilliamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page2of14lishedgeneticdifferencesbetweenpopulationsofF.
het-eroclitusmainlywithrespecttophylogeographicconstraints[28,29]orselection[30-38].
Thesestudiesusedmicrosatellite,mitochondrialDNA,andAFLPanal-ysesaswellastargetedgeneapproaches.
Singlenucle-otidepolymorphisms(SNPs)areausefulstartingpointtoscanlargeanddisparateregionsofthegenomeduetotheirabundanceinbothcodingandnon-codingregions,theirco-dominantnature,andlackofambiguity.
SNPshavebeenusedtoestablishdifferencesbetweenindividuals[39],populations[40-42]andspecies[43,44].
Theyalsoareusefulmarkersforpropensitytodisease[45-47],diseasestates[48],andevidenceofthegeneticbasisofadaptation[49-52].
Invertebrates,aSNPoccursonaverageevery100to1000basepairsandoftenisinlinkagedisequilibriumwithmanyotherSNPsalongthechromosome,formingstronghaplotypes,whichcanbeeasilyidentified[53].
Unfortunately,SNPresourcesarenotreadilyavailableinthemajorityofnon-modelspecieslackinggenomicresources.
Withthisinmind,wesetouttoestablishasetofSNPmarkerstoidentifydifferencesbetweenFunduluspopulationsandspecies.
MethodsSampleCollectionandExtractionF.
heteroclituswerecollectedusingminnowtrapsduringthespringof2005.
Spleenandtestesweresampledfrom20individualsfromeachoftencollectionsitesalongtheEastcoastoftheUnitedStates(Figure1).
F.
grandiswerecollectedusingminnowtrapsduringthewinterof2009(Figure1).
Finclipsweresampledfrom15individualsfromeachofthesixcollectionsitesalongtheGulfCoastoftheUnitedStates.
SpleenfromF.
majaliswasextractedfrom13individualsfromWoodsHole,Massachusettsand10individualsfromSapeloIsland,GA.
SpleenalsowasextractedfromF.
similiscollectedfromPensacola,Florida(3individuals)andCorpusChristi,Texas(8indi-viduals).
GenomicDNAfromspleenandtesteswasextractedbyphenolandchloroformasdescribedinWirginetal.
[54],andDNAwasresuspendedin50μL0.
1*TEbuffer.
GenomicDNAsfromfinclipswereextractedusingamodifiedversionofAljanabiandMartinez[55]andDNAwasresuspendedin50μL0.
1*TEbuffer.
Thisexperi-mentwasperformedaccordingtoanapprovedInstitu-tionalAnimalCareandUseCommitteeatNorthCarolinaStateUniversity.
DNAPyrosequencingF.
heteroclitusgenomicDNAs(500ng)fromeightindi-vidualsineachoftencollectionsites(allsitesexceptPointJudith,RI,Figure1A)weredigestedindividuallywith1UBspE1(NewEnglandBiolabs,MA)and1UEcoRI(NewEnglandBiolabs,MA).
Sampleswereincu-batedforthreehoursat37°Cinatotalvolumeof30μLcontainingBuffer3(NewEnglandBiolabs,MA).
Adap-tors(Table1)toeachoftherestrictionsites,25mMATP,and1UofT4DNAligase(Epicentre)wereaddedtoreac-tionsandincubatedat16°Covernight.
A2'O-methylblockwasaddedtothe3'cytosinebaseontheadapter.
ThisblockassuredthatonlythosefragmentsdigestedwithbothBspEIandEcoRIwouldbeamplifiedwithPCRandpreventedamplificationoffragmentswiththesametypeofrestrictionsiteonbothendsofthefragment.
PreselectivePCRreactionswithprimersspecifictoadaptors(Table1)wereperformedinatotalvolumeof25μLcontaining2μLofdiluted(1:10in0.
1*Tris-EDTAbuffer)ligationproductwithEcoRIprimer(IntegratedDNATechnologies;10pmol),BspE1primer(IntegratedDNATechnologies;10pmol)and1UTaq.
PCRcondi-tionswere20cyclesof94°Cfor10sec,49°for30sec,and72°Cforonemin.
Followingthepreselectiveamplifica-tion,aselectiveamplificationwascarriedouttodecreasethenumberoffragmentsamplifiedineachindividualtoapproximately200byextendingtheprimeronthe3'end.
PreselectivePCRproductswerediluted(1:10)and2μLofdilutedproductwasamplifiedwithprimers(Table1)toEcoRI+AAG(IntegratedDNATechnologies;10pmol)andBspEI+C(IntegratedDNATechnologies;10pmol)with1UTaqina25μLtotalvolume.
PCRconditionsinthefirstcyclewere94°Cfor10sec,65°Cfor30sec,and72°Cforoneminutewiththeannealingtemperaturereducedby0.
5°Cfor20cycles,then25cyclesof94°Cfor10sec,55°Cfor30sec,and72°Cforoneminute.
Primers(Table1)specifictotheEcoRIrestrictionsiteweregeneratedwiththegoalsoflabelingtheDNAfrag-mentsfromeachindividualwithspecificnucleotidebar-codes[56]andpreparingthosesamplesforemulsion-basedamplification.
Startingatthe5'end,19nucleotides(Table1)complementarytotheprimerontheDNAcap-turebeadsusedintheemulsionPCRreaction[57]weresynthesized(IntegratedDNATechnologies).
Followingthosenucleotides,eachprimerhadadistinct10basepairbarcode[56]usedtoidentifyindividuals(tenprimersintotal).
Thefinal19basepairsoftheprimerwerespecifictotheEcoRIadapter.
TheBspE1primer(Table1)startedatits5'endwith19nucleotides(Table1),whichwerecomplementarytotheprimerontheDNAcapturebeadsfollowedby18basepairsspecifictotheBspE1adapter(Figure2).
AllprimerswereHPLCpurified.
Amplifiedselectivefragmentswerediluted(1:10)andaddedtobothEcoRIandBspE1primers(IntegratedDNATechnologies;10pmol)ina25μLvolume.
PCRconditionswere94°Cfor10sec,50°Cfor30sec,and72°Cforoneminuteandwerecarriedoutfor30cycles.
PCRreactionswerepooledintoeightwells,whereeachofthetendistinctbarcodeswasrepresentedonlyonceineachofthepools.
EachpoolofPCRproductswaspurifiedusingQIAquickPCRPuri-Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page3of14Figure1SamplingsitesforFundulusspecies.
F.
heteroclituswascollectedalongtheeastcoastoftheUnitedStatesandF.
grandiswascollectedalongtheGulfofMexicocoast.
Wiscasset,MESandwich,MANewBedford,MAPointJudith,RIClinton,CTNewarkBay,NJTuckerton,NJMagotha,VAElizabethRiver,VAManteo,NCSapeloIsland,GAWeeksBay,ALDauphinIsland,ALLeeville,LAHackberry,LAPortO'Connor,TXPortAransas,TX200kmMACTNJVANCSCGAFLALMSLATXMEGulfofMexicoAtlanticOceanWilliamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page4of14ficationKit(Qiagen,USA).
PCRproductswerefurtherpurifiedwithAMPure(Agencourt).
EmulsionPCRwascarriedoutonPCRproductsasdescribed[57].
AmplificationofthePCRproductonthebeadwascontrolledforbyquantifyingandcalculatingthesizeoftheampliconpoolusingaBioanalyzer2100sothattherewasaminimumof2*106copiesofDNAthatrangedinsizefrom100to700basepairs.
SubsequentproductsweresequencedonaRoche/454LifeSciencesGSFLXSequencerattheUniversityofSouthCarolina'sEnvironmentalGenomicsCoreFacility.
ThePicoTiterplatewassubdividedintoeightregionswithanexpecta-tionof30,000readsperregion[58].
AssemblyofpyrosequencingsequencesandSNPDetectionSequencesweretrimmedoftheirbarcodes.
All626sequenceswithatleastoneambiguousbasewereremovedsincethepresenceofevenasingleambiguousbaseisaneffectiveindicatoroflow-qualitysequence[59].
Becauseshorterthanexpectedreadlengthsalsocorrelatestronglywithincorrectreads[60],anotherthreepercentofthesequences(whoselengthsweresmallerthan100bp)wereremoved.
TheremainingreadswerealignedusingCAP3[61].
Qualityscoreswererescaledtobecom-parabletotheusualPhredScoreusingARACHNE[62].
SNPswerecalledatboththeindividuallevelandpopu-lationlevel.
Attheindividuallevel,SNPswerecalledusingbothaBayesianmethodandalikelihoodratiotest(LRT)method.
FortheBayesianmethod,10-4wasusedasTable1:AdaptersandprimersusedintheamplificationofgenomicDNA.
AdaptersBspEI(5'to3')GACGATGAGTCCTGAGCCTGCTACTCTCAGGACTCGGGCCEcoRI(5'to3')CTAGAGTCCTAGTAGCACCTCGTAGACTGCGTACC*CATCTGACGCATGGTTAAPreselectivePrimersEcoRI(5'to3')CTGAGTCCTAGTAGCACCBspEI(5'to3')GACGATGAGTCCTGAGCSelectivePrimersEcoRI(5'to3')GACTGCGTACCAATTCAAGBspEI(5'to3')GACGATGAGTCCTGAGCCBarcodedPrimersEcoRI(5'to3')1GCCTCCCTCGCGCCATCAGAGCCTAAGCTGACTGCGTACCAATTCAAG2GCCTCCCTCGCGCCATCAGAGTTCAAGTCGACTGCGTACCAATTCAAG3GCCTCCCTCGCGCCATCAGACTTGAACTGGACTGCGTACCAATTCAAG4GCCTCCCTCGCGCCATCAGACGGTAACGTGACTGCGTACCAATTCAAG5GCCTCCCTCGCGCCATCAGATCCGAATCGGACTGCGTACCAATTCAAG6GCCTCCCTCGCGCCATCAGATGGCAATGCGACTGCGTACCAATTCAAG7GCCTCCCTCGCGCCATCAGCAGGTCCAGTGACTGCGTACCAATTCAAG8GCCTCCCTCGCGCCATCAGCATTGCCATGGACTGCGTACCAATTCAAG9GCCTCCCTCGCGCCATCAGCTAAGCCTAGGACTGCGTACCAATTCAAG10GCCTCCCTCGCGCCATCAGCGAATCCGATGACTGCGTACCAATTCAAGBspEI(5'to3')GCCTTGCCAGCCCGCTCAGGACGATGAGTCCTGAGCC*Starindicateslocationof2'O-methylblock.
Table1:AdaptersandprimersusedintheamplificationofgenomicDNA.
(Continued)Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page5of14thepriorforthemutationrate[63].
Atthepopulationlevel,foreachlocusonthecontig,wesimulatedtheerrormodelandmarkedalocusasapotentialSNPifithadalargernumberofsecondallelesincomparisontothecrit-icalvaluefromtheerrormodel.
Furthermore,apotentialSNPsitehadtohaveatleastthreeindividualssequencedto2*atthatlocusunlessanotherpotentialSNPsitewaswithinfivebasepairsorover90%oftheindividualshadbeenclassifiedasheterozygousattheindividuallevel.
Thiswasdonetominimizetherateoffalsepositivescausedbyhomologs.
BayesianandLRTmodelforSNPcallingatindividuallevelFortheBayesianmodel,foreachcontig,Prior=1*10-4representsthemutationrate;Nrepresentsthetotalnum-berofuniquemappinglociwithmultipleallelictypes;Aiandairepresent,respectively,themajorandminorallelesatlocusi;Nirepresentsthetotalnumberofallelesobservedforlocusi,andYjisthetypeofthejthallelecopyamongtheseNialleleswherej=0Ni;finally,ejistheprobabilityoferrorofthejthallelewheretheerrorproba-bilityiscomputedasandwhereQisthecorre-spondingqualityscoreafterrescaling.
Theposteriorprobabilityfortheithlocusbeinghomozygousorheterozygousis:Basedontheposteriorprobabilitiesfromabove,weclassifiedeachoftheseNlociashomozygousorheterozygousexclusively.
Ifalocuswasclassifiedasheterozygous,itwasfurthertestedusingalikelihoodratiotest(LRT)asfollows:Foraparticularlocusionthecontig:whereXjstandsforthetrueallelethatweshouldhaveobserved.
ForeachYj,wehaveanerrorprobabilityofejassociatedwithit.
Thenwehave:Thereforewehave:andBasedonalloftheabove,thelikelihoodoflocusIwascomputedas:1010QPPPPP(.
|)()~(|HeterodatadataHeteroHeterodatadatah=*eeteroHeteroPriorHomodatadataHomNi.
)(.
)~(.
|)(|**=PPP(0.
5)ooHomodatadataoHomoPrior.
)(.
)()~(|hom.
)(.
)~()(***PPPP11eeejYAjNijYajiji)()()11==∏*PP()()XApXapiiji====1PPPP(|)(|)(|)(YAXAeYaXaeYaXAeYiiiijiiiijiiiij=========11iiiiijAXae===|)PPYAXAepepYaXAepeiiiijjiiiijj111**()1pYBernoulliepepjjj111LepepepepjjNjIjjIjjj=111111Figure2Designof454pyrosequencingcontiggeneratedfromthedigestionofgenomicDNAwithrestrictionenzymes(EcoRIandBspEI),theadditionofrestrictionsitespecificlinkers,anindividualbarcodeanda454ampliconadapter.
EcoRIadapterBspEIadapterFadapterFbarcodeRadaptergDNA5'5'3'3'F-sequencingprimerEcoRIselectiveprimer*R-sequencingprimerBspEIselectiveprimerWilliamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page6of14WhereIj=1ifYj=Ai;andIj=0ifYj=aiTheLRTwasperformedwiththehypothesisofHO:p=0.
5versusHa:p>0.
5and-2*LRT~χ2(1).
ErrormodelsimulatingInordertocallSNPsatthepopulationlevel,wesimulatedtheerrormodelforeachlocuswithmultipleallelictypes;weassumedthataparticularlocuswashomozygouswithmajoralleleAiandrandomlysimulatedNinumberofallelescopiestobeAioranyoftheotherthreealleletypesfromauniformdistributionwithprobability(1-ej)andejrespectively.
Werepeatedthisprocess10,000timesandrecordedthedifferentnumbersofsecondallelesfoundinthesimulation.
Thecriticalvaluewaschosenasthenum-berofsecondalleleswitharight-sidep-valueof0.
001.
ValidationofSNPsMultiplexassaystargeting458SNPsin250F.
heteroclitusindividuals,90F.
grandisindividuals,23F.
majalisindi-viduals,and21F.
similisindividualswereattemptedusingtheSequenomMassARRAYtechnology.
Theseconsistedof81putativeSNPsidentifiedbytheF.
heteroclituspyrosequencing,350putativeSNPspreviouslyidentifiedinF.
heteroclitusESTs[64],and27putativeSNPsfrom22genescontaining,amongstothers,SNPsinthearylhydrocarbonreceptor[65],lactatedehydrogenaseB[29],andtheproximalpromoterofcytochromeP4501A(unpublished).
AssaysweredesignedusingtheMassAR-RAYAssayDesignSoftwarewiththegoalofmaximizingmultiplexingof36SNPsperwell(Sequenom,SanDiego,CA,USA).
OnlySNPswhere70basepairswereanno-tatedoneithersideofthepolymorphismwereincludedinthestudy.
Therewere14SNPspreviouslyidentifiedwith454pyrosequencingwherethiscriterionwasnotmet.
IfmultipleSNPswereproximal(95%ofF.
heteroclitusindividuals25961.
4SNPscalledin90%but95%ofallindividuals235.
4PolymorphicSNPscalledin>95%ofallindividuals16338.
6SNPscalledin90%ofallindividualsidentifiedin4544656.
8Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page9of14andF.
majalisandF.
similisclusteredtogetherapartfromotherspecies(Figure5b).
InF.
heteroclitus,AMOVAshowedthatmostofthevariationwasdistributedwithinpopulations(59.
05%),butanotherlargeproportionofvariation(31.
1%)wasdis-tributedamongnorthernandsouthernregions.
Theremaining9.
85%ofvariationwasexplainedbydifferencesamongpopulationswithinregions.
InF.
grandis,mostofthevariationwasdistributedwithinpopulations(82.
4%),andasmallerproportion(17.
6%)ofvariationwasdistrib-utedlongitudinallybetweenpopulationsacrosstheGulfofMexico.
AManteltestshowedsignificantisolationbydistanceamongF.
heteroclituspopulations(p95%ofindividuals)indicatingthatdifferencesinamplificationratebetweenspeciesledtotheloweroverallcallrate.
Inwhitespruce,91%ofSNPsverifiedwiththeIlluminaSNPbeadarrayplatform[71,72]weretrue.
ComparabletoF.
heteroclitus,70%ofSNPsinsprucewerecalledingreaterthan95%ofindivid-uals[52].
Overall,verificationofSNPswaspowerfulinprovidinginformationovermanymarkersandindividu-alsandwasabletoprovidedatatodeterminedifferenceswithinpopulations,betweenpopulationsandbetweenspecies.
Speciesdifferentiationwasdemonstratedusingprinci-plecomponentanalysis(PCA)aswellasSTRUCTUREanalysis.
BothanalysesshowedseparationbetweenF.
het-eroclitus,F.
grandisandF.
majalisandsimilisaswellaspopulationstructurewithinF.
heteroclitus(Figure4).
Theseanalysesprovidedthemostresolution(evenamongdistinguishingpopulations)inF.
heteroclitusbecausetheSNPswereoriginallyidentifiedinthisspecies(i.
e.
,duetoanascertainmentbias).
PCAandSTRUC-TUREdidnotdifferentiatesisterspecies,F.
similisandF.
majalis,fromeachotherorestablishpopulationstruc-turewithinthesespecies.
Smallsamplesizes(1to10indi-vidualsperpopulation),highlevelsofmonomorphism(averageof28%ofallSNPs),andthefactthatonly10%ofSNPallelesdifferedbetweenthesetwospecies,decreasedthepowertodetectsuchdifferenceswhenanalyzedinconjunctionwithF.
heteroclitusandF.
grandis.
PopulationstructurealsowasmaskedinF.
grandiswhendatawasanalyzedwithotherspecies.
However,whenF.
grandisindividualswereanalyzedseparately,theyalsoshoweddistinctpopulationstructure(datanotshown).
Oneotherstudyhasreportedmultiplefixeddifferencesinmito-chondrialsequencesbetweenF.
heteroclitusandF.
gran-dis[33],butnootherstudytodatehasevaluateddifferencesatmanylocibetweenallfourspeciesusedinthisstudy.
WithinF.
heteroclitusandF.
grandisspecies,within-populationfixationindices(FIS,averagedacrossallloci)rangedfrom0.
09to0.
32.
AmongF.
heteroclitus,allpopu-lationshadanoverallsignificantdeficiencyofheterozy-gotesindicatedbypositiveFISvalues.
Inthesepopulations,approximately10%oflocihadsimilarlyverylargeFISvalues(>0.
5)acrosspopulationscausingtheskewintheaverageFISvalueforeachpopulation.
Withinapopulation,theselociwerepredominatelyhomozygousforoneallelewithacompleteabsenceoftheheterozygoteandoneorafewindividualshomozygousforthealterna-tiveallele.
ThelociwhichpresentedthispatternwerecalledconservativelyatbothallelesbySequenomsoft-wareacrossallindividualsindicatingthatgenotypingerrorwasnotthemainreasonforthispattern.
Furthermore,allnorthernpopulationswerepredomi-natelyhomozygousforonealleleandallsouthernpopula-tionswerepredominatelyhomozygousforthealternativealleleindicatingstrongdemographicpatternsinthedata.
ThesamedemographicpatternwasnotfoundinF.
gran-dis.
AmongF.
grandispopulations,most(70%)SNPswithhighFISvaluesweredifferentbetweenpopulations.
ThisisincontrasttoF.
heteroclituspopulationswherelociwithhighFISvaluesweresharedacrosspopulations.
WithinanyoneF.
grandispopulation,oneallelewaspre-dominantasahomozygotewithoneorafewindividualswiththealternativehomozygote.
Themostparsimoniousexplanationisthatthereisundetectedsubstructure.
SNPsinHardy-Weinbergwereshowntobemoderatelypolymorphic(averageof60%)inF.
heteroclitus.
InF.
gran-dis,SNPswereshowntolackpolymorphism(7.
18%).
ThehigherpercentageofmonomorphiclociinF.
grandislikelyisduetoascertainmentbiasinSNPdiscoverycausedbyonlyusingF.
heteroclituspopulations.
Manyofthemonomorphicloci(24%)representfixeddifferencesbetweenF.
heteroclitusandF.
grandis.
Thus,whileSNPmarkersdevelopedinF.
heteroclitusarenotnecessarilypolymorphicinotherFundulusspecies,theystillcanbeusedtodifferentiateF.
heteroclitusfromotherspecies.
AmongF.
heteroclituspopulations,genotypedatarevealedstronglatitudinalclinesbetweentheNorthernandSouthernF.
heteroclituspopulations.
PCA,STRUC-TURE,FSTvalues,andtheisolationbydistancetestiden-tifiedthatindividualsfromNorthernpopulations(above40-41°N)weredistinctfromSouthernpopulations.
Thissplitiscenteredaroundthesouthern-mostextentoftheAtlanticcoastaladvancementduringthelatePleistocene[73].
Specifically,observedheterozygosityandallelicrichnessacrossalllociissignificantlylower(p=0.
043,p=0.
042,respectively)inthenorththaninthesouth.
Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page12of14Thesedifferenceshavebeenshownpreviouslyinmor-phologicalfeatures[74]numerousallozymeloci[34-36,75]andmicrosatellites[28].
Thelargerhistoricalpop-ulationsizeofF.
heteroclitusinthesouth[28]wouldmaintaingreaterheterozygosityandallelicrichnessatsharedloci;inthenorth,wherepopulationsizesaresmaller,locihaveahigherprobabilityofbecomingfixed.
FourSTRUCTUREclustersencompassthesixnorth-ernpopulationswhileonlytwoclustersencompassthefivesouthernpopulations(Figure5A).
Separatenorthernclustersmaybedrivenbysmallerpopulationsizesinwhichdriftisgreater.
Whengeneticdrifthasalargereffectitbecomeseasiertodistinguishpopulationsbecausetheaveragedifferenceinallelefrequenciesofamarkerindifferentpopulationswillbegreater.
ThisisillustratedbyalargeraverageFSTof0.
20amongnorthernpopulationsincomparisontothatofanaverageFSTof0.
10amongsouthernpopulations.
Thisstatisticisalsoevidentforthenorthandsouthsplit,wherepopulationsfromrespectiveregionshadanextremelyhighFSTvalueof0.
44whencomparedagainstoneanother.
SimilargeneticdivergencehasbeenreportedforF.
heteroclitususingmicrosatellites(0.
196amongnorthernpopulations,0.
117amongsouthernpopulationsand0.
330forthetwomostdivergentpopulations,NovaScotiaandGeorgia[28]).
Similardemographicpatternshavebeendescribedinfreshwaterfish[76]andmarinespeciessuchasgoby[77]andbluecrab[78],and,asinFundulus,thesepat-ternsareattributedtoPleistoceneevents.
AsimilarlatitudinalclineoccursbetweenpopulationsofF.
grandis,andaManteltestshowssignificantisolationbydistance.
However,therewerenosignificantdiffer-encesbetweeneitherlevelsofpolymorphismorobservedheterozygosityalonglatitudeorlongitude.
Williamsetal.
,2008reportedsignificantisolationbydistanceaswellasdecreasedallelicrichnesswithincreasinglatitude.
Inthis2008study,microsatelliteswereused,andtwoadditionalsitessoutherntothoseusedinourstudywereincluded.
SincemicrosatelliteshavemanymoreallelesthanSNPsandtwoadditionalsiteswerefoundtohaverelativelyhigherallelicrichnessincomparisontoallothersamplingsitesalongthegulf,thismayaccountforthedifferencesfoundinlevelsofpolymorphism.
ConclusionsBytargetingSNPscontainedinbothcodingandnon-codingareasofthegenome,weareabletobetterunder-standhowevolutionaryforcesareshapingtheFundulusgenome.
SimilarstudiesusinghighthroughputmethodstosequenceSNPmarkershavebeendevelopedinAtlan-ticcod[51],whitespruce[52],Eucalyptus[8],andswine[70].
Likeourstudy,thesestudiesexpandedtheirownspecies'knowledgebasewithrespecttopotentialmarkersforstudyingevolutionaryadaptation(inthecaseofcodandspruce),genome-wideassessmentofdiversity(Euca-lyptus)orforuseinbreedingprograms(swine)AdditionalmaterialAuthors'contributionsLMWdesignedexperiments,carriedoutlaboratoryandstatisticalanalyses,anddraftedthemanuscript.
XMcarriedoutSNPdetectionof454data.
ARBcarriedoutSNPdetectionof454data,assistedinstatisticalanalysesofMassARRAYdata,andprovidedcommentsonearlierversionsofthismanuscript.
CDBassistedindesignedexperimentsandhelpdevelopSNPdetectionsoftware.
MFOdesignedexperiments,assistedonstatisticalanalysis,andhelpedtodraftthemanuscript.
AllauthorsreadandapprovedthefinalmanuscriptAcknowledgementsTheauthorsthankG.
BozinovicandM.
EverettforassistanceinthecollectionofsamplesandD.
Crawfordforvaluableinputintomethodology.
PartofthisworkwascarriedoutbyusingtheresourcesoftheComputationalBiologyServiceUnitfromCornellUniversitywhichispartiallyfundedbyMicrosoftCorporation.
FundingwaspartiallyprovidedbyNIEHSTrainingGrantES525163awardfromtheDepartmentofEnvironmentalandMolecularToxicologyatNorthCarolinaStateUniversitytoLMW,NIHRO1ES011588toMFO,NSFDEB0948510toARB,andNIHR01HG003229CDB.
AuthorDetails1DepartmentofEnvironmentalandMolecularToxicology,Box7633,NorthCarolinaStateUniversity,Raleigh,NC27695-7633,USA,2DepartmentofBiologicalStatisticsandComputationalBiology,CornellUniversity,Ithaca,NY14853,USAand3RosenstielSchoolofMarineandAtmosphericSciences,UniversityofMiami,4600RickenbackerCauseway,Miami,FL33149,USAReferences1.
BordoniR,BonnalR,RizziE,CarreraP,BenedettiS,CremonesiL,StenirriS,ColomboA,MontrasioC,BonalumiS,etal.
:EvaluationofhumangenevariantdetectioninampliconpoolsbytheGS-FLXparallelPyrosequencer.
BmcGenomics2008,9:464.
2.
GarberM,ZodyMC,ArachchiHM,BerlinA,GnerreS,GreenLM,LennonN,NusbaumC:Closinggapsinthehumangenomeusingsequencingbysynthesis.
GenomeBiology2009,10:R60.
3.
IngmanM,GyllenstenU:SNPfrequencyestimationusingmassivelyparallelsequencingofpooledDNA.
EuropeanJournalofHumanGenetics2009,17:383-386.
4.
TurnerEH,LeeCL,NgSB,NickersonDA,ShendureJ:Massivelyparallelexoncaptureandlibrary-freeresequencingacross16genomes.
NatureMethods2009,6:315-316.
5.
ZhengJB,MoorheadM,WengL,SiddiquiF,CarltonVEH,IrelandJS,LeeL,PetersonJ,WilkinsJ,LinS,etal.
:High-throughput,high-accuracyarray-basedresequencing.
ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica2009,106:6712-6717.
6.
DeSchutterK,LinYC,TielsP,VanHeckeA,GlinkaS,Weber-LehmannJ,RouzeP,dePeerYV,CallewaertN:GenomesequenceoftherecombinantproteinproductionhostPichiapastoris.
NatureBiotechnology2009,27:561-U104.
7.
IaconoM,VillaL,FortiniD,BordoniR,ImperiF,BonnalRJP,Sicheritz-PontenT,DeBellisG,ViscaP,CassoneA,CarattoliA:Whole-genomepyrosequencingofanepidemicmultidrug-resistantAcinetobacterbaumanniistrainbelongingtotheEuropeancloneIIgroup.
AntimicrobialAgentsandChemotherapy2008,52:2616-2625.
8.
NovaesE,DrostDR,FarmerieWG,PappasGJ,GrattapagliaD,SederoffRR,KirstM:High-throughputgeneandSNPdiscoveryinEucalyptusgrandis,anuncharacterizedgenome.
BmcGenomics2008,9:.
Additionalfile1SNPminorallelefrequencies.
DistributionsofSNPminorallelefrequencies(MAF)withinF.
heteroclitusandF.
grandispopula-tions.
Received:5February2010Accepted:30April2010Published:30April2010Thisarticleisavailablefrom:http://www.
biomedcentral.
com/1471-2156/11/322010Williamsetal;licenseeBioMedCentralLtd.
ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(http://creativecommons.
org/licenses/by/2.
0),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
BMCGenetics2010,11:32Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page13of149.
VeraJC,WheatCW,FescemyerHW,FrilanderMJ,CrawfordDL,HanskiI,MardenJH:Rapidtranscriptomecharacterizationforanonmodelorganismusing454pyrosequencing.
MolecularEcology2008,17:1636-1647.
10.
BontellIL,HallN,AshelfordKE,DubeyJP,BoyleJP,LindhJ,SmithJE:WholegenomesequencingofanaturalrecombinantToxoplasmagondiistrainrevealschromosomesortingandlocalallelicvariants.
GenomeBiology2009,10:R53.
11.
WordenAZ,Panaud,Piegu:GreenevolutionanddynamicadaptationsrevealedbygenomesofthemarinepicoeukaryotesMicromonas(vol324,pg268,2009).
Science2009,325:147-147.
12.
LuikartG,EnglandPR,TallmonD,JordanS,TaberletP:Thepowerandpromiseofpopulationgenomics:Fromgenotypingtogenometyping.
NatureReviewsGenetics2003,4:981-994.
13.
DiazRJ:Overviewofhypoxiaaroundtheworld.
JournalofEnvironmentalQuality2001,30:275-281.
14.
DiazRJ,RosenbergR:Marinebenthichypoxia:Areviewofitsecologicaleffectsandthebehaviouralresponsesofbenthicmacrofauna.
OceanographyandMarineBiology-anAnnualReview1995,33:245-303.
15.
SmithKJ,AbleKW:Dissolvedoxygendynamicsinsaltmarshpoolsanditspotentialimpactsonfishassemblages.
MarineEcology-ProgressSeries2003,258:223-232.
16.
WirginI,WaldmanJR:ResistancetocontaminantsinNorthAmericanfishpopulations.
MutationResearch-FundamentalandMolecularMechanismsofMutagenesis2004,552:73-100.
17.
OleksiakMF,ChurchillGA,CrawfordDL:Variationingeneexpressionwithinandamongnaturalpopulations.
NatureGenetics2002,32:261-266.
18.
OleksiakMF,RoachJL,CrawfordDL:NaturalvariationincardiacmetabolismandgeneexpressioninFundulusheteroclitus.
NatureGenetics2005,37:67-72.
19.
PetersonJSK,BainLJ:Differentialgeneexpressioninanthracene-exposedmummichogs(Fundulusheteroclitus).
AquaticToxicology2004,66:345-355.
20.
MeyerJN,VolzDC,FreedmanJH,DiGiulioRT:DifferentialdisplayofhepaticmRNAfromkillifish(Fundulusheteroclitus)inhabitingaSuperfundestuary.
AquaticToxicology2005,73:327-341.
21.
WhiteheadA,CrawfordDL:Variationintissue-specificgeneexpressionamongnaturalpopulations.
GenomeBiology2005,6:R13.
22.
GonzalezHO,RolingJA,BaldwinWS,BainLJ:Physiologicalchangesanddifferentialgeneexpressioninmummichogs(Fundulusheteroclitus)exposedtoarsenic.
AquaticToxicology2006,77:43-52.
23.
RolingJA,BainLJ,Gardea-TorresdeyJ,BaderJ,BaldwinWS:Hexavalentchromiumreduceslarvalgrowthandaltersgeneexpressioninmummichog(Fundulusheteroclitus).
EnvironmentalToxicologyandChemistry2006,25:2725-2733.
24.
WhiteheadA,CrawfordDL:Neutralandadaptivevariationingeneexpression.
ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica2006,103:5425-5430.
25.
FisherMA,OleksiakMF:Convergenceanddivergenceingeneexpressionamongnaturalpopulationsexposedtopollution.
BmcGenomics2007,8:108.
26.
OleksiakMF:Changesingeneexpressionduetochronicexposuretoenvironmentalpollutants.
AquaticToxicology2008,90:161-171.
27.
PaschallJE,OleksiakMF,VanWyeJD,RoachJL,WhiteheadJA,WyckoffGJ,KolellKJ,CrawfordDL:FunnyBase:asystemslevelfunctionalannotationofFundulusESTsfortheanalysisofgeneexpression.
BmcGenomics2004,5:96.
28.
AdamsSM,LindmeierJB,DuvernellDD:Microsatelliteanalysisofthephylogeography,Pleistocenehistoryandsecondarycontacthypothesesforthekillifish,Fundulusheteroclitus.
MolecularEcology2006,15:1109-1123.
29.
BernardiG,SordinoP,PowersDA:Concordantmitochondrialandnuclear-DNAphylogeniesforpopulationsoftheteloestfishFundulusheteroclitus.
ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica1993,90:9271-9274.
30.
CrawfordDL,PowersDA:EvolutionaryAdaptationtoDifferentThermalEnvironmentsViaTranscriptionalRegulation.
MolecularBiologyandEvolution1992,9:806-813.
31.
McMillanAM,BagleyMJ,JacksonSA,NacciDE:Geneticdiversityandstructureofanestuarinefish(Fundulusheteroclitus)indigenoustositesassociatedwithahighlycontaminatedurbanharbor.
Ecotoxicology2006,15:539-548.
32.
WilliamsLM,OleksiakMF:Signaturesofselectioninnaturalpopulationsadaptedtochronicpollution.
BmcEvolutionaryBiology2008,8:282.
33.
WhiteheadA:Comparativemitochondrialgenomicswithinandamongspeciesofkillifish.
BmcEvolutionaryBiology2009,9:11.
34.
PowersDA,PlaceAR:BiochemicalgeneticsofFundulusheteroclitus(L).
Temporalandspatialvariationingene-frequenciesofLDH-B,MDH-A,GPI-B,andPGM-A.
BiochemicalGenetics1978,16:593-607.
35.
PowersDA,RopsonI,BrownDC,VanbenedenR,CashonR,GonzalezvillasenorLI,DimicheleJA:GeneticvariationinFundulusheteroclitus-geographicdistribution.
AmericanZoologist1986,26:131-144.
36.
CashonRE,VanbenedenRJ,PowersDA:BiochemicalgeneticsofFundulusheteroclitus(L).
Spatialvariationingene-frequenciesofIDH-A,IDH-B,6-PGDH-A,andEST-S.
BiochemicalGenetics1981,19:715-728.
37.
CrawfordDL,PowersDA:Molecular-BasisofEvolutionaryAdaptationattheLactateDehydrogenase-BLocusintheFishFundulus-Heteroclitus.
ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica1989,86:9365-9369.
38.
CrawfordDL,ConstantinoHR,PowersDA:LactateDehydrogenase-BCdnafromtheTeleostFundulus-Heteroclitus-EvolutionaryImplications.
MolecularBiologyandEvolution1989,6:369-383.
39.
GillP:Anassessmentoftheutilityofsinglenucleotidepolymorphisms(SNPs)forforensicpurposes.
InternationalJournalofLegalMedicine2001,114:204-210.
40.
WeirBS,CardonLR,AndersonAD,NielsenDM,HillWG:Measuresofhumanpopulationstructureshowheterogeneityamonggenomicregions.
GenomeResearch2005,15:1468-1476.
41.
PaschouP,ZivE,BurchardEG,ChoudhryS,Rodriguez-CintronW,MahoneyMW,DrineasP:PCA-correlatedSNPsforstructureidentificationinworldwidehumanpopulations.
PlosGenetics2007,3:1672-1686.
42.
Yamaguchi-KabataY,NakazonoK,TakahashiA,SaitoS,HosonoN,KuboM,NakamuraY,KamataniN:JapanesePopulationStructure,BasedonSNPGenotypesfrom7003IndividualsComparedtoOtherEthnicGroups:EffectsonPopulation-BasedAssociationStudies.
AmericanJournalofHumanGenetics2008,83:445-456.
43.
PrimmerCR,BorgeT,LindellJ,SaetreGP:Single-nucleotidepolymorphismcharacterizationinspecieswithlimitedavailablesequenceinformation:highnucleotidediversityrevealedintheaviangenome.
MolecularEcology2002,11:603-612.
44.
KongFR,TongZS,ChenXY,SorrellT,WangB,WuQX,EllisD,ChenS:RapididentificationanddifferentiationofTtichophytonspecies,basedonsequencePolymorphismsoftheribosomalinternaltranscribedspleacerregions,byrollingcircleamplification.
JournalofClinicalMicrobiology2008,46:1192-1199.
45.
JohnsonN,FletcherO,PallesC,RuddM,WebbE,SellickG,SilvaIDS,McCormackV,GibsonL,FraserA,etal.
:CountingpotentiallyfunctionalvariantsinBRCA1,BRCA2andATMpredictsbreastcancersusceptibility.
HumanMolecularGenetics2007,16:1051-1057.
46.
TomlinsonI,WebbE,Carvajal-CarmonaL,BroderickP,KempZ,SpainS,PenegarS,ChandlerI,GormanM,WoodW,etal.
:Agenome-wideassociationscanoftagSNPsidentifiesasusceptibilityvariantforcolorectalcancerat8q24.
21.
NatureGenetics2007,39:984-988.
47.
AmosCI,WuXF,BroderickP,GorlovIP,GuJ,EisenT,DongQ,ZhangQ,GuXJ,VijayakrishnanJ,etal.
:Genome-wideassociationscanoftagSNPsidentifiesasusceptibilitylocusforlungcancerat15q25.
1.
NatureGenetics2008,40:616-622.
48.
PoehlmannA,KuesterD,MeyerF,LippertH,RoessnerA,Schneider-StockR:K-rasmutationdetectionincolorectalcancerusingthePyrosequencingtechnique.
PathologyResearchandPractice2007,203:489-497.
49.
MauricioR,StahlEA,KorvesT,TianDC,KreitmanM,BergelsonJ:NaturalselectionforpolymorphisminthediseaseresistancegeneRps2ofArabidopsisthaliana.
Genetics2003,163:735-746.
50.
HoekstraHE,HirschmannRJ,BundeyRA,InselPA,CrosslandJP:Asingleaminoacidmutationcontributestoadaptivebeachmousecolorpattern.
Science2006,313:101-104.
51.
MoenT,HayesB,NilsenF,DelghandiM,FjalestadKT,FevoldenSE,BergPR,LienS:IdentificationandcharacterisationofnovelSNPmarkersinAtlanticcod:Evidencefordirectionalselection.
BmcGenetics2008,9:18.
Williamsetal.
BMCGenetics2010,11:32http://www.
biomedcentral.
com/1471-2156/11/32Page14of1452.
NamroudMC,BeaulieuJ,JugeN,LarocheJ,BousquetJ:Scanningthegenomeforgenesinglenucleotidepolymorphismsinvolvedinadaptivepopulationdifferentiationinwhitespruce.
MolecularEcology2008,17:3599-3613.
53.
VignalA,MilanD,SanCristobalM,EggenA:AreviewonSNPandothertypesofmolecularmarkersandtheiruseinanimalgenetics.
GeneticsSelectionEvolution2002,34:275-305.
54.
WirginII,DamoreM,GrunwaldC,GoldmanA,GarteSJ:GeneticDiversityatanOncogeneLocusandinMitochondrial-DNAbetweenPopulationsofCancer-ProneAtlanticTomcod.
BiochemicalGenetics1990,28:459-475.
55.
AljanabiSM,MartinezI:Universalandrapidsalt-extractionofhighqualitygenomicDNAforPCR-basedtechniques.
NucleicAcidsResearch1997,25:4692-4693.
56.
ParameswaranP,JaliliR,TaoL,ShokrallaS,GharizadehB,RonaghiM,FireAZ:Apyrosequencing-tailorednucleotidebarcodedesignunveilsopportunitiesforlargescalesamplemultiplexing.
NucleicAcidsResearch2007,35:e130.
57.
MarguliesM,EgholmM,AltmanWE,AttiyaS,BaderJS,BembenLA,BerkaJ,BravermanMS,ChenYJ,ChenZT,etal.
:Genomesequencinginmicrofabricatedhighdensitypicolitrereactors.
Nature2005,437:376-380.
58.
MeyerM,StenzelU,HofreiterM:Paralleltaggedsequencingonthe454platform.
NatureProtocols2008,3:267-278.
59.
HuseSM,HuberJA,MorrisonHG,SoginML,WelchDM:AccuracyandqualityofmassivelyparallelDNApyrosequencing.
GenomeBiology2007,8:R143.
60.
BrockmanW,AlvarezP,YoungS,GarberM,GiannoukosG,LeeWL,RussC,LanderES,NusbaumC,JaffeDB:QualityscoresandSNPdetectioninsequencing-by-synthesissystems.
GenomeResearch2008,18:763-770.
61.
HuangXQ,MadanA:CAP3:ADNAsequenceassemblyprogram.
GenomeResearch1999,9:868-877.
62.
BatzoglouS,JaffeDB,StanleyK,ButlerJ,GnerreS,MauceliE,BergerB,MesirovJP,LanderES:ARACHNE:Awhole-genomeshotgunassembler.
GenomeResearch2002,12:177-189.
63.
DuvernellDD,LindmeierJB,FaustKE,WhiteheadA:RelativeinfluencesofhistoricalandcontemporaryforcesshapingthedistributionofgeneticvariationintheAtlantickillifish,Fundulusheteroclitus.
MolecularEcology2008,17:1344-1360.
64.
QuackenbushJ,LiangF,HoltI,PerteaG,UptonJ:TheTIGRGeneIndices:reconstructionandrepresentationofexpressedgenesequences.
NucleicAcidsResearch2000,28:141-145.
65.
HahnME,KarchnerSI,FranksDG,MersonRR:ArylhydrocarbonreceptorpolymorphismsanddioxinresistanceinAtlantickillifish(Fundulusheteroclitus).
Pharmacogenetics2004,14:131-143.
66.
ExcoffierL,LavalG,SchneiderS:Arlequin(version3.
0):Anintegratedsoftwarepackageforpopulationgeneticsdataanalysis.
EvolutionaryBioinformatics2005:47-50.
67.
PritchardJK,StephensM,DonnellyP:Inferenceofpopulationstructureusingmultilocusgenotypedata.
Genetics2000,155:945-959.
68.
FalushD,StephensM,PritchardJK:Inferenceofpopulationstructureusingmultilocusgenotypedata:Linkedlociandcorrelatedallelefrequencies.
Genetics2003,164:1567-1587.
69.
RosenbergNA:DISTRUCT:aprogramforthegraphicaldisplayofpopulationstructure.
MolecularEcologyNotes2004,4:137-138.
70.
WiedmannRT,SmithTPL,NonnemanDJ:SNPdiscoveryinswinebyreducedrepresentationandhighthroughputpyrosequencing.
BmcGenetics2008,9:81.
71.
ShenR,FanJB,CampbellD,ChangWH,ChenJ,DoucetD,YeakleyJ,BibikovaM,GarciaEW,McBrideC,etal.
:High-throughputSNPgenotypingonuniversalbeadarrays.
MutationResearch-FundamentalandMolecularMechanismsofMutagenesis2005,573:70-82.
72.
FanJ-B,OliphantA,ShenR,KermaniBG,GarciaF,GundersonKL,HansenM,SteemersF,ButlerSL,DeloukasP,etal.
:HighlyParallelSNPGenotyping.
ColdSpringHarborSymposiaonQuantitativeBiology2003,68:69-78.
73.
MickelsonDMCL,FullertonDS,BornsHW,(Ed):Late-QuaternaryEnvironmentsoftheUnitedStates.
Minneapolis:UniversityofMinnesotaPress;1983.
74.
AbleKW,FelleyJD:GeographicalvariationinFundulusheteroclitus-Testsforconcordancebetweeneggandadultmorphologies.
AmericanZoologist1986,26:145-157.
75.
RopsonIJ,BrownDC,PowersDA:BiochemicalgeneticsofFundulusheteroclitus(L.
)6.
Geographicalvariationinthegenefrequenciesof15loci.
NationalGeographicResearch1990,44:16-26.
76.
BernatchezL,WilsonCC:Comparativephylogeographyofnearcticandpalearcticfishes.
MolecularEcology1998,7:431-452.
77.
GyselsES,HellemansB,PampoulieC,VolckaertFAM:Phylogeographyofthecommongoby,Pomatoschistusmicrops,withparticularemphasisonthecolonizationoftheMediterraneanandtheNorthSea.
MolecularEcology2004,13:403-417.
78.
McMillen-JacksonAL,BertTM:MitochondrialDNAvariationandpopulationgeneticstructureofthebluecrabCallinectessapidusintheeasternUnitedStates.
MarineBiology2004,145:769-777.
doi:10.
1186/1471-2156-11-32Citethisarticleas:Williamsetal.
,SNPidentification,verification,andutilityforpopulationgeneticsinanon-modelgenusBMCGenetics2010,11:32