A high-density linksage map for Astyanax mexicanus using genotyping-by-sequencing technology


previous item | next item

Citation
A high-density linksage map for Astyanax mexicanus using genotyping-by-sequencing technology

Material Information

Title:
A high-density linksage map for Astyanax mexicanus using genotyping-by-sequencing technology
Series Title:
Genes Genomes Genet
Creator:
Carlson, Brian M.
Onusko, Samuel W.
Gross, Joshua B.
Publisher:
G3: Genes, Genomes, Genetics
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Next Generation Sequencing ( local )
Qtl Analysis ( local )
Blind Mexican Cave Tetra ( local )
Regressive Phenotypic Evolution ( local )
Genre:
serial ( sobekcm )

Notes

Abstract:
The Mexican tetra, Astyanax mexicanus, is a unique model system consisting of cave-adapted and surface-dwelling morphotypes that diverged >1 million years (My) ago. This remarkable natural experiment has enabled powerful genetic analyses of cave adaptation. Here, we describe the application of next-generation sequencing technology to the creation of a high-density linksage map. Our map comprises more than 2200 markers populating 25 linksage groups constructed from genotypic data generated from a single genotyping-by-sequencing project. We leveraged emergent genomic and transcriptomic resources to anchor hundreds of anonymous Astyanax markers to the genome of the zebrafish (Danio rerio), the most closely related model organism to our study species. This facilitated the identification of 784 distinct connections between our linksage map and the Danio rerio genome, highlighting several regions of conserved genomic architecture between the two species despite ∼150 My of divergence. Using a Mendelian cave-associated trait as a proof-of-principle, we successfully recovered the genomic position of the albinism locus near the gene Oca2. Further, our map successfully informed the positions of unplaced Astyanax genomic scaffolds within particular linksage groups. This ability to identify the relative location, orientation, and linear order of unaligned genomic scaffolds will facilitate ongoing efforts to improve on the current early draft and assemble future versions of the Astyanax physical genome. Moreover, this improved linksage map will enable higher-resolution genetic analyses and catalyze the discovery of the genetic basis for cave-associated phenotypes.
Original Version:
Genes Genomes Genet, Vol. 5, no. 2 (2015-02-06).

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
Unknown
Resource Identifier:
K26-05220 ( USFLDC: LOCAL DOI )
k26.5220 ( USFLDC: LOCAL Handle )

USFLDC Membership

Aggregations:
University of South Florida
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

INVESTIGATIONAHigh-DensityLinkageMapfor Astyanax mexicanus UsingGenotyping-by-Sequencing TechnologyBrianM.Carlson,SamuelW.Onusko,andJoshuaB.Gross1DepartmentofBiologicalSciences,UniversityofCincinnati,Cincinnati,Ohio45221ABSTRACT TheMexicantetra, Astyanaxmexicanus ,isauniquemodelsystemconsistingofcave-adapted andsurface-dwellingmorphotypesthatdiverged . 1millionyears(My)ago.Thisremarkablenaturalexperimenthasenabledpowerfulgeneticanalysesofcaveadaptation.Here,wedescribetheapplicationofnextgenerationsequencingtechnologytothecreationofahigh-densitylinkagemap.Ourmapcomprisesmore than2200markerspopulating25linkagegroupsconstructedfromgenotypicdatageneratedfromasingle genotyping-by-sequencingproject.Weleveragedemergentgenomicandtranscriptomicresourcesto anchorhundredsofanonymous Astyanax markerstothegenomeofthezebra sh( Daniorerio ),themost closelyrelatedmodelorganismtoourstudyspecies.Thisfacilitatedtheidenti cationof784distinct connectionsbetweenourlinkagemapandthe Daniorerio genome,highlightingseveralregionsofconservedgenomicarchitecturebetweenthetwospeciesdespite 150Myofdivergence.UsingaMendelian cave-associatedtraitasaproof-of-principle,wesuccessfullyrecoveredthegenomicpositionofthealbinism locusnearthegene Oca2 .Further,ourmapsuccessfullyinformedthepositionsofunplaced Astyanax genomicscaffoldswithinparticularlinkagegroups.Thisabilitytoidentifytherelativelocation,orientation, andlinearorderofunalignedgenomicscaffoldswillfacilitateongoingeffortstoimproveonthecurrent earlydraftandassemblefutureversionsofthe Astyanax physicalgenome.Moreover,thisimprovedlinkage mapwillenablehigher-resolutiongeneticanalysesandcatalyzethediscoveryofthegeneticbasisforcaveassociatedphenotypes.KEYWORDSnext-generation sequencing QTLanalysis blindMexican cavetetra regressive phenotypic evolution TheblindMexicancavetetraisapowerfulsystemforunderstanding theevolutionarymechanismsgoverningregressivephenotypes.These animalswerediscoveredin1936andinitiallywereassignedtoanew genus — Anoptichthys ( “ bony shwithouteyes ” )(HubbsandInnes 1936).Breedingstudiesinthe1940sledtothediscoveryofviable hybridoffspringresultingfromcrossesbetweenthe(derived)blind cave-dwellingformsand(ancestral)surface-dwellingformsfromthe samegeographicalregionofnortheastMexico(Breder1943a,b).Both morphotypesarenowregardedasmembersofthesame(oraclosely related)species, Astyanaxmexicanus .Thissystemhasspurredmore thanhalfacenturyofcomparativeresearch(S x ado glu1956)focusing onunresolvedproblemsinevolution(Jeffery2001),development (Pottin etal. 2011),genetics(Schemmel1980),physiology(Salin etal. 2010),andbehavior(Burchards etal. 1985). Classicalandquantitativegeneticapproacheshaveprovidedclear evidencethatmanytroglomorphic(cave-associated)phenotypes evolvedthroughheritablegeneticchanges.Thesestudiescentered onbothMendelianandcomplexphenotypes,includingeyeregression (Yamamoto etal. 2004;Protas etal. 2007;Yoshizawa etal. 2012; O ’ Quin etal. 2013),feeding-relatedbehaviors(Schemmel1980;Yoshizawa etal. 2012),sleeploss(Duboué etal. 2011),schoolingbehavior (Kowalko etal. 2013),pigmentationloss(reviewedinJeffery2009), andintraspeci caggression(Elipot etal. 2013).QTLstudieshave identi edcandidategenesmediatingavarietyofthesetraits,suchas retinaldegeneration(O ’ Quin etal. 2013),ribnumber,eyesize(Gross etal. 2008),albinism( Oca2 )(Protas etal. 2006),andthe brown phenotype( Mc1r )(Gross etal. 2009). Copyright©2015Carlson etal. doi:10.1534/g3.114.015438 ManuscriptreceivedOctober30,2014;acceptedforpublicationDecember11, 2014;publishedEarlyOnlineDecember17,2014. Thisisanopen-accessarticledistributedunderthetermsoftheCreative CommonsAttributionUnportedLicense( http://creativecommons.org/licenses/ by/3.0/ ),whichpermitsunrestricteduse,distribution,andreproductioninany medium,providedtheoriginalworkisproperlycited. Supportinginformationisavailableonlineat http://www.g3journal.org/lookup/ suppl/doi:10.1534/g3.114.015438/-/DC11Correspondingauthor:DepartmentofBiologicalSciences,312CliftonCourt,711B RieveschlHall,Cincinnati,OH45221.E-mail:grossja@ucmail.uc.edu Volume5|February2015|241

PAGE 2

Genomicresourcesforthismodelsystem,however,havehistoricallybeenlimited.The rstlinkagemapwascalculatedbasedon recombinationfrequenciesofanexperimentalF1· Pachóncavebackcrosspedigreeusingmarkersgeneratedfromrandomampli edpolymorphicDNA(RAPD) ngerprinting(BorowskyandWilkens2002). Thismapwassupplantedbyahigher-resolutionmapusingmore individualsandmarkerscomposedofpolymorphicmicrosatellites identi edusingCANdinucleotiderepeats(Protas etal. 2006).Using thissecond-generationlinkagemap,Protas etal. (2008)discovered ageneticbasisforseveralcave-associatedphenotypicchangesincludingpigmentationregression,reducedribnumbers,slowerweight loss,andincreasedchemicalsensitivity.Earlycomparativegenomicanalysesutilizingthismap rstdemonstratedextensivesyntenyconservedbetween Astyanax and Daniorerio ,despite 150 Myofdivergence(Gross etal. 2008).The rstnext-generation sequencing(NGS)-basedlinkagemapusingrestriction-associated DNAsequencing(RAD-seq)technologywaspublishedbyO ’ Quin etal. (2013).Thismap,comprising698markerson25linkage groups,strengthenedtheevidenceforvastregionsofsyntenybetweenthegenomesof Astyanax andzebra shandidenti edseveral criticallociassociatedwithretinaldegeneration(O ’ Quin etal. 2013). Here,wepresentthemostdense,comprehensivelinkagemapto dateusinggenotyping-by-sequencing(GBS)technology.Thistechnologyenablesaccurateandhigh-throughputcollectionofmassive amountsofsequencedata(Davey etal. 2011),includingthousandsof single-nucleotidepolymorphisms(SNPs)segregatingbetweencavedwellingandsurface-dwellingmorphs.GBSutilizesdeepIllumina sequencingofrestrictionenzyme-nickedgenomicDNAlibrariesthat areuniquelybarcodedforeachmemberofanexperimentalpedigree. Thistechniqueisoptimizedtoavoidinclusionofrepetitiveportionsof thegenomeandisextremelyspeci candhighlyreproducible(Elshire etal. 2011).Fisharewell-representedamongstudiesusingGBSand otherRAD-seq – basedmethodologies(Rowe etal. 2011).However, amajorityofGBSstudiesin shhavefocusedonspeciesofcommercial(Everett etal. 2012;Houston etal. 2012;Li etal. 2014)orconservationalconcern(Hecht etal. 2013;Ogden etal. 2013;Hess etal. 2014;Larson etal. 2014).Here,weadaptedthistechnologytoconstructahigh-densitylinkagemapforevolutionaryanddevelopmental studiesinouremergingmodelsystem.Theresultinglinkagemapwill enablehigher-resolutiongenomicstudiesandinformtheassignment ofchromosomalbuildsfortheongoing Astyanax genomesequencing project(McGaugh etal. 2014). MATERIALSANDMETHODS Pedigree,husbandry,andgenomicDNAisolation LinkagemappingandQTLstudieswereperformedusinggenotypic andphenotypicdataobtainedfromtwoseparateF2hybridmapping populations(n=129;n=41)bredfromamalesurface shand femalecave shfromthePachóncave.Inaddition,surface(n=4), Pachóncave(n=4),andsurface · PachónF1hybrid(n=4)specimenswereusedtoevaluateandcodeGBSmarkersforusewithJoinMapsoftware(v.4.1;Kyazma;seebelow),butwerenotincludedin linkagemappingcalculations.ParentalspecimensbelongedtolaboratorypopulationsoriginallysourcedfromtheElAbraregionofnortheasternMexicoandall shusedweregenerouslyprovidedtoour laboratorybyDr.RichardBorowsky(NewYorkUniversity).Alllive sh usedinthisstudyweremaintainedas previouslydescribed(seeGross etal. 2013).Everyindividualfromthe “ Asty66 ” F2population(n=129) wasindividuallyrearedina1-litertank.Allphenotypicdatafromthe “ Asty12 ” F2population(n=41)wereobtainedfromparaformaldehyde-preservedspecimens. Genotyping-by-sequencing GenomicDNAwasextractedfromcaudaltail ntissueoflivesurface, cave,andF1andF2hybrid Astyanaxmexicanus specimensusingthe DNeasyBloodandTissueKit(Qiagen)aspreviouslydescribed(Gross etal. 2013).Twentygenomicsamplesweredigestedwith Eco RI,subjectedtogelelectrophoresisandimagedtoverifythatsamplequality, concentration,andrestrictionfragmentsizedistributionsweresuitable foruseindownstreamanalyses.DNAsampleswerethenpipettedinto individualwellsof96-wellplatesanddilutedtoa nalvolumeof30 m l (100ng/ m l).SampleswereprocessedbytheInstituteforGenomic Diversity(CornellUniversity),wheregenomiclibrarieswereconstructedandGBSwasperformedasdescribedelsewhere(Elshire etal. 2011;Lu etal. 2013). GBSmarkerselection Genotypesforeachof7956GBSmarkers(eachconsistingofasingle SNPina64-bp-longsequencefragment)werescreenedincaveand surface(parental)formstoassignthemorphotypicoriginofeach allele.F1individualswerethenevaluatedtocon rmheterozygosityat eachlocus.Themorphotypicoriginofeachallelewasassignedby consensus — ifthreeormore(outoffour)surfaceorcaveindividuals hadtheidenticalnucleotideataparticularlocus,thenthegenotype wasassignedtotheconsensusparentalpopulation.Likewise,atrue “ hybrid ” genotypewasassignedifthreeormoreF1individualsharboredthesameheterozygouscondition( e.g. ,M,R,S,W,Y,KSNP code)atagivenlocus.Thosegenotypeswithanambiguousmorphotypicoriginweredenoted “ NA. ” Markerswerethenscreenedforsuitabilityinlinkagecalculations. Markersweredeemedunsuitableanddiscardedfromfurtheranalysis ifneitherparentalgenotypecouldbeassigned( i.e. ,boththesurface andcavegenotypeswerescored “ NA ” )oriftheassignedsurfaceand cavegenotypeswereidentical;6006genomicmarkersweredeemed suitableandpreparedforlinkagemapcalculationusingthe “ crosspollination ” (CP)segregationcodingusedinJoinMap.Atthisstage, 107markerswerefoundtobeuninformative( i.e. ,asinglegenotype wassharedbyallF2individuals)anddiscardedfromfurtheranalysis. Wescreenedtheremainingset(n=5899)toidentifymarkersfailing toconformtopredictedgenotypicratios( e.g. ,1:2:1ratiosacrossthe entirepedigree);2896markersdemonstrateda x2valuemorethan50, implyingsigni cantdeparturefromthepredictedgenotyperatioand werediscardedfromfurtheranalysis.Our nalGBSmarkersetincluded3003markersevaluatedin170F2individuals. LinkagemapconstructionandQTLanalysis LinkagemapcalculationswereperformedusingJoinMap(v.4.1, Kyazma).Ourwork owusedprogramdefaultsettings,withthe followingexceptions:themaximumgroupingindependenceLOD valuewassetto50.0;linkagegroupswerecalculatedusingregression mapping;andlinkagemappingwasperformedusingtheKosambi method(Kosambi1943).LinkagegroupswereassignedbasedonindependenceLODscores.WeincreasedthemaximumgroupingindependenceLODvalueto50.0,becausethedefaultvalueof10.0did notallowsuf cientsubdivisionofourdataintoanappropriatenumberofgroups.Initialgroupingsidenti ed29groupspopulatedwith between10and225markers,withindependenceLODscoresranging from7.0to21.0.Thesegroupswerethenprocessedforformalmappingcalculations.242 |B.M.Carlson,S.W.Onusko,andJ.B.Gross

PAGE 3

The rstroundofmappingproduced28linkagegroupscomprisingatotalmaplengthof2956cM.Atthisstage,onelinkage group(comprising10markers,independenceLOD=19.0)failedto assembleintoaconsolidatedgroupandwasthereforeeliminated fromfurtheranalysis.Theremainingindividuallinkagegroupsranged inlengthfrom27.25to187.46cM,containingbetween10and225 markerswithanaverageintermarkerdistancebetween0.51and6.40 cM.Afterthisinitialroundofmapping,wefurtherscreenedexisting linkagestotargetthemostoptimal25groups( Astyanaxmexicanus haskaryotypicnumberof25;Kirby etal. 1977)andreducetheaverageintermarkerdistancetoatargetof 1cM.Accordingly,nine groups(10 # n # 45markers)wereremovedbecauseoflowmarker numberand/orunusuallyhighaverageintermarkerdistance.The ve largestgroups(154 # n # 225markers)werethensubdividedatthe lowestindependenceLODvalueresultingintwolinkagegroupscomprising20ormoremarkers.Throughoutmapping,welimitedthe in ationoftheoverallmaplengthbyeliminatingcertainmarkers sparselypopulatingdistalendsofotherwisedenselypopulatedlinkage groups.Thisresultedinsizereductionofthe velongestremaining linkagegroups(142.041 # n # 187.458cM)bysplittingthematthe lowestindependenceLODscoreatwhichagroup(comprising10or moremarkers)wasseparated.Inthesecases,thelargerofthetwo resultinggroupswasretained.Theresulting25linkagegroups(independenceLODscores10.0 # n # 24.0)weresubjectedtoadditionalmapping.Groupingsofmarkerseliminatedduringthisor asubsequentroundofmappingwereexcludedfromfurtheranalysis. Thesecondroundofmappingproduceda2556.6-cMlinkagemap composedof25linkagegroups,eachconsistingof25to171markers, ranginginlengthfrom31.18to142.78cM,withmeanintermarker distancesrangingfrom0.47to3.66cM.Usingthesamecriteria describedabove,anadditionalgroup(comprising25markersandan averageintermarkerdistanceof3.658)waseliminated.Adensely populatedgroupwithahighindependenceLOD(153markers;135.73 cM;independenceLODof24.0)wassplitand12linkagegroups (103.982 # n # 142.783cM)weretrimmed. Theresultofthisthirdand nalroundofmappingwasthen analyzedforgenomicsyntenysharedbetween Astyanaxmexicanus andthezebra shgenomeandusedtomapalbinismasaproofof concept.Albinismwasscoredasabinaryphenotypewhereinpresence ofmelanin(0)orabsenceofmelanin(1)wasassignedtoeachofthe membersofourexperimentalF2pedigrees.AllQTLanalysesofalbinismwereconductedusingR/qtl(Broman etal. 2003)runforeachof threescan-onemappingmethods:markerregression(MR),expect maximum(EM),andHaley-Knott(HK),accordingtothemethodologyinGross etal. (2014). Assignmentofgenomicsyntenybetweenthe Astyanax mexicanus and Daniorerio genomes Atpresent,physicalgenomeresourcesfor Astyanaxmexicanus arein theirearlydraftphases(McGaugh etal. 2014).Therefore,weanchoredourGBS-basedlinkagemaptothephysicalgenomeofthe mostcloselyrelated shmodelsystemwithcomprehensiveresources, Daniorerio . Astyanax and Danio aremembersofthesuperorder Ostariophysii,whichdiverged 150Myago(Briggs2005).Despite thisdistance,signi cantgenome-levelsyntenyremainsbetweenthese species(Gross etal. 2008;O ’ Quin etal. 2013).OurGBSmarkerset wasderivedfromendonucleaserestrictionsite-basedlibrariesandwas thereforeanonymous.We rstidenti edallGBSmarkersthatcould bedirectlylocalizedtoaconservedregioninthe D.rerio genome. Accordingly,weperformedBLASTsearchesofthe64-bpsequences comprisingourmarkersequencesdirectlyagainstthe Danio genome (downloadedfromtheEnsemblgenomebrowser; www.ensembl.org ). Theseandallsubsequentsearcheswereperformedusing aBLASTNscriptrunontheOhioSupercomputingCluster(OSC). Allqualitycontroldefaults,includinganexpectvalue(e-value)cutoff of10,weremaintained.Thescriptpermittedthereturnofalignments betweenagiven64-bpmarkersequenceandregionsofuptothree distincttargets( e.g. ,threedifferent Daniorerio chromosomes).In caseswhereasinglemarkersequencealignedmultipletimeswith thesametarget,rawresultswere lteredbye-value,retainingthe loweste-valuealignmentforeachmarker-targetpairing.Thereare two64-bpsequencesforeachGBSmarker,differingonlyinthateach containsoneofthetwoallelesfortheimbeddedSNP.Becausebothof thesesequenceswereincludedwhenBLASTsearchesusingthe64-bp markersequenceswereconducted,this lteringstepalsoservedto collapsetheseresultsintoasinglesetofresults,retainingthebetterof thetwoalignmentsforeachmarker-targetpairing. Insomeinstances,asinglequeriedsequencereturnedalignments withmultipletargets.Theseinstanceswereresolvedbysortingresults todeterminethe “ tophit, ” whichwasde nedashavingthelowest e-valueandhighestpercentidentity (incaseofane-valuetie)toaparticulartargetsequence.Ifthetargetofthetophit( i.e. ,thealignment withtheloweste-value)foragivenmarkersequenceagreedwiththe targetreportedforoneormoreothermarkersonthesamelinkage groupthatreturnedonlyasinglerobusthit,thenthetophitforthe markerinquestionwasconsidered “ supported ” andretained.Ifthe tophitwasnotsupportedinthisfashionbutadifferentBLASTresult was,thenthelatter “ nottophit,supported ” resultwasretainedin-stead.Ifnoneoftheresultsreturnedforamarkersequencewere supported,thenthetophitwasretained,despitethelackofsupport. Inrarecases,therewasnowaytoresolvewhichresultshouldbe retained.Resultsforthese “ unresolved ” markerswerediscarded. WhenusingBLASTsearchestoalignour64-bpmarkersdirectly tothe Daniorerio genomereturnedrelativelyfewhigh-qualityhits,we developedastrategywherebywe rstalignedourGBSmarkersequencestothe Astyanaxmexicanus genomeandtranscriptomedata.This informationwasthenusedtoidentifyhomologous Danio genomic andtranscriptomicsequences.Currentgenomicresourcesin Astyanax consistof . 10,000unplacedgenomicscaffolds(Bioproject PRJNA89115).Thecollectivesequencedataforthe Astyanax genome (GenBankAssemblyIDGCA_000372685)weredownloadedfrom Ensembl,alongwiththetranscriptsequencesfor23,042predicted genes.BLASTsearcheswereusedtodetermineputativelocations forthe64-bpsequencesofthe2235GBSmarkerscomprisingour nallinkagemapinthe Astyanax genomicandtranscriptomic datasets.Afterinitialsearcheswereperformedasdescribed, 2000-bpstretchesofgenomicsequenceharboringour64-bp GBSmarkersequenceswerealignedwiththe Danio genome.Similarly,fullsequencesforpredicted Astyanax transcriptstowhichour GBSmarkersalignedwerequeriedagainsta Danio cDNAdatabase downloadedfromEnsembl.Bothdatasetswerethen ltered(as described),yieldingasingle “ best ” Danio alignmentforeachinformativequery.Thisprocessenabledustoleveragedraftgenomicand transcriptomicdatatoaugmenttheamountofsequenceinformation associatedwithour64-bpGBSmarkersandtoidentifyhomologous genomicpositionsinawell-characterizedmodelsystem. AfterBLASTsearchesusingthedirect,genomic,andtranscriptomicalignmentmethodswerecompleted,the lteredresultsforall threewerecombined.Whenmultiplemethodsreturnedresultsforthe samemarker,asingleresultwaschosenandretainedusingthesame lteringprocessappliedtosingledatasets(above).TheCircos Volume5February2015|High-DensityGBS-BasedLinkageMapin Astyanax | 243

PAGE 4

program(Krzywinski etal. 2009)wasusedtovisualizecomparative genomicpositionsbetweenourlinkagemapandthe Astyanax and Daniorerio genomes. Positionidenti cationforpreviouslypublishedmarkers intheAstyanaxgenome PreviousmapspublishedbyGross etal. (2008)andO ’ Quin etal. (2013)wereusedtoexaminesyntenybetween Astyanax and Danio andtoprovideacomparisonbetweenthisstudyandpriorstudies. Theseauthorsprovidedpredicted Danio positionsforthemarkers usedintheiranalyses,butpositionsinthedraft Astyanax genome werenotdeterminedbecausethesestudiespredatedavailablegenomic resources.OurGBS-basedmapdoesnotshareanymarkerswiththe twopreviousmaps,soitwasnecessarytoidentifypositionsofpreviouslygeneratedmarkersin Astyanax toenablecomparisonbetween previousmappingeffortsandthosedescribedhere.Accordingly, microsatelliteandRAD-seqmarkersequences(whereavailable)for eachdatasetwerealignedwith Astyanax genomescaffoldsusingthe sameBLASTand lteringprotocolsusedforourowndata(above). Bothpreviousstudiesincludedmarkerslocatedincandidategenes. Thelocationsof Astyanax orthologsofthesecandidategeneswere identi edusingEnsembl. GBSmarkersequencesandgenotypingdataareavailablefromthe DryadDigitalRepository( http://dx.doi.org/10.5061/dryad.6s718 ). RESULTSANDDISCUSSION Ahigh-densitylinkagemapin Astyanaxmexicanus Here,wepresentadenselinkagemapfor Astyanaxmexicanus generatedusinggenotyping-by-sequencingtechnology.Thismapwas createdusing170experimentalF2individualsbasedongenotypic informationfor3003loci.Theconstructionofthismapultimately yielded25linkagegroups(thekaryotypicnumberfor Astyanax )comprising2235markersspanning2110.7cM,withanaverageintermarkerdistanceof1.052cM(Figure1,SupportingInformation, TableS1 ).Thestrategyweusedenablesapplicationofpowerful, cost-effective,next-generationsequencingtechnologytofacilitategeneticstudiesinemergingornonmodelsystems. Cross-generamarkeridenti cationwasgreatlyfacilitatedby alignment rsttodraft Astyanax genomicandtranscriptomicresources,followedbysearchesofthehomologoussequencesin Danio (Figure2,A – C).AlthoughdirectBLASTsearchesofour64-bpGBS markersequencesreturnedresultsforfewofthemarkersinourmap (1.2%),successratesweremuchhigherwhenusing Astyanax genomic (26.5%)ortranscriptomic(13.3%)sequencesasanintermediary(Table1).Each Daniorerio chromosomewasrepresentedinourcomparativegenomicanalysis,with Astyanax linkagegroupscontaining 14 – 52markers(average=30.84)comprisingancientsyntenicblocks sharedwitheachof25zebra shchromosomes(Figure2D).Ofthe 2235GBSmarkersthatconstituteourlinkagemap,784marker sequences(35.1%)weresuccessfullyidenti edinthe Daniorerio genome(Figure3A). Weperformedaproof-of-conceptanalysisusingthealbinism phenotypetovalidatetheutilityofourGBS-basedlinkagemap (Figure3,B – D).Accordingly,wemappedthemonogenictraitof albinismusingtheR/qtlpackagetoevaluatephenotypicandgenotypic dataforthe170F2hybridindividualsusedtoconstructourmap.We identi edapeakLODscoreof20.68onlinkagegroup13,associated withmarkerTP71406.Thismarkerandthesurroundingregionform asyntenicblockwithinaregionof Daniorerio chromosome6.This genomicintervalcontainsthegene Oca2 ,previouslydemonstratedto bethecausativelocusforalbinismin Astyanax cave sh.Thissupports previous ndingsofconservedsyntenyinclusiveofsigni cantportionsofchromosome6in Danio (Gross etal. 2008;O ’ Quin etal. 2013)andimpliesourdenselypopulatedmapwillenablefutureQTL studiesoftraitevolutionin Astyanax . Conservedgenomicarchitecturebetween Astyanax and Danio basedonGBSmarkers Ouranalysisofsyntenybetween Astyanax and Danio illustratesvariablelevelsofgenomicconservationacrosslinkagegroups(Figure2D, Figure3A).Certainchromosomes,forinstance,appeartohave changedlittlesincethedivergenceoftheseteleostspecies( e.g. , Danio chromosomes6and23in Astyanax linkagegroups13and15,respectively).However,other Danio chromosomesappearscattered acrossseverallinkagegroups,withoutaconsensusrepresentation foranyparticulargroup( e.g. , Danio chromosomes2and5). Webelievethese ndingsmostlikelyre ectgenomicrearrangementsthathaveoccurredsincethedivergenceofthesetwospecies. Figure1 AGBS-basedlinkagemapintheMexicancavetetra, Astyanaxmexicanus .Weanalyzed3003SNPmarkersin170individuals usinggenotyping-by-sequencingtechnology.Thislinkagemapconsistsof2235markersin25linkagegroups( A.mexicanus karyotype number=25),spanningatotaldistanceof2110.7cM(meanintermarkerdistance=1.052cM). Astyanax linkagegroup8(redbox) illustratestypicalmarkerdensityobservedinmostgroups.Thisgroup consistsof52GBSmarkersspanning67.061cMwithameanintermarkerdistanceof1.315cM. 244 |B.M.Carlson,S.W.Onusko,andJ.B.Gross

PAGE 5

However,this ndingcouldalsobeattributedtolowrepresentationof particular Danio chromosomeswithinourGBSmarkerset.Weexaminedthispossibilitybyassessingthenumberofsynteniclinks betweenourGBS-basedlinkagemapandeach Danio chromosome. Wewouldanticipatethatlongerchromosomeswouldnaturallyharbormoresynteniclinks.Valueswerethereforeexpressedasaratioof synteniclinkspermegabase(mean=0.59GBSmarkers/Mb).Althoughthemeanvalueforchromosomesthatwerenotstronglyrepresentedonanyparticularlinkagegroupinourmap( i.e. ,hadfewer than10synteniclinkswitheachlinkagegroup,mean=0.52GBS Figure2 ShortGBSsequencesidentifysyntenicstretchesbetweentwoOstariophysianfreshwater shspecies.Torevealsyntenicregions between Astyanaxmexicanus and Daniorerio ,we rstidenti edstretchesofthe Danio genomeharboringhomologoussequencestoour anonymousGBSmarkersequences(A).Individual64-bpsequencesforthe2235GBSmarkersinourlinkagemapwerecomparedwiththe Danio genomebothdirectlyandby rstaligningtolarger Astyanax genomicscaffoldsandpredictedgenetranscripts(B),followedbyalignment ofsomeorallofthelargersequencetothe Danio genomebasedonBLASTsequenceanalysis(C).Thisresultedinidenti cationofhomologous sequencesfor784 Astyanax GBSmarkerswithinthe Danio genome.Themarkerssharedbetween Danio chromosomesand Astyanax linkage groupsarerepresentedusinganOxfordplot(D). n Table1SummaryofBLASTresultsandidenti cationofmarkersusedin Astyanax -toDanio syntenicanalysis GBSMarkers to Danio GenomeaGBSMarkersto Astyanax GenomebAstyanax Genometo Danio GenomecGBSMarkersto Astyanax TranscriptomedAstyanax Transcriptometo Danio TranscriptomeeTotalno.ofBLASTqueries 2235223520882235572 BLASTresultcategories Singlerobusthit14(0.6%)1838(82.2%)255(12.2%)508(22.7%)110(19.2%) Tophit,withpositionalsupport0(0.0%)173(7.7%)92(4.4%)15(0.7%)120(21.0%) Tophit,withoutpositionalsupport10(0.4%)71(3.2%)138(6.6%)60(2.7%)61(10.7%) Nottophit,withpositionalsupport2( , 0.1%)6(0.3%)108(5.2%)2( , 0.1%)7(1.2%) Unresolved4(0.2%)14(0.6%)4(0.2%)12(0.5%)0(0.0%) Noresult2205(98.7%)133(6.0%)1491(71.4%)1638(73.3%)274(47.9%) Identi edsyntenicmarkers between Astyanax and Danio 26N/A593N/A298aResultsof64-bpGBSmarkersBLASTeddirectlytothe Daniorerio genome.bResultsof64-bpGBSmarkersBLASTeddirectlytothe Astyanax genomedraftassembly.cResultsof 2-kbgenomicintervalsharboring64-bpGBSmarkersBLASTedtothe Daniorerio genome.dResultsof64-bpGBSmarkersBLASTeddirectlytothe Astyanax predictedtranscriptome.eResultsof Astyanax transcriptsharboring64-bpGBSmarkersBLASTedtothe Daniorerio transcriptome. Volume5February2015|High-DensityGBS-BasedLinkageMapin Astyanax | 245

PAGE 6

markers/Mb,n=8)waslowerthanthatforchromosomesdemonstratingstrongsyntenywithaparticularlinkagegroup(mean=0.61 GBSmarkers/Mb,n=17),therewasnotasigni cantdifference betweenthetwogroups(t23=0.5809, P =0.5670).Thisleadsusto concludethat,althoughrepresentationofparticularchromosomesin ourdatasetmaybeacontributingfactor,itisunlikelythatthisisthe primarycauseofthedifferencesinchromosomalrepresentationpatternsobserved. Alternatively,BLASTresultsfor Astyanax GBSmarkers(orthe larger Astyanax sequencestowhichtheywerealigned)mayinclude paralogousgenesorotherwiseambiguousresultsthatcouldleadto erroneouslinksbetweenalinkagegroupanda Danio chromosome. Althoughwecannotruleoutthispossibility,wefeelourstrategy prioritizedthe “ optimal ” BLASTresultamongmultiplehitsforasingle markerleadingtoalignmentsthatagreewithnearbyunambiguous results(Table1).Asaresult,ofthe784markersinourmapforwhich aputative Danio positionwasdetermined,only15.9%(n=125)of nalcallswereunsupportedbytheresultsforothermarkersbelonging tothesamelinkagegroup( TableS1 ).Giventhatchromosomal arrangementshaveoccurredoverthe 150Mysincedivergence,we feeloursystematicapproachbestidenti esparalogousgenesand otherpotentialsourcesofambiguity. Erroneousorambiguousgenotypingdatamayhaveledto incorrectassignmentof “ cave ” and “ surface ” allelesforparticular markers.Theseerroneousassignmentscouldhaveadverselyaffected downstreamefforts,causingmarkerstobeincorrectlyplacedduring thegroupingand/ormappingstagesoflinkagemapconstruction.All effortsweremadetoensureallelicidenti cationwasaccurateusing astringentscreeningprocess(see MaterialsandMethods );however, wereliedonarelativelysmallnumberofcave,surface,andF1hybrid individuals(n=4each)toidentifyparentalallelicorigin.Similarly, therelativelysmallnumberofmeioticeventsrepresentedbythe170 Figure3 Whole-genomesyntenybetween Astyanax and Danio andaproof-of-conceptanalysisofalbinism.SynteniclinksbetweenourGBSmap andthe Danio genomewerevisualizedusingCircos(A).Eachlinerepresentsaconnectionbetweenthepositionofaparticularmarkerinour linkagemap(black;scaleincM)andahomologoussequencein Danio (variouscolors;scaleinMb).Wescoredalbinism,aMendeliantrait associatedwiththe Oca2 geneincave-dwelling Astyanax (C),andperformedQTLanalysisusingR/qtl.Eachofthreemappingmethods(MRin red;EMinblue;HKinblack)revealedpeakLODscoresof 20(LODat0.001 a threshold=6.75)at,oradjacentto,GBSmarkerTP71406on Astyanax linkagegroup13(B).HomologoussequencestoTP71406andseveralofitsneighborson Astyanax linkagegroup13areclustered togetheron Danio chromosome6nearthe Oca2 gene.AphenotypiceffectplotformarkerTP71406revealedthepredictedassociationbetween thehomozygous “ cave ” condition(genotype CC )andalbinisminF2individuals(D). n Table2Comparisonof Astyanax linkagemapsandsyntenicstudieswith Daniorerio Gross etal. 2008O ’ Quin etal. 2013CurrentAnalysis Totalno.oflinkagegroups282525 Totalno.ofgenomicmarkers4006982235 Linkagemaplength1783cM1835.5cM2110.7cM Markerdensity0.224percM0.380percM1.06percM MarkertypeMicrosatelliteMicrosatellite+RAD-seqGenotyping-by-sequencing No.of Astyanax genomicscaffoldsrepresented bymap 227350598 No.ofsyntenicmarkersidenti edbetween Astyanax and Danio 155173784 246 |B.M.Carlson,S.W.Onusko,andJ.B.Gross

PAGE 7

n Table3Comparisonofsyntenicanalysesbetween Astyanax linkagemapsandtheirassociationwiththe Daniorerio genomeacrossmultiplestudies Gross etal. 2008O ’ Quin etal. 2013CurrentAnalysis Daniorerio Chromosome No.of Syntenic LinksaRepresented Linkage Group(s)bNo.of Represented Astyanax Genome ScaffoldscNo.of Syntenic Links Represented Linkage Group(s) No.of Represented Astyanax Genome Scaffolds No.of Syntenic LinksRepresentedLinkageGroup(s) No.of Represented Astyanax Genome Scaffolds 1135, 8 ,21715 4 ,5,9,18,21,2311421,5,8,9, 12 ,13, 14 ,18,19, 21 ,2526 262,14,15,22267,12,13,16,235231,3,5,14, 16 , 17 ,18,19,22,24,2520 361,4,19644,15,253282,6,10,12, 13 ,14,15,19,22, 23 ,2523 436,72434243,5, 7 ,9,12,14,20,2417 515 1 ,5,9,10,20913 2 ,8,16,17,1911222,3,7,8,9,11,15,19,22,24,2517 69 4 ,134201,2,11,16,1816374,6,12, 13 ,16,17,22,2423 71117,22,24,2610613,22,23,256314, 5 ,7,9,10,11, 13 , 16 ,20,2425 849,12477,14, 17 6423, 6 ,7,9,10,13,14,20,21, 22 ,23,2429 953,174810, 11 7311,2,5,9,12, 13 , 15 , 19 ,22,2523 10317,18348,10,144143,5,7,16, 20 ,23,2512 110 — 0514,17,225363,4,5,7,9,15,17,18, 21 , 22 ,24,2524 127 10 ,1646 24 4261,2,3, 4 ,5,8,11,12,13,14,15,18,2018 13111, 5 67 4 ,124341,2,4,7,10, 13 ,14,21,22, 2521 1466,7493, 6 ,15,196242,5,8,10,13,14, 24 16 155 2 581,7, 12 ,14627 1 , 2 ,6,9,13,16,20,24,2517 1631337 8 ,197232,8, 10 ,11,16,17,19,20,22, 23 17 1763, 23 21201521,2,3, 5 ,6, 9 ,10,12,14,15, 20 ,22,2334 188 11 37 5 3323,5,6,7,13,14,15,16, 18 ,19,21,2426 193193125142 5 , 6 ,9,10,11,15,17,20, 23 ,2528 207 1 ,2731,23212, 3 ,9,13,14,18,19,2115 21315,17162, 7 61612, 14 ,18,1912 22412,204714,18,226421,3, 5 ,6, 7 ,9,10,11,12,13,14,15,16,18,19, 22 ,2332 236 26 3714,16,18,256333,4,7,10,14, 15 ,16,2217 2481,13,15692,3,8,118381,2, 3 , 7 , 10 ,11, 13 ,15,16, 18 ,2325 2536,7333,63313,5,7, 8 ,9,13,16,17, 20 ,2419Bold indicatesthatalistedlinkagegroupharbors veormorelinkswithagiven Danio chromosome.aIndicatesthenumberofsynteniclinksidenti edbetween Astyanax linkagemapsandeachlisted Daniorerio chromosome.bIndicatestheidentityof Astyanax linkagegroupsharboringsynteniclinkswitheachlisted Daniorerio chromosome.cIndicatesthenumberof Astyanax genomescaffoldsharboringconnectionswitheachlisted Daniorerio chromosome. Volume5February2015|High-DensityGBS-BasedLinkageMapin Astyanax | 247

PAGE 8

n Table4Representativeanalysisoflinkagegroupequivalenceandqualitybasedonhighlysyntenicchromosomesin Daniorerio andlinkagegroupsin Astyanaxmexicanus Gross etal. 2008O ’ Quin etal. 2013CurrentAnalysis Daniorerio Chromosome Principal Represented Linkage GroupaNo.of Syntenic LinksbIdentityofRepresented Astyanax Genome ScaffoldscPrincipal Represented Linkage Group No.of Syntenic Links IdentityofRepresented Astyanax Genome Scaffolds Principal Represented Linkage Group No.of Syntenic Links IdentityofRepresented Astyanax Genome Scaffolds 648KB871811.1, KB882115.1 , KB882122.1 , KB882161.1 , KB882172.1 , KB882176.1 116 KB871670.1 ,KB871811.1, KB871878.1 , KB872044.1 , KB872200.1, KB882115.1 , KB882120.1 , KB882122.1 , KB882161.1 , KB882172.1 , KB882176.1 , KB882185.1 1328 KB882256.1 , KB882253.1 , KB882235.1 ,KB882230.1, KB882185.1 , KB882171.1 , KB882161.1 , KB882152.1 , KB882122.1 , KB882120.1 , KB882115.1 , KB882082.1 , KB872595.1, KB871670.1 893KB871816.1,KB871923.1, KB882105.1 175 KB871601.1 , KB871607.1 , KB871684.1 ,KB871923.1, KB872214.1 617 KB882289.1 , KB882113.1 , KB882105.1 ,KB872252.1, KB871939.1 ,KB871817.1, KB871684.1 , KB871601.1 ,KB871595.1 1359 KB871819.1 , KB872081.1 , KB872296.1 , KB882107.1 , KB882118.1 , KB882125.1 46KB881455.1, KB872296.1 , KB882107.1 , KB872081.1 2517 KB882261.1 , KB882210.1 , KB882154.1 , KB882118.1 , KB882109.1, KB882107.1 , KB872296.1 , KB872081.1 , KB871838.1 ,KB871652.1, KB871591.1 17235 KB882084.1,KB882233.1, KB882265.1 201 KB882265.1 921 KB882265.1, KB882243.1 , KB882233.1, KB882179.1 , KB882158.1, KB882153.1 , KB882117.1, KB882084.1, KB872047.1,KB871726.1, KB871695.1 23266KB872166.1,KB880082.1, KB882102.1 ,KB882242.1 184 KB882098.1 , KB882102.1 , KB882128.1 1520KB882214.1, KB882138.1 , KB882128.1 , KB882102.1 , KB882098.1 ,KB872132.1, KB872075.1,KB871985.1Bold indicatesgenomicscaffoldscontainingsyntenicmarkersontheprincipalrepresentedlinkagegroupinourGBS-basedmapandoneormorepreviousmap s. Italic letteringindicatesscaffoldsthatcontainasyntenic markerintheGBS-basedmapandareassociatedwiththeprincipallinkagegroup(s)inpreviousmap(s)butdonotcontainasyntenicmarker(and viceversa ).aIndicatesthemostcommon( i.e. , “ principal ” )linkagegroupanchoringtotheindicated Daniorerio chromosome.bIndicatesthenumberofpointsofsyntenybetweentheprincipallinkagegroupfromthisarticleandtheindicated Daniorerio chromosome.cListstheidentityof Astyanax genomicscaffoldstowhicheachpointofsyntenyidenti es.248 |B.M.Carlson,S.W.Onusko,andJ.B.Gross

PAGE 9

F2individualsmayhaveresultedinlinkagemapinaccuracies(Gross etal. 2008).Futurecomparisonsbetweenthemapwepresenthereand a nished-grade Astyanax genomewillclarifyifregionslackingsyntenybetween Astyanax and Danio areattributabletoerrorsinour linkagemaporgenomicrearrangementsthathaveoccurredsincethe divergenceofthesetaxa. Unplaced Astyanax genomescaffoldscanbeanchored toournewlinkagemap Positionallocationsinthecurrentdraftofthe Astyanax genomewere establishedfor93.6%(n=2091)ofthe2235GBSmarkerspresentin ourmap.Thesemarkerswerelocalizedtopositionsspreadacross598 different Astyanax genomescaffolds.Our25 Astyanax linkagegroups containmarkersrepresentingbetween12(linkagegroups8and22) and55(linkagegroup3)genomescaffoldseach,withamap-wide averageof27.64scaffolds/linkagegroup.Individualgenomescaffolds containedbetween1and31GBSmarkersappearinginour nalmap, withanaverageof3.50markersperscaffold.GBSmarkerslocatedon thesamegenomicscaffoldcolocalizedtoasinglelinkagegroup87.3% ofthetime.Thissuggeststhatourrecombinationmappingsuccessfullyrecapitulatedthetruegenomicpositionsofthemarkersusedto constructourmap. Improvedlinkagemappingresourcesin Astyanax Wesoughttocompareourlinkagemapwithmapspreviously publishedbyGross etal. (2008)andO ’ Quin etal. (2013)thatalso examinedsyntenybetween Astyanax and Danio .Metricssuchasthe numberoflinkagegroups,totalmaplength,numberofmarkers,and markerdensityarecommonlyusedtocomparelinkagemapswithin species.BothourGBS-basedmapandtheRAD-seqandmicrosatellitebasedmappublishedbyO ’ Quin etal. (2013)consistof25linkage groups,matchingthe Astyanaxmexicanus karyotypenumberof25. Themicrosatellite-basedmappresentedbyGross etal. (2008)contains28groups(Table2).Althoughourmapisofcomparablelength, itrepresentsadramaticincreaseinmarkernumber(+559%compared withthatofGross etal. 2008;+320%comparedwiththatofO ’ Quin etal. 2013)andmarkerdensity(+473%comparedwiththatofGross etal. 2008;+279%comparedwiththatofO ’ Quin etal. 2013)relative topreviouslypublishedlinkagemapsforthissystem.Asaresult,we sawasubstantialincreaseinthenumberofsynteniclinksbetweenour mapand Danio (+506%comparedwiththatofGross etal. 2008; +453%comparedwiththatofO ’ Quin etal. 2013)andanincrease inthenumberofunplaced Astyanax scaffoldsthatcanbeanchoredto ourmap(+263%comparedwiththatofGross etal. 2008;+171% comparedwiththatofO ’Quin etal. 2013). Ourmapcontainsatotalof784linksbetweenourlinkagegroups andthe Daniorerio genomeandanaverageof30.84links(minimum =14,maximum=52)per Daniorerio chromosome(Table3).This representsaconsiderableimprovementovertheresultspresentedby Gross etal. (2008)(155totallinks,averagelinksper Danio chromosome=6.20,minimum=0,maximum=15)andO ’ Quin etal. (2013) (173totallinks,averagelinksper Danio chromosome=6.92,minimum=1,maximum=20).Additionally,althoughinstancesofsyntenystronglyrepresentedinpreviousmapswerealsoidenti edinthis analysis,ourmapdemonstratedincreasedrepresentationofcertain Danio chromosomespoorlyrepresentedinpreviousmaps.Forexample,Gross etal. (2008)didnotidentifylinksbetweentheirmapand Daniorerio chromosome11;however,weidenti ed36linksbetween ourmapandchromosome11.Similarly, Danio chromosomes17and 19areeachrepresentedonceinthemapofO ’ Quin etal. (2013).We identi edsubstantiallinksbetweenthesechromosomesandourlinkagegroups9(n=21)and23(n=15),respectively. Ourlinkagemapusesanentirelydifferentmarkersetthan thoseusedinpreviousmaps.Therefore,itwasnotpossibleto makedirectcomparisonswiththelinkagegroupsacrosspriorstudies. However,wecouldindirectlycomparemapsbyexaminingconnectionsbetween Astyanax genomicscaffoldsandeachlinkagemap.We examinedthe vestrongestsynteniclinksbetweensinglelinkage Figure4 Colinearitybetween Astyanax linkagegroups andgenomescaffolds.Wevisualizedthe “ anchoring ” ofsevenunplaced Astyanax genomescaffolds(various colors)tolinkagegroup23(black)inour Astyanax linkagemap.Forclarity,onlyscaffoldsharboringfouror moreGBSmarkerswereincluded.Scaffoldscorrespond todiscrete,colinearsectionsofthelinkagegroupwith minimaloverlap.Thelineararrangementofmarkersis largelypreservedbetweenthescaffoldandthelinkage group.Thescalefor Astyanax scaffoldsisinMb;thescale forlinkagegroup23isshownincM. Volume5February2015|High-DensityGBS-BasedLinkageMapin Astyanax | 249

PAGE 10

groupsinourGBS-basedmapandsingle Danio chromosomesand thenidenti edanalogousconnectionsbetweenthosechromosomes andspeci clinkagegroupsinthemapspresentedbyGross etal. (2008)andO ’ Quin etal. (2013). Astyanax genomicscaffoldsharboringmarkersassociatingeach linkagegroupwithaparticular Danio chromosomewerethencompared(Table4).Wefoundthatmanyoftheidenti ed Astyanax genomicscaffoldscolocalizetoputativelyanalogouslinkagegroups inbothourGBS-basedmapandthoseofGross etal. (2008)and/or O ’ Quin etal. (2013).However,ineverycaseexamined,ourlinkage groupswereinclusiveofamuchhighernumberof Astyanax genomic scaffoldscomparedwithpriorstudies.Thus,whilethelinkagegroups inourmaprepresentgenomicintervalssimilartothoserepresentedin priormaps,ourmapachievesahigherlevelofdetailandresolution. Theseresultsalsosuggestthatfuturemappingeffortsin Astyanax may bene tbycombiningGBSmarkerdiscoverywiththosemarkersused byGross etal. (2008)andO ’ Quin etal. (2013)togeneratethemost comprehensivelinkagemappingresource. High-densityGBS-basedlinkagemappingwillinform theAstyanaxgenomesequencingproject Preliminary Astyanax genomicresourcesenabledustolocate64-bp, anonymousGBSmarkers,andtoassessthequalityandreliabilityof our Astyanax linkagemap.Thisemergingresourcedidnotallowusto determinehowwellthe25 Astyanax chromosomesarerepresentedin ourmap.However,theseresourcesallowedustodetermineifmarkers predictedtooccurinthesamegenomescaffoldsalsoco-occurinour GBS-basedlinkagemap.Overall,weobservedahighlevelofagreementbetweenourlinkagegroupsandoneormoreunplaced Astyanax genomicscaffolds. Inmanycases,markerspresentonthesamescaffoldclustered togetheroveraportionofalinkagegroupwithlittleornointerruption fromunplacedmarkersormarkersfromotherscaffolds(Figure4).We expecttheseresultswillhelpinformchromosomalpositionsofscaffolds, giventhatlinkagemapshavebeensuccessfullyusedtoaugmentgenomicresourcesinother shspecies,includingseveralspeciesofcat sh (Liu2011;Ninwichian etal. 2012),rainbowtrout(Palti etal. 2011;Palti etal. 2012),andAtlanticsalmon(Lorenz etal. 2010).Webelieveour high-densityGBS-basedmapresour ceswillbothprovidearesourcefor morere nedQTLanalysesandinformthegenomicarchitectureofthe Astyanax genomesequencingproject. CONCLUSIONS Weconstructedahigh-densitylinkagemapfor Astyanaxmexicanus basedonhigh-throughputgenotyping-by-sequencingdata.Weleveragedemerging Astyanax genomicandtranscriptomicresourcesand Daniorerio genomicandtranscriptomicdatatolocatesyntenic regionssharedbetweenourmapandtheDanio genome.These ndingswerebasedonthephysicalpositionofhomologous(64-bp)GBS markersequences.Asexpected,basedonthesigni cantdivergence betweenthesespecies,werecoveredvaryinglevelsofsyntenybetween portionsofour Astyanax linkagegroupsandregionsofthe Danio genome.Asaproofofconcept,wesuccessfullymappedastrongQTL associatedwithalbinismanddemonstratedsigni cantconservedgenomicarchitectureintheregionssurroundingthegene Oca2 ,between Astyanax and Danio .Wesuccessfullyanchoredemerging Astyanax genomicinformationtoourGBS-bas edlinkagemap,identifyingthe putativelocationofthousandsof anonymousGBSmarkersequences withinunplaced Astyanax genomescaffolds.Thisstrategyrevealedsigni cantcolinearitybetweengenomicscaffoldsandourlinkagemap, anditdemonstratedtheutilityofhigh-density,GBS-basedlinkage mapstoinformandimprovenascentgenomicresources.Multiple comparisonswithpreviouslypublishedmapssuggestthatourGBSbasedmapoffersahigherlevelofresolutionandagreaternumberof connectionsbetween Astyanax and Danio genomes.Wehopethatthis resourceandtechnologywillacceleratethesearchandidenti cationof genesmediatingcave-associatedtraitsin Astyanax ,facilitatethegenomicassemblyforthissystem,andpro veusefultoothernaturalmodel systemsofevolutionaryandbiomedicalrelevance. ACKNOWLEDGMENTS TheauthorsthankAmandaKrutzler,BethanyStahl,andmembersof theGrosslaboratoryforvaluableeffortandinput.Theauthorsare alsogratefultoWesleyWarren,SuzanneMcGaugh,andtheGenome InstituteatWashingtonUniversityforprovidingaccesstothedraft genomeassembly(BioprojectPRJNA89115NCBIaccessionnumber APWO00000000;supportedbyNIHgrantR24RR032658-01to W.W.).Additionally,theauthorsthankSuzanneMcGaughforprovidingB.M.C.andothermembersoftheGrosslaboratoryforinstruction inscript-basedBLASTsearchmethods.Thisprojectwassupported byNationalInstitutesofHealth(NationalInstituteofDentaland CraniofacialResearch)grantDE022403(toJ.B.G.)andaCaveResearch FoundationGraduateStudent ResearchGrant(toB.M.C.). LITERATURECITEDBorowsky,R.,andH.Wilkens,2002Mappingacave shgenome:polygenic systemsandregressiveevolution.J.Hered.93:19 – 21. Breder,C.M.,Jr,1943aApparentchangesinphenotypicratiosofthe Characinsatthetypelocalityof Anoptichthysjordani HubbsandInnes. Copeia1943:26 – 30. Breder,C.M.,Jr,1943bProblemsinthebehaviorandevolutionofaspecies ofblindcave sh.Trans.N.Y.Acad.Sci.5:168 – 176. Briggs,J.,2005Thebiogeographyofotophysan shes(Ostariophysi:Otophysi):Anewappraisal.J.Biogeogr.32:287 – 294. Broman,K.W.,H.Wu,S.Sen,andG.A.Churchill,2003R/qtl:QTL mappinginexperimentalcrosses.Bioinformatics19:889 – 890. Burchards,H.,A.Dölle,andJ.Parzefall,1985Aggressivebehaviourofan epigeanpopulationof Astyanaxmexicanus (Characidae,Pisces)andsome observationsofthreesubterraneanpopulations.Behav.Processes11:225 – 235. Davey,J.W.,P.A.Hohenlohe,P.D.Etter,J.Q.Boone,J.M.Catchen etal. , 2011Genome-widegeneticmarkerdiscoveryandgenotypingusing next-generationsequencing.Nat.Rev.Genet.12:499 – 510. Duboué,E.R.,A.C.Keene,andR.L.Borowsky,2011Evolutionaryconvergenceonsleeplossincave shpopulations.Curr.Biol.21:671 – 676. Elipot,Y.,H.Hinaux,J.Callebert,andS.Retaux,2013Evolutionaryshift from ghtingtoforaginginblindcave shthroughchangesintheserotoninnetwork.Curr.Biol.23:1 – 10. Elshire,R.J.,J.C.Glaubitz,Q.Sun,J.A.Poland,K.Kawamoto etal. , 2011Arobust,simplegenotyping-by-sequencing(GBS)approachfor highdiversityspecies.PLoSONE6:e19379. Everett,M.V.,M.R.Miller,andJ.E.Seeb,2012Meioticmapsofsockeye salmonderivedfrommassivelyparallelDNAsequencing.BMCGenomics13:521. Gross,J.B.,M.Protas,M.Conrad,P.E.Scheid,O.Vidal etal. ,2008Synteny andcandidategenepredictionusingananchoredlinkagemapof Astyanax mexicanus .Proc.Natl.Acad.Sci.USA105:20106 – 20111. Gross,J.B.,R.Borowsky,andC.J.Tabin,2009Anovelrolefor Mc1r in theparallelevolutionofdepigmentationinindependentpopulations ofthecave sh Astyanaxmexicanus .PLoSGenet.5(1):e1000326.DOI: 10.1371/journal.pgen.1000326. Gross,J.B.,A.Furterer,B.M.Carlson,andB.A.Stahl,2013Anintegrated transcriptome-wideanalysisofcaveandsurfacedwelling Astyanax mexicanus .PLoSONE8(2):e55659.DOI:10.1371/journal.pone.0055659. Gross,J.B.,A.J.Krutzler,andB.M.Carlson,2014Complexcraniofacial changesinblindcave-dwelling sharemediatedbygeneticallysymmetric andasymmetricloci.Genetics196:1303 – 1319. 250 |B.M.Carlson,S.W.Onusko,andJ.B.Gross

PAGE 11

Hecht,B.C.,N.R.Campbell,D.E.Holecek,andS.R.Narum, 2013Genome-wideassociationrevealsgeneticbasisforthepropensity tomigrateinwildpopulationsofrainbowandsteelheadtrout.Mol.Ecol. 22:3061 – 3076. Hess,J.E.,N.R.Campbell,M.F.Docker,C.Baker,A.Jackson etal. , 2014Useofgenotyping-by-sequencingdatatodevelopahighthroughputandmulti-functionalSNPpanelforconservationapplications inPaci clamprey.Mol.Ecol.Resour.Advanceonlinepublication.Epub 2014Jun5.doi:10.1111/1755 – 0998.12283. Houston,R.D.,J.W.Davey,S.C.Bishop,N.R.Lowe,J.C.Mota-Velasco etal. ,2012CharacterisationofQTL-linkedandgenome-widerestrictionsite-associatedDNA(RAD)markersinfarmedAtlanticsalmon. BMCGenomics13:244. Hubbs,C.L.,andW.T.Innes,1936The rstknownblind shofthefamily Characidae:AnewgenusfromMexico.Occas.Pap.Mus.Zool.Univ. Mich.342:1 – 7. Jeffery,W.R.,2001Cave shasamodelsysteminevolutionarydevelopmentalbiology.Dev.Biol.231:1 – 12. Jeffery,W.R.,2009Regressiveevolutionin Astyanax cave sh.Annu.Rev. Genet.43:25 – 47. Kirby,R.F.,K.W.Thompson,andC.Hubbs,1977Karyotypicsimilarities betweentheMexicanandblindtetras.Copeia1977:578 – 580. Kosambi,D.D.,1943Theestimationofmapdistancesfromrecombination values.Ann.Eugen.12:172 – 175. Kowalko,J.E.,N.Rohner,S.B.Rompani,B.K.Peterson,T.A.Linden etal. , 2013Lossofschoolingbehaviorincave shthroughsight-dependent andsight-independentmechanisms.Curr.Biol.23:1874 – 1883. Krzywinski,M.,J.Schein,I.Birol,J.Connors,R.Gascoyne etal. , 2009Circos:Aninformationaestheticforcomparativegenomics.GenomeRes.19:1639 – 1645. Larson,W.A.,L.W.Seeb,M.V.Everett,R.K.Waples,W.D.Templin etal. , 2014Genotypingbysequencingresolvesshallowpopulationstructure toinformconservationofChinooksalmon( Oncorhynchustshawytscha ). Evol.Appl.7:355 – 369. Li,C.,G.Waldbieser,B.Bosworth,B.H.Beck,W.Thongda etal. ,2014SNP discoveryinwildanddomesticatedpopulationsofbluecat sh, Ictalurus furcatus ,usinggenotyping-by-sequencingandsubsequentSNPvalidation.Mol.Ecol.Resour.14:1261 – 1270. Liu,Z.,2011Developmentofgenomicresourcesinsupportofsequencing, assembly,andannotationofthecat shgenome.Comp.Biochem.Physiol.D6:11 – 17. Lorenz,S.,S.Brenna-Hansen,T.Moen,A.Roseth,W.S.Davidson etal. , 2010BAC-basedupgradingandphysicalintegrationofageneticSNP mapinAtlanticsalmon.Anim.Genet.41:48– 54. Lu,F.,A.E.Lipka,J.Glaubitz,R.Elshire,J.H.Cherney etal. , 2013Switchgrassgenomicdiversity,ploidy,andevolution:Novelinsights fromanetwork-basedSNPdiscoveryprotocol.PLoSGenet.9:e1003215. McGaugh,S.E.,J.B.Gross,B.Aken,M.Blin,R.Borowsky etal. ,2014The cave shgenomerevealscandidategenesforeyeloss.Nat.Commun.5:5307. Ninwichian,P.,E.Peatman,H.Liu,H.Kucuktas,B.Somridhivej etal. , 2012Second-generationgeneticlinkagemapofcat shanditsintegrationwiththeBAC-basedphysicalmap.G3(Bethesda)2:1233 – 1241. O ’ Quin,K.E.,M.Yoshizawa,P.Doshi,andW.R.Jeffery,2013Quantitative geneticanalysisofretinaldegenerationintheblindcave sh Astyanax mexicanus .PLoSONE8:e57281. Ogden,R.,K.Gharbi,N.Mugue,J.Martinsohn,H.Senn etal. , 2013Sturgeonconservationgenomics:SNPdiscoveryandvalidation usingRADsequencing.Mol.Ecol.22:3112 – 3123. Palti,Y.,C.Genet,M.C.Luo,A.Charlet,G.Gao etal. ,2011A rst generationintegratedmapoftherainbowtroutgenome.BMCGenomics 12:180. Palti,Y.,C.Genet,G.Gao,Y.Hu,F.M.You etal. ,2012Asecondgenerationintegratedmapoftherainbowtrout( Oncorhynchusmykiss )genome:Analysisofconservedsyntenywithmodel shgenomes.Mar. Biotechnol.(NY)14:343 – 357. Pottin,K.,H.Hinaux,andS.Rétaux,2011Restoringeyesizein Astyanax mexicanus blindcave shembryosthroughmodulationoftheShhand Fgf8forebrainorganisingcentres.Development138:2467 – 2476. Protas,M.E.,C.Hersey,D.Kochanek,Y.Zhou,H.Wilkens etal. , 2006Geneticanalysisofcave shrevealsmolecularconvergenceinthe evolutionofalbinism.Nat.Genet.38:107 – 111. Protas,M.,M.Conrad,J.B.Gross,C.Tabin,andR.Borowsky, 2007RegressiveevolutionintheMexicancavetetra, Astyanaxmexicanus .Curr.Biol.17:452 – 454. Protas,M.,I.Tabansky,M.Conrad,J.B.Gross,O.Vidal etal. , 2008Multi-traitevolutioninacave sh, Astyanaxmexicanus .Evol. Dev.10:196 – 209. Rowe,H.C.,S.Renaut,andA.Guggisberg,2011RADintherealmofnextgenerationsequencingtechnologies.Mol.Ecol.20:3499 – 3502. S x ado glu,P.,1956ApreliminaryreportonthegeneticsoftheMexicancave characins.Copeia1956:113– 114. Salin,K.,Y.Voituron,J.Mourin,andF.Hervant,2010Cavecolonization withoutfastingcapacities:Anexamplewiththe sh Astyanaxfasciatus mexicanus .Comp.Biochem.Physiol.A156:451 – 457. Schemmel,C.,1980Studiesonthegeneticsoffeedingbehaviourinthecave sh Astyanaxmexicanus f. anoptichthys .Anexampleofapparent monofactorialinheritancebypolygenes.Z.Tierpsychol.53:9 – 22. Yamamoto,Y.,D.W.Stock,andW.R.Jeffery,2004Hedgehogsignalling controlseyedegenerationinblindcave sh.Nature431:844 – 847. Yoshizawa,M.,Y.Yamamoto,K.E.O ’ Quin,andW.R.Jeffery, 2012Evolutionofanadaptivebehavioranditssensoryreceptorspromoteseyeregressioninblindcave sh.BMCBiol.10:108.Communicatingeditor:A.S.McCallion Volume5February2015|High-DensityGBS-BasedLinkageMapin Astyanax | 251


printinsert_linkshareget_appmore_horiz

Download Options

close


  • info Info

    There are only PDFs associated with this resource.

  • link PDF(s)



Cite this item close

APA

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

MLA

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

CHICAGO

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

WIKIPEDIA

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