内容简介： 本研究介绍了一种应用重测序方法进行数量性状基因座(QTL)定位的方法,该方法选用的研究对象可以是具有极端性状的一对品种杂交后获得的近交重组系群体(RILs)或F2代群体,选择群体中具有两种不同极端性状的20-50个个体分别构建DNA混池后进行重测序,通过对比两个混池的SNP位点的测序深度相关的一个参数(SNP-index)来定位QTL。该方法能用于群体遗传学研究,能快速识别人工选育及自然选择发生的染色体区域。

WholegenomesequencingrevealsQTL175

TraditionallyQTLshavebeenidenti edbylinkageanaly-sisofprogenyderived

fromacrossbetweenparentsshow-ingcontrastingphenotypesforatraitofinterest.To

performlinkageanalysis,DNA

markerscapableofdiscrimi-natingparentalgenomesarerequired.Duetothisrequire-ment,parents

forcrossesareselectedfromgeneticallydistantlyrelatedcultivars.ThisentailsthatparentsmaybedifferentinmanyQTLscontrollingagivenphenotype,complicatingtheisolationofindividualloci.Ontheotherhand,whenevercloselyrelatedparentsareused,identi -analysismethodtoidentifyDNAmarkerstightlylinkedtothecausalgeneforagivenphenotype(Giovannonietal.,1991;Michelmoreetal.,1991).Followingacrossbetweenparentallinesshowingcontrastingphenotypes,theresultingF2progenyarescoredforsegregationofthephenotype.TwobulkedDNAsamplesaregeneratedfromtheprogenyshowingcontrastingphenotypes,andDNAmarkersexhibitingdif-ferencesbetweenthetwobulksarescreened.Intheorigi-nalreports,DNAsbulkedfromF2progenywerescreenedwithrestrictionfragmentpolymorphisms(RFLPs)andran-domampli edpolymorphicDNA(RAPD)markerstoiden-tifythemarkerslinkedtothetraitsofinterest.Later,BSAwasappliedtoidentifyQTLs(Mansuretal.,1993;DarvasiandSoller,1994),whichissometimescalled‘selectiveDNApooling’.However,intheseanalyses,theavailabilityofDNAmarkerswasthemainfactorlimitingeffectivenessofthemethods.Furthermore,genotypingofeachmarkerforthetime-consumingandcostly.sequencinghasacceleratedtheanalysisofQTLsinyeast,amodelorgan-ismwitharelativelysmallgenomesize(12.5Mb).Ehrenr-eichetal.(2010)madeacrossbetweentwodiploidyeaststrainsandobtainedalargenumberofhaploidprogeny.TheythenappliedBSAtoselecttwopopulationswithextremephenotypes,andgenotypedthebulkedDNAwithasinglenucleotidepolymorphism(SNP)microarrayandwholegenomesequencing,whichsuccessfullyidenti edthelocationofQTLsinvolvedinresistancetovariouschemicalcompounds.TheproposedmethodiscalledX-QTLsinceanextremelylargenumberofprogenywereusedineachbulk.Similarapplicationsofwhole-genomesequencingtoBSAarereportedinyeastwithsuccessfulidenti cationofQTLsforxyloseutilization(Wengeretal.,2010),heattolerance(Partsetal.,2011),andethanoltoler-ance(Swinnenetal.,2012).However,theapplicationofwholegenomesequencingtoBSAforidentifyingQTLsinwhole-genomeresequencingoftwoDNAbulksofprogeny

(eachwith20–50individuals)showingextremephenotypicvaluesbynext-generationsequencing(NGS)technology.SincethisapproachhasawideapplicabilityinQTLidenti cationinplantspecies,includingcrops,weproposetonamethemethodQTL-seqasspeci callyappliedtoplantspecies.BecauseitdoesnotrequireDNAmarkerdevelopmentandgenotyping,themosttime-consumingandcostlyprocedureneededfortheconventionalQTLanalysis,QTL-seqallowstherapididenti cationofQTLs.RESULTS

PrincipleofQTL-seq

TheprincipleofQTL-seqisshowninFigure1andisexplainedbytakingriceasanexample.QTL-seqcombinesbulked-segregantanalysis(Giovannonietal.,1991;Michel-moreetal.,1991;Mansuretal.,1993;DarvasiandSoller,1994)andwhole-genomeresequencingforrapididenti ca-tionofthegenomicregionsthatdifferbetweenthetwoparentsusedinageneticcrossandalsocontributetothehigherandlowervaluesofthetraitsofinterestamongtheresultingprogeny.ForQTLmappingusingQTL-seq,we rstgenerateamappingpopulationbycrossingtwoculti-canbeusedforQTL-seqdependingonthetraitstobestudied.Recombinantinbredlines(RILs)anddoubledhap-loids(DH)showahighdegreeofhomozygosity,andindi-vidualsineachlinecanberegardedasproxyclonesthatallowreplicatedmeasurementsofthephenotype,andthusaresuitablefordetectingQTLsofminoreffects.TheadvantageofusinganF2populationistheashorttimerequiredforitsgeneration.However,noreplicatedmea-surementsarepossibleforeachgenotype.Asaresult,theapproachisnotsuitablefordetectingminoreffectQTLs.Aftertheprogenyofamappingpopulationaremeasuredforthefocusedtrait,wescoresegregationofthepheno-type.IfthenumberofQTLsinvolvedinthetraitvariationismultiple,frequencydistributionofmeasuredvalueswillbeclosetothenormal(Gaussian)distribution(Figure1a).Here,wefocusedonthemultipleprogenyshowingextremephenotypes,i.e.thoseexhibitingthehighestandthelowestextremevalues.WesampledDNAfrom10to20individualsfromeachextremityandbulkedthemtogener-ate‘Highest’bulkand‘Lowest’bulk.EachofthebulkedDNAswasappliedtowholegenomeresequencingwitha>6xgenomecoverage.WeexpectthebulkedDNAtocontaingenomesfrombothparentsina1:1ratioforthemajorityofgenomicregions.However,weshoulddetectunequalrepresentationofthegenomesfromthetwopar-entsinthegenomicregionsharboringQTLforthepheno-typicdifferencebetween‘Highest’and‘Lowest’bulks.

©2013TheAuthors

ThePlantJournal©2013BlackwellPublishingLtd,ThePlantJournal,(2013),74,174–183