GATK使用说明及流程

GATK使用说明及流程1.要分析的序列名称中，一般不要有空格2.准备Reference文件(需为fasta格式)及比对1）Index处理：生成一个ref.fasta.fai的文件2）生成.dict文件：samtoolsfaidxreference.fastajava-jarpicard-tools/CreateSequenceDictionary.jarR=reference.fastaO=reference.dictbwaindex-abwtswref.fastabwamem-t16-Mref.fastaread.fqmates.fqsample.sam转换结果文件到bam格式java-jarpicardtools/SamFormatConvertI=xx.samo=xx.bamorsamtoolsview-bSxx.sam-oxx.bam3.准备样本的BAM文件1）Sortthealignedreadsbycoordinateorder2）Markduplicates3）Addreadgroupinformation(同时具有，sam2bam转换、sort功能，可合并setp1)4）IndextheBAMfilejava-jarpicardtools/SortSam.jarINPUT=unsorted_reads.bamOUTPUT=sorted_reads.bamSORT_ORDER=coordinate(input可以输入sam文件，output输出bam，省去上述的格式转换)java-jarpicardtools/MarkDuplicates.jarINPUT=sorted_reads.bamOUTPUT=dedup_reads.bamMETRICS_FILE=sample01.dedup.metricsMAX_FILE_HANDLES=1000注意：MAX_FILE_HANDLES=Integer，参数由“ulimit-n”获得极限值。Duringthesequencingprocess,thesameDNAmoleculescanbesequencedseveraltimes.Theresultingduplicatereadsarenotinformativeandshouldnotbecountedasadditionalevidencefororagainstaputativevariant.Theduplicatemarkingprocess(sometimescalleddeduppinginbioinformaticsslang)identifiesthesereadsassuchsothattheGATKtoolsknowtoignorethem.java-jarpicardtools/AddOrReplaceReadGroups.jarI=dedup_reads.bamO=addrg_reads.bamID=group1LB=lib1PL=illuminaPU=unit1SM=sample1ID=StringReadGroupIDDefaultvalue:1.Thisoptioncanbesetto'null'toclearthedefaultvalue.LB=StringReadGroupLibraryRequired.PL=StringReadGroupplatform(e.g.illumina,solid)Required.PU=StringReadGroupplatformunit(eg.runbarcode)Required.SM=StringReadGroupsamplenameRequired.java-jarpicardtools/BuildBamIndexI=addrg_reads.bamorsamtoolsindexaddrg_reads.bam例如1）bwa比对bwaindex-abwtswref.fastabwamem-t16-Mref.fastaread.fqmates.fqsample.sam2)转换sam到bamsamtoolsview-bSsample01.sam-osample01.bamjava-jarpicardtools/SamFormatConvertI=xx.samo=xx.bam3）排序java-jarpicardtools/SortSam.jarI=sample.bamO=sample.sorted.bamsort_order=coordinate4）去重复java-jarpicardtools/MarkDuplicates.jarINPUT=sorted_reads.bamOUTPUT=dedup_reads.bamMETRICS_FILE=sample01.dedup.metricsMAX_FILE_HANDLES=10005）分组java-jarpicardtools/AddOrReplaceReadGroups.jarI=sample.sorted.bamO=group.bamID=group1LB=lib1PL=illuminaPU=unit1SM=sample16）index样品java-jar~/my_bin/picardtools1.94/BuildBamIndex.jarI=group.bam4.使用参数--------------------------------------------------------------------------------TheGenomeAnalysisToolkit(GATK)v2.6-4-g3e5ff60,Compiled2013/06/2414:48:56Copyright(c)2010TheBroadInstituteForsupportanddocumentationgoto(requiredString)Typeofanalysistorun.-I,--input_fileinputfile(s),SAMorBAM-rbs,--read_buffer_sizeNumberofreadsperSAMfiletobufferinmemory--BQSR/-BQSR(File)Theinputcovariatestablefilewhichenableson-the-flybasequalityscorerecalibration(intendedforusewithBaseRecalibratorandPrintReads).Enableson-the-flyrecalibrateofbasequalities.ThecovariatestablesareproducedbytheBaseQualityScoreRecalibratortool.Pleasebeawarethatoneshouldonlyrunrecalibrationwiththecovariatesfilecreatedonthesameinputbam(s).-K,--gatk_keyGATKKeyfile.Requiredifrunningwith-etNO_ET.Pleasesee-home-and-how-does-it-affect-me#latestfordetails.--intervals/-L(List[IntervalBinding[Feature]])Oneormoregenomicintervalsoverwhichtooperate.Canbeexplicitlyspecifiedonthecommandlineorinafile(includingarodfile).UsingthisoptiononecaninstructtheGATKenginetotraverseoveronlypartofthegenome.Thisargumentcanbespecifiedmultipletimes.Onemayusesamtools-styleintervalseitherexplicitly(e.g.-Lchr1or-Lchr1:100-200)orlistedinafile(e.g.-LmyFile.intervals).Additionally,onemayspecifyarodfiletotraverseoverthepositionsforwhichthereisarecordinthefile(e.g.-Lfile.vcf).TospecifythecompletelyunmappedreadsintheBAMfile(i.e.thosewithoutareferencecontig)use-Lunmapped.-XL,--excludeIntervalsOneormoregenomicintervalstoexcludefromprocessing.Canbeexplicitlyspecifiedonthecommandlineorinafile(includingarodfile)--reference_sequence/-R(File)Referencesequencefile.--num_threads/-nt(Integerwithdefaultvalue1)Howmanydatathreadsshouldbeallocatedtorunningthisanalysis..Howmanydatathreadsshouldbeallocatedtothisanalysis?DatathreadscontainsNcputhreadsperdatathread,andactascompletelydataparallelprocessing,increasingthememoryusageofGATKbyMdatathreads.Datathreadsgenerallyscaleextremelyeffectively,upto24cores......5.分析流程1）MappingandDuplicateMarking2）LocalRealignment3）BaseQualityRecalibration该步骤的运行，需要使用已知的snp/indel信息做参考。若没有已知信息，可以先用GATK和samtools初步获得，取其一致snp/indel信息，作为参考。具体可参考他人博客：:variantscallingbyGATKjava-jarGenomeAnalysisTK.jar\-Rref.fasta\-TUnifiedGenotyper\-Isample01.realn.bam\-osample01.gatk.raw.vcf\-stand_call_conf30.0\-stand_emit_conf0\-glmBOTH\-rfBadCigar这边有一个-rf参数，是用来过滤掉不符合要求的reads，这边是把包含错误的Cigar字符串的reads给排除掉，关于Cigar字符串可以参考关于sam文件的说明(TheSAMFormatSpecification)，sam文件的第六行就是这边的Cigar字符串，-rf的其他参