strelkaSomatic

Strelka variant caller in somatic mode

Overview

Dependencies

Usage

Cromwell

java -jar cromwell.jar run strelkaSomatic.wdl --inputs inputs.json

Inputs

Required workflow parameters:

Parameter	Value	Description
`tumorBam`	File	Input BAM file with tumor data
`tumorBai`	File	BAM index file for tumor data
`normalBam`	File	Input BAM file with normal data
`normalBai`	File	BAM index file for normal data
`reference`	String	Reference assembly id
`snvsVcfGather.refIndex`	String	fai of the reference assembly, determines the ordering by chromosome
`indelsVcfGather.refIndex`	String	fai of the reference assembly, determines the ordering by chromosome

Optional workflow parameters:

Parameter	Value	Default	Description
`bedFile`	String?	None	BED file designating regions to process
`numChunk`	Int?	None	If BED file given, number of chunks in which to split each chromosome
`outputFileNamePrefix`	String	"strelkaSomatic"	Prefix for output files

Optional task parameters:

Parameter	Value	Default	Description
`splitIntervals.gatk`	String	"$GATK_ROOT/bin/gatk"	GATK executable path
`splitIntervals.splitIntervalsExtraArgs`	String?	None	Additional arguments for the 'gatk SplitIntervals' command
`splitIntervals.nonRefModules`	String	"gatk/4.1.2.0"	Environment modules other than the genome refence
`splitIntervals.memory`	Int	32	Memory allocated for job
`splitIntervals.overhead`	Int	8	Memory overhead for running on a node
`splitIntervals.timeout`	Int	72	Hours before task timeout
`convertIntervalsToBed.modules`	String	"python/3.7"	Environment modules
`convertIntervalsToBed.memory`	Int	16	Memory allocated for job
`convertIntervalsToBed.timeout`	Int	4	Hours before task timeout
`configureAndRunParallel.nonRefModules`	String	"python/2.7 samtools/1.9 strelka/2.9.10"	Environment module names other than genome reference
`configureAndRunParallel.jobMemory`	Int	16	Memory allocated for job
`configureAndRunParallel.threads`	Int	4	Number of threads for processing
`configureAndRunParallel.timeout`	Int	4	Hours before task timeout
`snvsVcfGather.modules`	String	"gatk/4.1.2.0"	Environment module names and version to load (space separated) before command execution
`snvsVcfGather.gatk`	String	"$GATK_ROOT/bin/gatk"	GATK to use
`snvsVcfGather.memory`	Int	16	Memory allocated for job
`snvsVcfGather.timeout`	Int	12	Hours before task timeout
`indelsVcfGather.modules`	String	"gatk/4.1.2.0"	Environment module names and version to load (space separated) before command execution
`indelsVcfGather.gatk`	String	"$GATK_ROOT/bin/gatk"	GATK to use
`indelsVcfGather.memory`	Int	16	Memory allocated for job
`indelsVcfGather.timeout`	Int	12	Hours before task timeout
`configureAndRunSingle.regionsBed`	String?	None	BED file designating regions to process
`configureAndRunSingle.nonRefModules`	String	"python/2.7 samtools/1.9 strelka/2.9.10"	Environment module names other than genome reference
`configureAndRunSingle.jobMemory`	Int	16	Memory allocated for job
`configureAndRunSingle.threads`	Int	4	Number of threads for processing
`configureAndRunSingle.timeout`	Int	4	Hours before task timeout

Outputs

Output	Type	Description
`snvsVcf`	File	VCF file with SNVs, .gz compressed
`indelsVcf`	File	VCF file with indels, .gz compressed

Commands

This section lists command(s) run by strelkaSomatic workflow

Running strelkaSomatic

Main task, configuring and running Strelka2

	set -eo pipefail

	~{writeBed}
	~{indexFeatureFile}

	configureStrelkaSomaticWorkflow.py \
	--normalBam ~{normalBam} \
	--tumorBam ~{tumorBam} \
	--referenceFasta ~{refFasta} \
	~{regionsBedArg} \
	--runDir .

	./runWorkflow.py -m local -j ~{threads}

Converting interval files to .bed

.bed files use 0-based index, we need to do a proper conversion of interval files

        python3 <<CODE
        import os, re
        intervalFiles = re.split(",", "~{sep=',' intervalFiles}")
        for intervalFile in intervalFiles:
            items = re.split("\.", os.path.basename(intervalFile))
            items.pop() # remove .interval_list suffix
            bedName = ".".join(items)+".bed"
            with open(intervalFile, 'r') as inFile, open(bedName, 'w') as outFile:
                for line in inFile:
                    if not re.match("@", line): # omit the GATK header
                        fields = re.split("\t", line.strip())
                        fields[1] = str(int(fields[1]) - 1)
                        outFile.write("\t".join(fields)+"\n")
        CODE

Splitting intervals

SplitIntervals is used to produce a requested number of files to make the analysis parallel, decreasing demand for resources per chunk and improving speed

	set -eo pipefail

	mkdir interval-files
	ln -s ~{refFai}
	ln -s ~{refDict}
	~{gatk} --java-options "-Xmx~{memory-8}g" SplitIntervals \
	-R ~{refFasta} \
	~{intervalsArg} \
	~{scatterArg} \
	--subdivision-mode BALANCING_WITHOUT_INTERVAL_SUBDIVISION \
	-O interval-files \
	~{splitIntervalsExtraArgs}

	cp interval-files/*.interval_list .

Combining vcf files after scatter

This task uses GatherVcfs tools from GATK suite which needs the files to be ordered. No overlap between covered features allowed.

   set -eo pipefail

   ~{gatk} GatherVcfs \
   -I ~{sep=" -I " vcfs} \
   -R ~{refFasta}
   -O ~{outputName}

   gzip ~{outputName}

Support

For support, please file an issue on the Github project or send an email to gsi@oicr.on.ca .

Generated with generate-markdown-readme (https://github.com/oicr-gsi/gsi-wdl-tools/)

oicr-gsi / strelkaSomatic