dragenGermline

A workflow for calling SNVs on fastq inputs in germline mode

Overview

Dependencies

• gsi hg38 modules : hg38-dbsnp 138

Usage

Cromwell

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

Inputs

Required workflow parameters:

Parameter	Value	Description
sampleInputs	Array[InputGroup]	Input structure with tumor fastq files and read group strings
outputFileNamePrefix	String	Prefix for output files
reference	String	The genome reference build. For example: hg19, hg38, mm10

Optional workflow parameters:

Parameter	Value	Default	Description

Optional task parameters:

Parameter	Value	Default	Description
extractInfoLine.timeout	Int	4	Timeout for the job
extractInfoLine.jobMemory	Int	4	Job allocated RAM
composeList.jobMemory	Int	4	Job allocated RAM
composeList.timeout	Int	4	Timeout for the job
runDragenGermline.enableDupMarking	Boolean	true	Flag for duplicate marking, true by default
runDragenGermline.enableTargeted	Boolean	true	Flag for enabling calling on targets like HBA, GBA etc. clusters
runDragenGermline.additionalParameters	String?	None	Additional dragen parameters
runDragenGermline.timeout	Int	96	Hours before task timeout

Outputs

Output	Type	Description	Labels
unfilteredVcf	File	SNV calls before applying any filters	vidarr_label: unfilteredVcf
filteredVcf	File	SNV calls with filter information attached	vidarr_label: filteredVcf
ploidyVcf	File?	Ploidy vcf file	vidarr_label: ploidyVcf
targetedVcf	File?	Targeted vcf file	vidarr_label: targetedVcf

Commands

This section lists command(s) run by dragenGermline workflow

• Running dragenGermline

dragenGermline is a workflow which launches DRAGEN SNV calling pipeline. It creates input list based on information passed by the respective olive and then aligns all reads using input fastq files, calling SNVs in Germline mode after that. It applies a number of filters and adds annotations from dbSNP database, if available

Extracting information from RG line

     python3<<CODE
     import json
     import re
     jsonInput = "~{write_json(fastqInput)}"
     with open(jsonInput, "r") as ji:
         inputData = json.load(ji)
     ji.close()
 
     try:
         myPattern = r'\S+?\:\S+'
         rgs = re.findall(myPattern, inputData['readGroup'])
         for rgroup in rgs:
             if rgroup.startswith("ID:"):
                 RGID = rgroup.split(":")[1]
                 Lane = rgroup.split("_")[-2]
             if rgroup.startswith("SM:"):
                 RGSM = rgroup.split(":")[1]
             if rgroup.startswith("LB:"):
                 RGLB = rgroup.split(":")[1]
         fastqR1 = inputData['fastqR1']
         fastqR2 = inputData['fastqR2']
         myResult = ",".join([RGID, RGSM, RGLB, Lane, fastqR1, fastqR2])
         print(myResult)
     except:
         print("Error parsing string")
     CODE 

Composing input lists

    python3<<CODE
    l = "~{sep=' ' inputLines}"
    inLines = l.split()
    linesToPrint = ["RGID,RGSM,RGLB,Lane,Read1File,Read2File\n"]
    for inputString in inLines:
        inputString.rstrip()
        if not inputString.startswith("Error"):
            linesToPrint.append(inputString + "\n")
 
    with open("~{outputFileName}", "w") as tl:
        tl.writelines(linesToPrint)
    tl.close() 
    CODE

Running SNV caller

       dragen -f -r ~{refDir} \
       --fastq-list ~{sampleFastqList} \
       --enable-duplicate-marking ~{enableDupMarking} \
       --enable-variant-caller true \
       --enable-targeted ~{enableTargeted} \
       --dbsnp ~{dbSNP} \
       --output-directory . \
       --output-file-prefix ~{outputFileNamePrefix} ~{additionalParameters}

Support

For support, please file an issue on the Github project or send an email to [email protected] .

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