Implement new mutect2 variant calling strategy

[syansanofi]

Issue

Update Mutect2 variant calling to include latest gatk (4.2.0.0) recommended best practices.

Approach

• Update mutect2 template
• Add or modify existing templates to include post mutect2 clean up templates (executed in this order)
  • GetPileupSummaries
  • CalculateContamination
  • LearnReadOrientation
  • SortBamout
  • FilterMutectCalls required step for downstream consensus variant calling
  • FilterAlignmentArtifacts

Related to this change, additional germline references are required (many are seen in mutect1 list, however not all are listed)

• Sequence Dictionary: seq dict of reference
• BWA img: .img file from this tool https://gatk.broadinstitute.org/hc/en-us/articles/360040095652-BwaMemIndexImageCreator
• Panel of normal
• Germline resource file
• Contamination VCF For more information, please refer to this tutorial.

Important Points

• One important point for this new approach is the lack of germline resource file for mouse (mm10) genomes, meaning only FilterMutectCalls should be run for mouse genomes. This is required for somaticseq implementation because the vcf headers are changed by this tool.
• Realignment should also NOT BE PERFORMED for gatk4 mutect2. The realignment step has been internalized as described by authors. So for example in PostAlignment stage, AmpliconGatkRealign or AbraRealign should not be applied to mutect2 samples.

[syansanofi]

Mutect2 template

https://github.com/epam/fonda/blob/4a651caa0ab4bdb4ff92516d2294331c9723f134/src/main/resources/templates/mutect2_template.txt#L2

This line should probably be changed to the following:

• I did not include flags
• Mutect2 can/should handle tumor only calling now so that instead of haplotypecaller, mutect2 is run with only a tumor sample, ie controlBam can be optional
• If controlBam is used as normal, then the normal sample name should be provided (the sample name within bam, usually the file name without .bam file extension

gatk --java-options $javaOptions Mutect2 -I $controlBam -I $bam -normal $controlSampleName -R $refGenome -L $bed -pon $panelOfNormal --germline-resource $germlineResource --bam-output $bamout --f1r2-tar-gz ${f1r2.tar.gz} -O $vcf

Refs https://gatk.broadinstitute.org/hc/en-us/articles/360037593851-Mutect2

[syansanofi]

PileupSummaries

New template for PileupSummaries tool. This tool should produce pileup tables used in CalculateContamination tool.

gatk --java-options $javaOptions GetPileupSummaries -I $bam -L $bed -R $refGenome -V $contaminationVCF --sequence-dictionary $sequenceDictionary --interval-set-rule INTERSECTION -O $tumorPileupTable

[syansanofi]

CalculateContamination

New template for CalculateContamination tool. This tool is a spiritual replacement for ContEst in older gatk versions.

• --tumor-segmentation was not included in early gatk4 versions but has reappeared, some documentation could have this discrepancy.

gatk --java-options $javaOptions CalculateContamination -I $tumorPileupTable -L $bed --tumor-segmentation $segmentsTable -O $contaminationTable

[syansanofi]

SortSam for bamout

SortSam applied to bamout of mutect2 for later stages.

• Probably can re use picard version of this tool but have not tested using that version. Sort order here should be coordinate

gatk --java-options $javaOptions SortSam -I $bamout -O $bamoutSorted -SO coordinate

[syansanofi]

LearnReadOrientationModel

New template for LearnReadOrientationModel.

• Uses prior artifacts to look at sequencing bias. ${f1r2.tar.gz}

gatk --java-options $javaOptions LearnReadOrientationModel -I ${f1r2.tar.gz} -O ${artifactsPriors.tar.gz}

[syansanofi]

FilterMutectCalls

New template for FilterMutectCalls.

• Only required field should be -R -V -O
• Optional fields (for mm10 genome that skip previous steps)
  • --contamination-table
  • --tumor-segmentation
  • --ob-priors
  • -stats
  • --filtering-stats

gatk --java-options $javaOptions FilterMutectCalls -R $refGenome -V $vcf -O $filteredVCF --contamination-table $contaminationTable --tumor-segmentation $segmentsTable --ob-priors ${artifactsPriors.tar.gz} -stats $vcfStats --filtering-stats $filteringStats

[syansanofi]

FilterAlignmentArtifacts

New template for FilterAlignmentArtifacts. Further filtering after FilterMutectCalls.

gatk --java-options $javaOptions FilterAlignmentArtifacts -I $bamoutSorted -V $filteredVCF -O $filteredVCFartifacts -R $refGenome --bwa-mem-index-image $bwaImage`