I am just starting to learn to use bioinformatics tools. My university has a limited and expensive bioinformatics team, so I'm mostly on my own except for big questions.

I am planning to use GATK to run 58 cancer control/normal pairs of Exome sequencing data (Illumina) from FASTQ or BAM file format, through the pipeline, with an output of a VCF & MAF format for analysis.

The current GATK pipeline is used for disease but not cancer, so I was wondering if anyone knew if there should be changes made for cancer. Here's the current pipeline starting with BAM files:

  • (Non-GATK) Picard Mark Duplicates or Samtools roundup
  • Indel Realignment (Realigner TargetCreator + Indel Realigner)
  • Base Quality Score Reacalibration (Base Recalibrator + PrintReads)
  • HaplotypeCaller
  • VQSR (VariantRecalibrator and ApplyRecalibrator in SNP and INDEL mode)
  • Annotation using Oncotator (?)

I'd like some verification that this pipeline will output what I need to run my samples on MuTect, MutSig, or some other analysis program. I appreciate any advice.

  • 1
    $\begingroup$ I typically deal with the sequence files upstream of this step, or the VCF files downstream--otherwise I would be more comfortable attempting to provide some helpful feedback. At an extremely superficial level, what you have outlined make sense, but I have just never used GATK firsthand (lots of colleagues have and do). What about posting this verbatim on these two sites: SeqAnswers, and BioStars? I have seen lots of NGS-related Q&A posted there (your question is essentially a very detailed bioinformatics protocol). $\endgroup$
    – mdperry
    Jun 14, 2015 at 17:35

2 Answers 2


This post goes over what MuTec requires as input. mark duplicates and indel realignment will probably have to be done on bam file to use it as input. BQSR is optional and does not change the quality too much. HaplotypeCaller is used for germline not somatic variant calling.

If you have followup bioinformatics questions, you might be able to find answers more quickly on biostars or GATK forums

  • $\begingroup$ Thank you! I didn't know that about Haplotype Caller. The team at my school suggested it. I will try biostars or GATK forums. $\endgroup$ Jun 15, 2015 at 18:31

MuTect2 was just released into beta as part of GATK 3.5. It's based on HaplotypeCaller but makes somatic SNV and INDEL calls. You can find more information about MuTect2 on the GATK blog and ask any additional questions on the forum.

As a note: IndelQualityRecalibration is not needed with Mutect2, and there is no VQSR available for somatic calls.

MarkDuplicates -> BQSR -> Mutect2 -> Oncotator is a good basic workflow for somatic variant calling.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .