calculateContamination step with multi-tumor samples
Can you please provide
a) GATK version used gatk-4.1.4.1
b) Exact GATK commands used gatk CalculateContamination -I getpileupsummaries.table -O calculatecontamination.table -tumor-segmentation segments.table
c) The entire error log if applicable.
Dear GATK team,
I was under impression you need to process each tumor and generate 1) segments and 2) calculate contamination tables.
Now how do you feed them to FilterMutectCalls if you have multiple tumor samples? I was not able to locate the tutorial describing this.
gatk FilterMutectCalls -V unfiltered.vcf \ [--tumor-segmentation segments.table] \ [--contamination-table contamination.table] \ --ob-priors read-orientation-model.tar.gz \ -O filtered.vcf
I would greatly appreciate it if GATK dev members (or anyone else who encountered this before) could guide me how to proceed.
JC
-
jungmin choi Multi-sample mode in Mutect2 is a relatively new feature, so please pardon that out tutorials haven't caught up yet. We do, however, have up-to-date documentation in our github repo: https://github.com/broadinstitute/gatk/blob/master/docs/mutect/mutect.pdf. Your questions are answered on page 4 in the sample command line for FilterMutectCalls.
The short answer, though, is that you must specify `--tumor-segmentation` and `--contamination-table` multiple times, once for each sample you wish to apply contamination filtering to. Usually the best practice would be once per tumor sample, but this is not necessary. Furthermore, it is allowable to have tumor segmentation without the contamination table and vice versa. For example, in joint calling with three tumor samples, the following is valid.
gatk FilterMutectCalls -V unfiltered.vcf \ --tumor-segmentation segments1.table \
--tumor-segmentation segments2.table \ --contamination-table contamination2.table \ --ob-priors read-orientation-model.tar.gz \ -O filtered.vcf -
By the way, this works because the segmentation and contamination tables contain sample information, which FilterMutectCalls recognizes. And thank you for trying out joint calling with Mutect2!
-
Thank you and great to have the latest Mutect2 documentation now.
-
I am a bit perplexed to see the results like below. The order of the samples is tumor1, normal, and tumor2 from the same patient. Is the tool genotyping the tumor automatically heterozygous (0/1) regardless of the allelic fraction (0,0)?
1 26707938 . G A . PASS CONTQ=93;DP=56;ECNT=1;GERMQ=56;MBQ=34,24;MFRL=160,192;MMQ=60,60;MPOS=10;NALOD=0.782;NLOD=5.80;POPAF=6.00;ROQ=64;SEQQ=61;STRANDQ=52;TLOD=12.36 GT:AD:AF:DP:F1R2:F2R1:SB 0/1:23,9:0.288:32:11,4:10,3:17,6,7,2 0/0:20,0:0.045:20:6,0:13,0:12,8,0,0 0/1:0,0:0.500:0:0,0:0,0:0,0,0,0
Thank you!
-
jungmin choi Yes, the GT field is Mutect2 output is not meaningful.
-
Thanks for confirming this. I am also getting errors when I specify -bamout parameter like below. Is this expected as well?
-bamout <pair>/gatk4.1_multisample/<pair>.mutect2.<chunk>.bam
-
I think we might never have tested -bamout in Mutect2's multi-sample mode, so it's quite possible. It's not the highest priority, but it would be nice to fix and I bet it's not that hard. What error and stack trace are you getting?
-
I am no longer reproducing the bamout errors. It must have been a syntax mistake on my side. Please disregard my question.
Thanks!
Please sign in to leave a comment.
8 comments