Calculating variations between two tumor states
Hello:
Given that I have two related tumor samples at different timepoints, I want to see changes between both. I have used mutect2 with one of the samples as "tumor" and the other as "normal", I have done:
Mutect2 -R hg38_no_alt.fa -I tumor1.bam -I tumor2.bam -normal LH00280_124_227TT5LT4_2_CTACCACC --germline-resource af-only-gnomad.hg38.vcf.gz --panel-of-normals 1000g_pon.hg38.vcf.gz --native-pair-hmm-threads 4 --f1r2-tar-gz tumor1.tar.gz -O variants_tumor1_vs_tumor2.vcf.gz
I then do LearnReadOrientationModel, getPileupSummaries and calculateContamination as indicated in the best practices and finally:
FilterMutectCalls -R hg38_no_alt.fa -V variants_tumor1_vs_tumor2.vcf.gz --tumor-segmentation tumor1.segments.table --contamination-table tumor1.contamination.table --ob-priors tumor1.readorientationmodel.tar.gz --stats tumor1.stats --filtering-stats tumor1.stats -O filtered_tumor1.vcf.gz
However, there are a couple of aspects that I want in my results and I am not sure if this is the right approach.
- I want to assume any difference between both tumors is relevant, in this context I don't want to take into account whether a variant is germline or somatic, in the sense that I don't have a healthy sample to determine whether the variant is germline.
- My understanding is that by tumor1 vs. tumor2 (tumor1 being the pseudonormal) I am getting mutations (somatic) present in tumor2 but I am not getting those being present in tumor1 but not in tumor2 which I also want.
- Ideally I would want a VCF file with all the above and the allele frequencies for all samples of all changes (present in tumor1 and not in tumor2 and present in tumor2 but not in tumor1).
Thank you.
-
This is probably not a single command job but may require additional steps for calling, filtering merging and comparing for both samples. This goes beyond our scope of solving GATK related problems and issues.
Regards.
Please sign in to leave a comment.
1 comment