Estimate tumor contamination in normal sample

darioberdariober Cambridge UKMember

Hello again- I'd like to hear your thoughts about estimating the tumour contamination in a normal sample. (This is useful to know when the matched normal is tissue adjacent to the tumor).

From Estimate cross-sample contamination using GetPileupSummaries and CalculateContamination I see that the combination of getPileupSummaries+CalculateContamination is not appropriate for this task.

However, I see in mutect.pdf, section VII, a proposed strategy for estimating tumor in normal. Do you have any experience/recommendations about it?

Thanks a lot!
Dario

Answers

  • shleeshlee CambridgeMember, Broadie ✭✭✭✭✭

    Hi @dariober,

    about estimating the tumour contamination in a normal sample.
    From Estimate cross-sample contamination using GetPileupSummaries and CalculateContamination I see that the combination of getPileupSummaries+CalculateContamination is not appropriate for this task.

    Currently, the GATK does not offer a strategy towards this. However, there is an external tool, also from the Broad Institute, called deTiN that can estimate tumor contamination in the normal.

  • davidbendavidben BostonMember, Broadie, Dev ✭✭✭

    @dariober Precisely because deTiN already exists, this proposal is a low priority. We'll probably implement later this year.

  • darioberdariober Cambridge UKMember

    Thank you @shlee, @davidben for the pointers. I haven't run deTiN yet but I must say the pipeline described in mutect.pdf seems more straightforward both in the method and in the implementation since deTiN requires quite a few input files which I would have to create (e.g. the CNV calls). Maybe I'll try implementing the "mutect.pdf"-strategy and see how it goes...

Sign In or Register to comment.