WES or WGS analysis of cancer samples with no matched germline

Question

[Bogdan]

Dear all,

although this question has been asked before (~ 2 years ago), if I may ask it again, as I am looking for some updated workflows, strategies, ideas :

"what would be the acceptable pipeline for processing WES or WGS data from primary tumors that do not have matched germlines sequences ?"

many thanks,

-- bogdan

[WouterDeCoster]

What is the aim of the analysis?

[Bogdan]

Hi Wouter, thanks : we aim to identify the SNV and CNV in the tumor samples (and some of these samples do not have a matched normal sample).

Answer 1

[markus.riester]

MuTect and FreeBayes for SNVs are probably the most used tools for tumor-only analyses. Copy number, if you have some normals, you can try CNVkit, cn.mops (I think) and GATK4beta. If you need purity/ploidy and want to classify variants as germline vs somatic, there isn't much out yet, so we wrote our own tool PureCN (Bioconductor, try the much improved developer version if you like).

[Bogdan]

Dear Marcus, thank a lot for your suggestions. Yes, we have been using MUTECT2 and cn.MOPS for detecting SNV and CNV in germline-tumor pairs. However, when we have only a TUMOR sample (with no GERMLINE sequence), may I ask, how would you run MUTECT2 ?

[markus.riester]

Just run in without the normal BAM. The GATK3 version of MuTect2 doesn't officially support tumor-only analyses, but will do it. For GATK4beta, it's described in the documentation.

[Bogdan]

Thanks Marcus ! thanks, it is working in GATK 3.8.