Entering edit mode
6.5 years ago
ATpoint
82k
May I ask for your suggestions to correct allele frequencies for SNVs from WGS samples. In more detail, tumors and matched normal data were aligned to hg38 with BWA mem, followed by variant calling with VarScan. SNVs were then filtered with the recommended fpfilter and required not to overlap with the latest 1KG variants. As pathology estimates for the tumor samples range between 50 and 70%, the obtained allele frequencies should be corrected for these impurities and copy number. I browsed through the posts out there on this topic and as always, 1 one question, many answers. Could you therefore give your opinions on which tool currently sets the standard for SNV classification?
I don't know of any tool that does this, but I've been in and out of cancer research since 2013 and the field has changed a lot since then with the revelation of tumour clonality, etc. What about doing something cool like modelling the level of impurity based on the frequencies of germline variants detected, and then 're-adjusting' the somatic variant frequencies based on this model? This is a very broad-sweeping statement and I can't give you the exact implementation yet. I'm thinking something along the lines of what sva does when it aims to correct for batch. If you go this route, mention me in the ACKs!