Germline variants in Cancer

Question

[prs]

I need to design a workflow to identify pathogenic germline variants from tumor-only or tumor-normal sample data.Please suggest a workflow design or any specific tools that can be used for this purpose.

[Ram]

pathogenic

You need ClinVar, SIFT and PolyPhen2 annotations.

germline variants

Use MuTect2 with the option --genotype-germline-sites to call variants. If necessary, use --genotype-pon-sites as well.

[prs]

And I think I should add a step to filter out the common polymorphism before proceeding to find the pathogenic variants. Also, is there any database which has listed pathogenic variants for different cancer type?

[Ram]

Look at COSMIC - that's also a good source to annotate with. ClinVar is good for germline. Post gnomAD annotation, I use gnomAD non-cancer allele frequencies to filter variants.

[prs]

So, if I am developing a pipeline only to find pathogenic germline variants, this is how I should start --> while variant calling, I can give known variant sites from dbSNP/gnomAD -->VCF --> removing common polymorphic site from VCF (using gnomAD non cancer alleles) --> Annotation of remaining variants using clinvar. Please let me know if this is a good approach or if I need to improve at any step.

[Ram]

removing common polymorphic site from VCF

I'd retain those sites and just add a flag. That way, you can filter them later if you need to. The pipeline looks fine otherwise.

[prs]

Okay, thanks!