COSMIC

[aardes]

Hi
I have a question regarding COSMIC integration. Since PCGR relies on Ensembl VEP for variant annotation, and VEP supports COSMIC via its dedicated plugin, I was wondering:

1. Does PCGR currently support COSMIC annotations (either directly or via the reference bundle)?
2. If not, is it technically possible to enable COSMIC annotation by supplying a custom VEP plugin configuration
3. Are there any architectural or licensing considerations that intentionally exclude COSMIC from the reference bundle?

We are exploring adding a separate somatic annotation layer based on COSMIC and would like to understand whether PCGR can accommodate this within its existing framework.

Best regards

[sigven]

Hi @aardes,

Thanks for reaching out and your interest in using PCGR!

You raise important questions - there are several aspects to this. I'll try to answer your questions below:

1. Here, you need to more specifically define what you mean by COSMIC annotations. What is a variant COSMIC annotation in your view? PCGR provides variant links to COSMIC through VEP, that's its only COSMIC "annotation". (I have even asked VEP maintainers regarding the future of these annotations after COSMIC's updated terms of use - although I don't think anything is settled there yet - VEP's COSMIC annotations - terms and conditions Ensembl/ensembl-vep#1875)
2. One possible way to "push" COSMIC annotations (with a scope that is yet to be defined) through the PCGR pipeline is to have those annotations in your input VCF (as INFO tags), and then retain this information in the output files from PCGR. This means that you need to do your own processing with COSMIC (if you have a license and data available) before proceeding with PCGR analysis. However, note that this strategy will give you COSMIC annotations in the TSV/Excel files, not in the HTML report. There is generally no custom plugin option to enable COSMIC annotation, also since the COSMIC annotation is not well defined.
3. Definitely - you are to the point here. PCGR cannot distribute COSMIC data in the public domain (through its reference bundle) due to the current licensing terms. This is a general challenge with developing tools in this landscape; curated resources are increasingly being subject to restrict licensing terms (intellectual property), which prevents them from being bundled and made available in an open-source tool like PCGR.

Hope this clarifies your questions somewhat.

best,
Sigve

[aardes]

@sigven

Thanks a lot for detailed answer.

I am exploring the opportunities to annotate SOMATIC data and I am aware that we can not use COSMIC anymore unless we do have the licence.

Kind regards
Amin