Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

How to Access COSMIC Signature Decomposition in Individual Solutions #260

Open
hilmklnc opened this issue Oct 24, 2024 · 3 comments
Open
Assignees

Comments

@hilmklnc
Copy link

Hi,

First of all, thank you for developing and maintaining this valuable tool!

I have two questions regarding the use of SigProfilerExtractor:

1. How can I identify which COSMIC signatures are decomposed into individual signatures in the result files?
After running SigProfilerExtractor, I can see the individual signature outputs in the 'All_Solutions' directory (e.g., All_Solutions/SBS96_11_Signatures/ containing SBS96A, SBS96B,..., SBS96K). However, I am unable to determine which COSMIC signatures these individual signatures have been decomposed from. I can only see their COSMIC signatures in the suggested solution, not individual solutions. Is there a way to access this information within the existing output files? If not, how can I use SigProfilerAssignment or another approach to map these individual signatures to COSMIC signatures?

2. Clarification on the 'All_solutions_stat.csv' file:
In this file, the suggested solution is marked with a star (*) and, for example, it shows 12 signatures. However, I noticed there is also a p-value marked as "Most Stab Sigs" for the next solution (13 signatures). Could you explain the significance of this p-value and what it means in the context of the stability of the suggested solution?

Any help or guidance on these issues would be greatly appreciated!

Best regards,
Hilmi

@mdbarnesUCSD
Copy link
Collaborator

Hi @hilmklnc,

  1. How can I identify which COSMIC signatures are decomposed into individual signatures in the result files?

The signatures that are labeled SBS96A, SBS96B, ..., etc. are the de novo signatures. The results from decomposing the de novo signatures into COSMIC signatures can be found in the suggested solution directory (within subdirectory is COSMIC_SBS96_Decomposed_Solution). You can use SigProfilerAssignment's decompose_fit function to decompose your de novo signatures into COSMIC signatures.

  1. Clarification on the 'All_solutions_stat.csv' file

The selection of the signatures is determined using only the average and minimum stabilities. @rvangara or @marcos-diazg could you please share some insight into the p-value? Thanks!

@mdbarnesUCSD mdbarnesUCSD self-assigned this Oct 29, 2024
@hilmklnc
Copy link
Author

Hi,

Thank you for your response. I know that the de novo signatures and that their decomposition into COSMIC signatures can be found in the solution directory within the suggested solution.

However, I am still a bit unclear on how I can directly map each of these de novo signatures (e.g., SBS96A, SBS96B, ...) in the specific signature solution file to their corresponding COSMIC signatures.
For example, I want to specifically obtain COSMIC signatures of SBS_signature_11 result coming from the all solution. For decompose_fit function which input file should I use to decompose de novo signatures of a certain solution file.

Thank you,
Hilmi

@mdbarnesUCSD
Copy link
Collaborator

You will want to use the de novo signatures file from the All_Solutions directory for input to decompose_fit as well as your samples matrix. The de novo signatures file that you use will depend on the rank of the solution that you chose from the All_Solutions directory.

After you run decompose_fit in the results directory there will be a file Decompose_Solution/De_Novo_map_to_COSMIC_SBS96.csv that details for each of the de novo signatures (ie SBS96A) the COSMIC signatures that it is composed of.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants