Skip to content

Conversation

@krassowski
Copy link

Hi @gravitino,

First off, thank you for maintaining this package!

During my MRes project I implemented FDR calculation for cudaGSEA. I was wondering if you are interested in merging my changes in. I have tested the results against the GSEADesktop (from Broad Institute) and it looks quite good (the grey line is the GSEADesktop FDR; cuda_full_fdr - the second facet - is the implementation that I propose in this PR):

cuda_gsea_benchmark

I also implemented an "approximate" FDR procedure (one-sided, please see with_approximate_fdr branch, here is the diff) which is slightly faster but is not a part of this PR. The other branch also comes with a command line option for switching the FDR calculation - I can include it in this PR if you wish.

Please let me know what you think - I may be able to find some time to polish the code if you consider it worth merging.

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

Successfully merging this pull request may close these issues.

1 participant