search

z0on / GO_MWU

Rank-based Gene Ontology analysis of gene expression data
37 stars 17 forks source link

bestGOs - selection criteria #22

Closed mewojciechowski closed 3 months ago

mewojciechowski commented 3 months ago

Hi Misha,

I have another question :-). This one concerns the bestGOs. What is the criterium for selecting and ordering these? I've noticed that they're pretty general, at least for my data.

Thanks again

Marek

z0on commented 3 months ago

HI Marek - This is a rarely used module tbh, I don't think I ever used it myself. But the idea is to reduce the list of your significant GO terms (many of which are significant because they share the same genes) to the list that would only have largely non-overlapping, most significant GOs. We just take the GO tree that you see plotted by the gomwuPlot, cut it at height 0.9, and for each of the resulting groups choose the GO that has the lowest pvalue (if there are many with the same low pvalue, choose the one containing the most "good genes"). Misha

On Wed, May 8, 2024 at 3:52 PM mewojciechowski @.***> wrote:

Hi Misha,

I have another question :-). This one concerns the bestGOs. What is the criterium for selecting and ordering these? I've noticed that they're pretty general, at least for my data.

Thanks again

Marek

— Reply to this email directly, view it on GitHub https://github.com/z0on/GO_MWU/issues/22, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABZUHGCJ2XMJZOP4JPL6SMDZBKF7VAVCNFSM6AAAAABHNVK3Y6VHI2DSMVQWIX3LMV43ASLTON2WKOZSGI4DMNBTHE4TQMA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

mewojciechowski commented 3 months ago

Thanks!

Marek