A question about calculating the adjusted p-value

[RacconC]

Hello,

As I understand it, the calculation of adjusted p-value in rmats is done for each type of alternative splicing such as ES or IR. I'm wondering why the adjusted p-value is calculated separately for each gene in SUPPA, and I would like to understand what the advantages of this approach are.

Thank you for your help.

[EduEyras]

Hi,

Yes, multiple test correction (MTC) adjust the p-value for the fact that your tests are not completely independent of each other.

rMATS corrects per event type because you have the option to analyse only one type of event, but you could also do the MTC across all events.

In SUPPA you can operate the same way, and do the MTC per event or all together, by uploading all p-values into R and applying Benjamini-Hochberg approach or similar.

We decided to include a per-gene correction, which would be less strict than the transcriptome-wide correction but still accounts for the fact that all the events in the same gene will be using data from the same or similar set of transcripts and compared with a similar set of other transcripts with similar expression levels.

I hope this helps

E.

On Mon, 28 Aug 2023 at 23:38, RacconC @.***> wrote:

Hello,

As I understand it, the calculation of adjusted p-value in rmats is done for each type of alternative splicing such as ES or IR. I'm wondering why the adjusted p-value is calculated separately for each gene in SUPPA, and I would like to understand what the advantages of this approach are.

Thank you for your help.

— Reply to this email directly, view it on GitHub https://github.com/comprna/SUPPA/issues/170, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADCZKB53ZHCOVGKWJBBVU2DXXSNMZANCNFSM6AAAAAA4BOCS4M . You are receiving this because you are subscribed to this thread.Message ID: @.***>