Comparison of multiple groups in rMATS

[mamarkevi]

Good day!
Could you please specify if there is any pipeline where rMATS could be used for comparison of several groups with one condition? I'm a new in AS analysis and have the issue to compare AS events within several brain regions. As i understood i could obtain PSI for these samples separately, but without p-value and dPSI. Can i somehow to use rMATS for fully results?

[EricKutschera]

If you have multiple groups that you want to compare you can follow: https://github.com/Xinglab/rmats-turbo/tree/v4.3.0?tab=readme-ov-file#running-the-statistical-model-separately

By combining all of your samples together in --task post all the alignments are used to detect a single set of splicing events. Then you can specify any two groups of samples to run the statistical analysis

Here's a similar post: #96 (comment)

[mamarkevi]

Thank you, very much for your explanation. Could you please also suggest how better to obtain region-specific events, including dPSI? I mean comparison 1 vs others, 2 vs others with considering same lists events.

[EricKutschera]

After you run all the samples in a post step you'll have the read counts for all samples for the event list. After that you can choose any two subsets of samples to run --task stat which will output p-value and delta PSI. The README section linked above describes running pairwise comparisons (1 to 2, 1 to 3, 2 to 3), but you can run 1 vs others, 2 vs others, ...

[mamarkevi]

Thank you!