How to plot PCA on the PSI values using rMATS output?

[Motoufiq]

Hi,

I recently ran the nf-core rnasplice pipeline https://nf-co.re/rnasplice/1.0.4/, rMATS output is generated as part of the this pipeline for instance it looks like this: RNAsplice_pipeline/star/rmats/OY/rmats_post/, and the respective files:

Since there are many files on which file(s) PCA makes sense? For instance, If I am interested in *.JCEC, then, should I pool all the files ending with *.JCEC, and plot PCA on it? Is there a script for plotting pca on rMATS output?

-rw-r--r--  1 TT test  4007431 Nov 20 05:29 A3SS.MATS.JCEC.txt
-rw-r--r--  1 TT test  3972661 Nov 20 05:29 A3SS.MATS.JC.txt
-rw-r--r--  1 TT test  3106603 Nov 20 05:29 A5SS.MATS.JCEC.txt
-rw-r--r--  1 TT test  3075223 Nov 20 05:29 A5SS.MATS.JC.txt
-rw-r--r--  1 TT test  2542064 Nov 20 05:29 fromGTF.A3SS.txt
-rw-r--r--  1 TT test  1926476 Nov 20 05:29 fromGTF.A5SS.txt
-rw-r--r--  1 TT test  4585271 Nov 20 05:29 fromGTF.MXE.txt
-rw-r--r--  1 TT test   677592 Nov 20 05:29 fromGTF.novelJunction.A3SS.txt
-rw-r--r--  1 TT test   558920 Nov 20 05:29 fromGTF.novelJunction.A5SS.txt
-rw-r--r--  1 TT test  3794567 Nov 20 05:29 fromGTF.novelJunction.MXE.txt
-rw-r--r--  1 TT test   181812 Nov 20 05:29 fromGTF.novelJunction.RI.txt
-rw-r--r--  1 TT test  9757588 Nov 20 05:29 fromGTF.novelJunction.SE.txt
-rw-r--r--  1 TT test      102 Nov 20 05:29 fromGTF.novelSpliceSite.A3SS.txt
-rw-r--r--  1 TT test      102 Nov 20 05:29 fromGTF.novelSpliceSite.A5SS.txt
-rw-r--r--  1 TT test      140 Nov 20 05:29 fromGTF.novelSpliceSite.MXE.txt
-rw-r--r--  1 TT test      108 Nov 20 05:29 fromGTF.novelSpliceSite.RI.txt
-rw-r--r--  1 TT test      104 Nov 20 05:29 fromGTF.novelSpliceSite.SE.txt
-rw-r--r--  1 TT test  2450270 Nov 20 05:29 fromGTF.RI.txt
-rw-r--r--  1 TT test 16875667 Nov 20 05:29 fromGTF.SE.txt
-rw-r--r--  1 TT test   863222 Nov 20 05:29 JCEC.raw.input.A3SS.txt
-rw-r--r--  1 TT test   666392 Nov 20 05:29 JCEC.raw.input.A5SS.txt
-rw-r--r--  1 TT test  1686893 Nov 20 05:29 JCEC.raw.input.MXE.txt
-rw-r--r--  1 TT test   960057 Nov 20 05:29 JCEC.raw.input.RI.txt
-rw-r--r--  1 TT test  5688230 Nov 20 05:29 JCEC.raw.input.SE.txt
-rw-r--r--  1 TT test   849991 Nov 20 05:29 JC.raw.input.A3SS.txt
-rw-r--r--  1 TT test   654178 Nov 20 05:29 JC.raw.input.A5SS.txt
-rw-r--r--  1 TT test  1652487 Nov 20 05:29 JC.raw.input.MXE.txt
-rw-r--r--  1 TT test   924464 Nov 20 05:29 JC.raw.input.RI.txt
-rw-r--r--  1 TT test  5620263 Nov 20 05:29 JC.raw.input.SE.txt
-rw-r--r--  1 TT test  8165111 Nov 20 05:29 MXE.MATS.JCEC.txt
-rw-r--r--  1 TT test  8071880 Nov 20 05:29 MXE.MATS.JC.txt
-rw-r--r--  1 TT test  4408945 Nov 20 05:29 RI.MATS.JCEC.txt
-rw-r--r--  1 TT test  4370003 Nov 20 05:29 RI.MATS.JC.txt
-rw-r--r--  1 TT test 26011754 Nov 20 05:29 SE.MATS.JCEC.txt
-rw-r--r--  1 TT test 25839251 Nov 20 05:29 SE.MATS.JC.txt
-rw-r--r--  1 TT test      442 Nov 20 05:29 summary.txt

Thank you,
Toufiq

[EricKutschera]

This post has a script for making a PCA plot from rMATS output: #398 (comment)

It's set to use the JC files but you could change this line to use JCEC:

rmats-turbo/rMATS_P/extract_psi_for_pca.py

Line 128 in 449aa83

suffix = '.MATS.JC.txt'

[Motoufiq]

@EricKutschera Thank you. I will change to include *.JCEC.txt files. For next step, plotting PCA, should we consider all event types or only specific events for instance A3SS or RI etc.,

[EricKutschera]

The script uses all event types (SE, A5SS, A3SS, MXE, RI), but making a separate plot for each event type also sounds reasonable

[Motoufiq]

@EricKutschera Thanks much.