lnclevel and sample clinical information matched

[XXuxi]

hi rmats time,
After inputting a large number of samples, I have correctly obtained the lnclevel of the clipping event, and I want to consult how to match the value in lnclevel with the sample information. I guess that the sample order of -b1 and -b2 for calculating the input of rmats is the sample order represented by the output value of lnclevel. I want to get your confirmation.

[XXuxi]

Another question is whether the splicing events generated in the results based on the GTF file are the lists obtained in different transcripts with or without that exon or intron in the same gene. I wanted to investigate the differences between transcripts in whether this splicing event occurred or not for analysis. Looking forward to your reply.

[EricKutschera]

Yes the values in IncLevel1 are for the samples in the same order as --b1. Similarly for IncLevel2 and --b2: #263

I'm not sure exactly what your second question is. This post has some details about how events are detected using the GTF and reads: #161 (comment)

[XXuxi]

I apologize for not describing the problem clearly. In fact, when observing *.mats.JC, there is a problem of mismatch between exon sites and GTF.


As shown in the above images, each exon start (ES) site is 1 value larger than that shown by gtf. I found corresponding patterns in other splicing events.

Actually, I used gtf mainly to find out what transcripts each splicing event might occur on, and the correspondence is provided in the gtf file. So I hope you can answer my above doubts. Or do you have a better way of identifying what transcripts the splicing event might be on, and I really need your help.

Finally, I tried to create a new id for each splicing event, like ENSG00000225190.12; SE:chr17:45475726-45475888:45476073-45477900:-. For now I plan to use past0() to create. Perhaps you can provide us with a better way to generate a characteristic id for the rmats results?

[EricKutschera]

The start coordinates in the rmats output files are zero-based and the end coordinates are one-based (like BED file format): #316 (comment)

Looking for transcripts that use the event coordinates in the gtf file makes sense. There is an example script in this post #148

Making an event ID like you did by joining the event coordinates is reasonable