Question about fromGTF.novelJunction.[AS_Event].txt

[Cjy0219]

Hi! Thanks for your excellent tool!
I have learned that "An event in fromGTF.novelJunction.[AS_Event].txt must have at least one junction which is not annotated in the GTF file, but which is supported by the RNA data". However, I found that not all events in "fromGTF.novelJunction.[AS_Event].txt" have results in "[AS_Event].MATS.JC.txt" or "[AS_Event].MATS.JCEC.txt". I am very confused, since it is supported by RNA data, why is it not detected in the results?
Default values ​​are used for many arguments, and my code is:
python3.6 rmats.py --b1 ${SRP}_case.txt --b2 ${SRP}_control.txt --gtf Mus_musculus.GRCm39.109.gtf -t paired --readLength $length --variable-read-length --nthread 8 --od ${ResultPath}/output/ --tmp ${ResultPath}/tmp_output/
Look forward to your reply. Thank you for your attention to this matter.

[EricKutschera]

rMATS has a filter that removes any events that do not have the minimum required reads to run the statistical test. The statistical test requires at least 1 read for each isoform and at least 1 read for each sample group. If rMATS is run with --statoff then the minimum requirements for the statistical test don't need to be enforced because the test will not be run. If you run with --statoff then all events in the fromGTF file will be in the MATS output file and you can check what the read counts are. Here's the code for the filter: https://github.com/Xinglab/rmats-turbo/blob/v4.2.0/rmats.py#L304

Any event reported in a fromGTF.novelJunction file must have at least 1 read supporting at least 1 junction in the event. However, rMATS may have gotten all of the other exon and junction information from the GTF file. In that case the event won't be able to pass the filter

[Cjy0219]

I compared the results with and without the '--statoff' parameter, and indeed, the difference lies in the requirements of the statistical test. I greatly appreciate your assistance.