Entering edit mode
6 months ago
Megan Noonan ▴ 10
I've tried running rMATS with either fastq or bam files for my experiment (paired reads) and I keep getting zero used reads. I checked my fastq files and the read length is 150 but a lot of my reads are under "NOT_EXPECTED_READ_LENGTH". Can someone help me troubleshoot this? Here is my code for running with bam files:
python rmats.py --b1 /home/data/Megan/AS_analysis/rMATS/ScrVeh.txt --b2 /home/data/Megan/AS_analysis/rMATS/ScrIns.txt --gtf /home/data/Megan/AS_analysis/rMATS/gencode.v41.annotation.gtf -t paired --readLength 150 --nthread 4 --od /home/data/Megan/AS_analysis/rMATS/ScrVeh_ScrIns --tmp /home/data/Megan/AS_analysis/rMATS
read outcome totals across all BAMs USED: 0 NOT_PAIRED: 10463908 NOT_NH_1: 756430308 NOT_EXPECTED_CIGAR: 6469470 NOT_EXPECTED_READ_LENGTH: 540038736 NOT_EXPECTED_STRAND: 0 EXON_NOT_MATCHED_TO_ANNOTATION: 0 JUNCTION_NOT_MATCHED_TO_ANNOTATION: 0 CLIPPED: 137075821 total: 1450478243
Okay, I did that and now here is my output. Definitely have used reads now but am not sure if the other parameters are within a decent range (specifically it seems like exons and junctions not matched are high).
are you sure your reads were aligned to the same reference + GTF as the one used for rmats? if not it could be problematic... either re-align with the same annotation as rmats or trying to build or own rmats reference. or maybe you aligned allowing for novel splice junctions??? anyway you can at least check out your results from the reads that did work and see if they make sense
Good question, I'll have to check. If I start with fastq files this wouldn't matter, would it?
If rmats does the alignment then yes you should be fine