All output files have only headers

[happypiggyzjx]

Hi, I am trying to run rmats from fastq, here is my code:

rmats.py --s1 /cold_data/zhaojiaxin/rmats/lab_sfiles/labsm8/lab6.sm8 --s2 /cold_data/zhaojiaxin/rmats/lab_sfiles/labsd6/lab6.sd6 --gtf /hot_warm_data/wangduo/reference/Homo_sapiens.GRCh38.109.gtf --bi /cold_data/zhaojiaxin/genomeIndex -t paired --readLength 150 --nthread 8 --libType fr-unstranded --od /cold_data/zhaojiaxin/rmatscorrect/lab6 --tmp /cold_data/zhaojiaxin/rmatscorrect/lab6

The run screen doesn't report any errors, everything is displayed correctly, the final folder outputs a full number of files, but each file has only the title. This is very confusing for me as the summary of results even counts the number of each event detected, what is the problem and what do I need to do to fix it?Below is a summary of the results from the run screen:

gtf: 21.97975993156433
There are 62802 distinct gene ID in the gtf file
There are 252890 distinct transcript ID in the gtf file
There are 38782 one-transcript genes in the gtf file
There are 1648370 exons in the gtf file
There are 26259 one-exon transcripts in the gtf file
There are 23459 one-transcript genes with only one exon in the transcript
Average number of transcripts per gene is 4.026783
Average number of exons per transcript is 6.518130
Average number of exons per transcript excluding one-exon tx is 7.157498
Average number of gene per geneGroup is 8.658762
statistic: 0.03295469284057617

read outcome totals across all BAMs
USED: 0
NOT_PAIRED: 0
NOT_NH_1: 152845714
NOT_EXPECTED_CIGAR: 6130770
NOT_EXPECTED_READ_LENGTH: 477327818
NOT_EXPECTED_STRAND: 0
EXON_NOT_MATCHED_TO_ANNOTATION: 0
JUNCTION_NOT_MATCHED_TO_ANNOTATION: 0
CLIPPED: 0
total: 636304302
outcomes by BAM written to: /cold_data/zhaojiaxin/rmatscorrect/lab6/2024-03-13-14_53_31_088102_read_outcomes_by_bam.txt

novel: 595.0189492702484
The splicing graph and candidate read have been saved into /cold_data/zhaojiaxin/rmatscorrect/lab6/2024-03-13-14_53_31_088102_*.rmats
save: 0.0005075931549072266
loadsg: 0.08091044425964355

==========
Done processing each gene from dictionary to compile AS events
Found 58416 exon skipping events
Found 4647 exon MX events
Found 22059 alt SS events
There are 13359 alt 3 SS events and 8700 alt 5 SS events.
Found 9739 RI events

ase: 4.6443188190460205
count: 0.9242393970489502
Processing count files.
Done processing count files.

Here are the results in the output folder:

[happypiggyzjx]

Then I tried adding the parameter "--variable-read-length" alone, "--statoff" alone, and both "-- variable-read-length" "--statoff", all result in only the headers

[EricKutschera]

For the initial run it looks like the issue was:
USED: 0
NOT_EXPECTED_READ_LENGTH: 477327818

Running with --variable-read-length should resolve NOT_EXPECTED_READ_LENGTH. When you ran with --variable-read-length and --statoff did you also run with new --od and --tmp directories? If old --od and --tmp directories are used then rMATS can report errors related to finding old files. If you still get only the header lines in the output files then I would expect the read outcomes section to show something or there to be an error message

[happypiggyzjx]

A million thanks for your reply, please forgive me for not replying in time, I tried 3 ways before I saw your message and also learnt from previous questions that I need brand new folders. Of these 3 ways, this one with the addition of the parameter --variable-read-length solved my current problem.

But now I came to check the results and a new problem appeared, one of the commands has the following results, I checked the files ending with .MATS.JCEC.txt and all of them have only one line of results, but the code shows an overview of the results clearly a lot of results (attached at the end of the article), meanwhile, here is the code I am running: (I made sure to use the new folder as well as the correct parameters)

rmats.py --s1 /cold_data/zhaojiaxin/rmats/lab_sfiles/labsm8/lab29.sm8 --s2 /cold_data/zhaojiaxin/rmats/lab_sfiles/labsd6/lab29.sd6 --gtf /hot_warm_data/wangduo/reference/Homo_sapiens.GRCh38.109.gtf --bi /cold_data/zhaojiaxin/genomeIndex -t paired --readLength 100 --nthread 8 --libType fr-firststrand --od /cold_data/zhaojiaxin/rmatscorrect/lab29 --tmp /cold_data/zhaojiaxin/rmatscorrect/lab29 --variable-read-length

Now I plan to try again, but Still hoping for your professional reply, and again, sorry for bothering you.

read outcome totals across all BAMs
USED: 2973
NOT_PAIRED: 0
NOT_NH_1: 3208
NOT_EXPECTED_CIGAR: 74
NOT_EXPECTED_READ_LENGTH: 0
NOT_EXPECTED_STRAND: 0
EXON_NOT_MATCHED_TO_ANNOTATION: 36
JUNCTION_NOT_MATCHED_TO_ANNOTATION: 7573
CLIPPED: 0
total: 13864
outcomes by BAM written to: /cold_data/zhaojiaxin/rmatscorrect/lab29/2024-03-15-14:01:05_681369_read_outcomes_by_bam.txt

novel: 0.07744216918945312
The splicing graph and candidate read have been saved into /cold_data/zhaojiaxin/rmatscorrect/lab29/2024-03-15-14:01:05_681369_*.rmats
save: 0.0042917728424072266
loadsg: 0.013990402221679688

==========
Done processing each gene from dictionary to compile AS events
Found 58417 exon skipping events
Found 4647 exon MX events
Found 22059 alt SS events
There are 13359 alt 3 SS events and 8700 alt 5 SS events.
Found 9739 RI events

ase: 2.80719256401062
count: 0.9718458652496338
Processing count files.
Done processing count files.

[happypiggyzjx]

Also I have a small question, does adding or not adding the parameter --variable-read-length affect the results? That is to say, do I need to re-run the results of the data that I didn't add the parameter to before, by adding the parameter?

[EricKutschera]

read outcome totals across all BAMs
USED: 2973
NOT_NH_1: 3208
JUNCTION_NOT_MATCHED_TO_ANNOTATION: 7573
total: 13864

It looks like you are using different --s1 and --s2 files from before and those files only have 13864 total alignments. The earlier post with different files had total: 636304302

Unless run with --statoff the final output files are filtered to events with reads supporting each sample group and each isoform. Since not many reads were USED, most of the detected events probably were detected just from the --gtf and those events were filtered because they didn't have supporting reads: #89 (comment)

--variable-read-length disables the filter that requires alignments to match --readLength. Disabling that filter could allow more reads to be used which would affect the results

[happypiggyzjx]

--statoff

Thank you very much for your reply! I've now re-uploaded the data with the parameters as you said, and hopefully I'll reap good results first thing in the morning~