JWR_checker.py ERROR:Invalid input

[eltonjrv]

Dear NanoSplicer developers,
Thanks for what seems to be a really useful tool!

I'm encountering an issue right on the beginning of the pipeline, as my bam file doesn't seem to be in a "valid" format as the error message states: ERROR:Invalid input

The following is my cmd line:
$ python /mnt/scratch/fbsev/bioinformatics-tools/NanoSplicer/bin/JWR_checker.py --chrID=NC_048595.1 --genome-loc=427780000-427819000 --output_csv=Q5C7_NC_048595_plcd4_JWRchecker.csv --threads=16 VL_P431_Q5C7concat_vs_CriGriNCBI.bam Q5C7_NC_048595_plcd4_JWRchecker.h5

The input bam file is 29Gb in size and was generated through minimap2 (-ax splice -p 0 -N 40).
There is an indexed bam (VL_P431_Q5C7concat_vs_CriGriNCBI.bam.bai) in the directory where I'm running it.

Could anyone shed a light on this?

Thanks in advance,
Elton

[zeyjenm]

Hello @eltonjrv ,

Glad to hear that you find NanoSplicer useful!
The invalid input may be due to the --output_csv argument. This argument was not designed for users to choose their own file name. If you leave it as just '--output_csv', this should hopefully resolve the issue. The name of the csv file that is output is chosen from the name of the h5 file in the input, so it should still have the same name.

Regards,
Zeynep

[eltonjrv]

Thanks a lot, Zeynep.
It worked out well now and I've reached the end of the 3 executions (JWR_checker.py, JWR_subset.py, and NanoSplicer.py).
However, I'm only getting the probability table output ("output_prob_table.tsv").
No "output_jwr.bed" with corrected splice junctions for each JWR has been created.
I'd really like this bed file to load on IGV.
Any clue on what could be the issue?
I'm running NanoSplicer with all default settings, except by setting -F p, since I have pod5 raw data instead of fast5.
Thanks again,
Looking forward to hearing back from you,
Best,
Elton

[zeyjenm]

Hi @eltonjrv

I've looked into the issue and identified the cause. I'll get back to you in a week after we have discussed the best way to address it. Thank you for letting us know about this problem!

Best regards,
Zeynep

[eltonjrv]

Thanks Zeynep.
Looking forward to hearing back from you then with a solution to this issue.
Best,
Elton

[zeyjenm]

Hello @eltonjrv ,

Are you using DNA data with R10 chemistry? Unfortunately, at the moment, NanoSplicer is only compatible with R9 DNA data. We are working on adding compatibility for R10 chemistries. I will update you once this is complete.

Best regards,
Zeynep

[eltonjrv]

Thanks for this quite useful information, Zeynep.
Yes, our chemistry was indeed R10.
Would you still recommend relying on the content of the "output_prob_table.tsv" file that appeared to be properly generated through our NanoSplicer execution?
Best,
Elton

[youyupei]

Hi @eltonjrv,

Unfortunately, I wouldn’t recommend it, as the squiggle signal from R10 differs significantly from that of R9. I’d be surprised if the probability calculations remained accurate under these conditions. @zeyjenm is currently working on improving compatibility for both cDNA R10 and direct RNA. We are happy to update you about any progress.

[eltonjrv]

Thanks for letting me know, Yupei.
I'll wait for your updates then.
Your efforts on this tool is appreciated!
Best,
Elton