transcripts absent in TSV but present in GTF

[defendant602]

@andrewprzh

This issue is relevant to issue #337 , and I found 136 transcripts were present in the *.transcript_models.gtf file, but absent from the *.discovered_transcript_grouped_tpm.tsv file.

The data was from a human sample transcriptome, and I used GRCh38 as reference and IsoQuant version 3.10.0.

I have sent my output files to this Email addr isoquant.rna@gmail.com. Looking forward to your reply, thanks!

[andrewprzh]

Dear @defendant602

Unfortunately, I didn't get any emails. Could it be because of the attachment size?
Could you double check please, maybe upload them to google drive or somewhere else if possible?

Best
Andrey

[andrewprzh]

Yes, the email went to the spam folder of isoquant.rna account and was not forwarded. Will try to reproduce it on my end.

[andrewprzh]

Dear @defendant602

I reproduced the issue and I the issue only relates to "grouped" counts and TPMs. Non-grouped counts are fine.
If the transcript has a 0 count it will not be printed in grouped counts.

The other, and a more major question, is why are these transcripts in the GTF if 0 is reported in counts.
I'll continue the investigation.

[andrewprzh]

I will rework and revamp transcript model construction algorithm along with the quantification strategy. I'll keep you posted.