problem viewing protein alignments in tappAS with gff3 created via isoAnnotLite

[kmattioli]

Hello -

Thanks for all your work on tappAS! I am loving the tool as an easy way to visualize isoforms.

I am having one issue that I can't figure out. I have PacBio data that I analyzed via SQANTI3 and created the required gff3 file using the --isoAnnotLite flag. I am noticing that in many cases, when I try to load the protein alignment view, I am getting the error:

[ERROR] Gene protein annotations must have 'protein' as first entry. Entry: ~nt_flushdata

I have gotten the above error in 4/7 genes that I've tried to load so far.

The only potential pattern I can see with these 4/7 in the gff3 file is that the first isoform listed in the transcripts view also has a predicted NMD feature in the gff3.

Would love to know any solutions or workarounds for the time being if this ends up being a bug fix! Thanks!

-kaia

[kmattioli]

Following up: this issue went away when I filtered out all "NMD" features in the gff3. For what it's worth, it also looks like the gff3 includes predicted NMD features that are not predicted by SQANTI3.

Leaving open as it still seems like a potential bug, but there is a workaround!

[psalguerog]

Hello @kmattioli,

I am sorry for the problem you have been having with tappAS and also for the delay in the answer.

It seems to be a bug on IsoAnnotLite. As your annotation comes from SQANTI3, it would be really helpful if you could run the newest version of IsoAnnotLite (2.7) manually and let me know if the error persists.

You can find the link of download on the next webpage: IsoAnnotLite

To run IsoAnnotLite, you have to use the following SQANTI3 output files:

• my_corrected.gtf
• my_classification.txt
• my_junctions.txt

Furthermore, you have to download the same GFF3 annotated file you used in SQANTI3 with the argument -ggf3 annotation.gff3. The argument -novel is highly recommended.

After running IsoAnnotLite, try to create a new project on tappAS with the same input you did. If the problem persists, I would have to ask you to share your SQANTI3 files used in IsoAnnot to debug the tool.

Thank you for contacting us.

Regards, Pedro Salguero

[aarzalluz]

Closing unless OP reopens or users report similar issues.

[aydemiro]

Hi,

I'm having the same issue with a gff created by SQANTI3 v5.1.1. Whenever a gene has at least one transcript annotated with NMD, the protein is not shown in TappAS with: [ERROR] Gene protein annotations must have 'protein' as first entry. Entry: ~nt_flushdata

The gff file is from the QC step.

Best, Ozkan