interpretation of voila mobilize constitutive.tsv

[d bogdanoff]

Hello, 

Thank for maintaining MAJIQ. I'm currently using Voila's modulize tool to annotate and associate splice junctions in my data with their associated exons. This approach has worked well for alternatively utilized junctions - those included in the individual splicing-type .tsv files that voila modulize outputs.

I'd like to extend this approach to constitutively spliced-in exons, i.e, for each constitutively included junction, get reference_exon and sliced_with exon coordinates. I tried running voila modulize with the options --show-all and --keep-constitutive to generate the constitutive.tsv file. This seems to provide a semblance of what I'm looking for for genes on the negative strand,  however for positively stranded genes the reference_exon_coord and splice_with_coords values do to map to exons and are largely provided in a reverse order with a larger coordinate provided before a smaller coordinate. 

Does this sound like the expected constitutive.tsv output? If so, what's the correct interpretation of these coordinates in relation to the junction_coord value? Let me know if I can provide any more information about the question at hand - and thanks again for this great tool. 

Derek

[San Jewell]

Hello Derek,

After looking over a set of coordinates I can confirm it looks like there is a strand specific issue with the exon coordinates in constitutive.tsv file. I have made a patch to correct the issue, you may re-clone or reinstall from the repository to receive the corrected version. Please let me know if you are able to see the changes as you imagined.

Thank you for reporting this!

-San

[d bogdanoff]

Hi San,

Thanks for the quick response and changes made to voila's code. Looking at the temporary constitutive.tsv files while the new code is running, It looks like the output for constitutive junctions on both strands are now consistent with each other and follow the general pattern of the other viola splicing-type outputs. 

Will let you know if I have any further questions, but for now things look good. Thanks again for the help, 

Derek

[Maoting Chen]

Hi,

I am also interested in the constitutive junctions. There is a situation where the gene can use alternative promoters and all the junctions involved are constitutive junctions, which means one isoform is longer and the other is part of it. In this case, I want to know the read counts mapped to the junctions unique in the longer isoforms. So is there a way to know the read counts for those constitutive junctions?

Thank you,

Maoting

[San Jewell]

Hi Maoting,

I'm sorry for the long reply, I was away on medical leave. Are you asking to export the read counts per junction in the tsv files? If so, you may simply use the switch --show-read-counts while running voila tsv or voila modulize, to achieve output of these extra columns.

Thanks,

-San