Few questions about majiq and voila

[chris chris]

Hi, I just had a few questions about majiq and voila.

1. My data is fr-firststrand, so I'm using strandedness=reverse. I've tried using strandedness=None as well, and I'm getting more LSVs. Assuming this is normal, is it safe to assume that the additional LSVs observed from strandedness=None can be ignored?

2. The number of unique gene names I'm seeing from the voila tsv files are different from the output files from "voila modulize". Is there a reason why this is the case? I assumed "voila modulize" would be categorizing the exact same LSVs observed by "voila tsv".

3. How do I save static html files from voila view? I'm not able to find a flag or command for this.

Thanks.

Chris

[Paul Jewell]

For 1 I will need to check the exact definition and get back to you, however if you know that your data is stranded I would generally not trust any results using unstranded config or vice versa.

For 2, Both voila tsv and voila modulize contain many filters. I would assume that if you use --show-all with both runs, then you will get the entire data sets and all gene-ids (assuming all events are not further filtered by other flags, ex, --decomplexify-psi)

For 3, There is a small floppy disk icon in the top right of each gene page that can be used to save that gene as a static file.

Thanks.

[Joseph Aicher]

With respect to strandedness:

It's not a big deal if you have stranded data but run it unstranded. You will get a few more LSVs which are spurious (especially for overlapping anti sense genes), but overall counts/quantifications should be very similar.

It's a waste if you have unstranded data and you run as stranded. Then, you are essentially working with half the read depth. But overall should get similar values of PSI with more uncertainty.

Worst case is stranded data analyzed with wrong strand (forward vs reverse). Then reads will be assigned to wrong genes. This is the only misspecified case I really wouldn't trust. But generally would expect far fewer reads/LSVs in this case.

Best,

Joseph 

--
You received this message because you are subscribed to the Google Groups "majiq_voila" group.
To unsubscribe from this group and stop receiving emails from it, send an email to majiq_voila...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/majiq_voila/d0e3b576-e8b9-4de5-898e-1fdb34fc8214n%40googlegroups.com.

[chris chris]

Thanks a lot for the answers.

Does MAJIQ automatically detect/handle paired-end or single-end data?

Thanks.

Chris

[Joseph Aicher]

Dear Chris,

MAJIQ detects usage of different junctions/retained introns directly from split/unsplit alignments per read. It does not distinguish/use information between paired reads, so it does not need to detect/handle differences between single vs paired-end data.

Best,

Joseph

To view this discussion on the web visit https://groups.google.com/d/msgid/majiq_voila/44e8ff05-333e-4127-b5eb-ec04a13fddc0n%40googlegroups.com.

--

Joseph Aicher

MD/PhD Candidate

Perelman School of Medicine

Genomics and Computational Biology Graduate Group