Accessing LSV String(s) via CLI for MAJIQ-SPEL

[Charles-Alexandre Roy]

Hello,

From what I can tell from reading over the documentation, the only way to obtain the LSV strings needed for running MAJIQ-SPEL (on Galaxy or locally) is by manually copying the LSV from a Voila view visualization of an LSV of interest. This is quite impractical if there are many LSVs one would like to process with MAJIQ-SPEL.

Is there a way of programmatically extracting the LSV strings needed for MAJIQ-SPEL from the file(s) Voila uses to create the visualizations? Ideally, one could extract the LSVs of interest with a set of filters (e.g. gene names, target regions, etc) and then use MAJIQ-SPEL to create a series of index.html files for those LSVs that could then be viewed in a web browser.

Currently, the MAJIQ-SPEL README (which incidentally, hasn't been updated to reflect the fact that majiq spel was updated to Python 3 two years ago) mentions that one needs to use an http daemon to view the HTML files output by MAJIQ-SPEL and provides a command like: majiq_spel --debug --httpd --lsv "<very long lsv string>" to accomplish this. Can anyone provide separate commands for A, generating the index.html files and B, view them with an http daemon?

Thanks so much for any and all help!

Best,
Charles

[San Jewell]

Hi Charles,

Thanks for reaching out. There is currently no feature for generating voila view output pages on the command line, as they require a web browser for proper rendering and function. Based on this request I will likely make a new feature in upcoming releases, which optionally allows majiq-spel lsv-text data to be written as a column to the output of the "voila tsv" command.

I believe the reason that this information is not mass generated in general is because our use case for primer design experiments is usually centered around one or a small number of lsvs of interest. Can you describe your use case in a little more depth so that we can better consider which features to add to help the software serve our users better?

Thanks!

-San

[Charles-Alexandre Roy]

Hi San,

Thank you for confirming. I think that having the option to write the lsv-text data as a column would be a useful start.

For our use case, we have an experiment with ~50,000 different LSVs for each of several comparisons and don’t necessarily know which genes or even how many are going to be differentially spliced (according to some p-value cutoff) and would rather not have to have to look at them all.

For such cases, the more filtering and data extraction options available on the command line, the better.

Best,

Charles

[Matthew Gazzara]

Hi Charles,

The main use case for MAJIQ-SPEL is to generate pairs of RT-PCR primers that flank an LSV of interest for experimental validation and generate the expected product lengths amplified by those primers. Typically, you want to manually look at the LSVs you are trying to validate and be aware of the product sizes for when you run them out on a gel, so for that reason SPEL was designed to focus on one LSV at a time. Are you intending to generate that many primers for experimental validations? 

If you are simply running comparisons between conditions and want to know which LSVs are changing, you would want to use voila tsv and/or voila view commands to make a text file that you can parse further to count the number of genes/events that are changing and visualize those genes of interest on your web browser.

We can help you get either of these things working, depending on what your use case is and computational set up, so let us know if this sounds right for your situation. 

All the best,

Matt Gazzara