PSI value for each replicate

[Maoting Chen]

Hi,

I have a question regarding to how to get PSI value per replicates form the same condition. 

I ran the build command with a configuration file clarifying each condition and the replicates belonging to each condition, like condition A = R1, R2, R3, R4. And then I got the output containing .majiq file for each replicate. 

i calculated both group/condition level PSI and individual sample PSI using the .majiq files I got. But I notice in some individual sample PSI files, some LSVs are missing compared to the LSVs in the group/condition level PSI file because those LSVs in those individual samples have too low read counts and are filtered out. 

So how could I get all the individual sample PSI values for all the LSVs identified in the group PSI files? For example, I know if I run voila tsv on the group/condition level PSI files (.voila) and use --show-all-read-count, I can see the read counts for the LSV in all replicates. Is there similar option to show all PSI as well?

Thank you,

Maoting

[San Jewell]

Hi Maoting,

I'm making a few assumptions so please correct me if I'm wrong. First of all in both the majiq build and also the majiq psi run steps, there will be certain filters applied by default as you have seen. I am assuming that you have disabled some portion of these filters so that a less confident PSI will be determined per each output .voila file  in the end. You can use voila view to check a gene of interest with all psi files specified and see that you are capturing in a manner filtered to your criterion.

After this, it sounds like you are asking for a kind of voila tsv mode with an output column per individual psi. This is not currently supported -- voila tsv mode only works on one input quantification file at a time. However, the output data from modulizer, despite being more comprehensive, should contain what you expect if you provide it with all of the input psi.voila files and again disable some of the filters to your needs. You can look at the "junctions.tsv" output file, and group on lsv_id, to get something more similar to tsv output.

Let me know if it helps.

Thanks,

-San