single-end and paired-end RNA seq

[Ammar Naqvi]

Hi,

Does rMATs require the libraries from the two conditions (ctrl vs case) to be of the same library type (either single-end or paired-end)? In other words could you run it when one of the libraries is single-ended, while the other is paired-end? 

thanks,

ammar

[Thomas Danhorn]

rMATS will certainly not complain if you try to do that. I have not
looked or tried if it works from FASTQs, but I am pretty sure if make your
own BAMs with STAR that rMATS would use them.

The better question is, should you? For a scientific experiment to be
meaningful, you have to keep all the variables you can control the same
(except for the one you are testing) and hope that those you cannot
control are not biased toward either of your groups, so that they just
result in random noise. Clearly this is not the case if you use
completely different sequencing technologies (and maybe library protocols,
not to speak of RNA extraction protocols, etc.) for each of your groups.

If you get a significant event, you will not know for sure if it was
"real" or cause by the different protocols. So if you really must do
this, my suggestion is that you use this only for hypothesis generation.
Once you find events that are of interest for your question, you go back
to the lab, design primers and test those alternative isoforms and their
prevalence in your samples with qPCR (or something similar). Only of
those tests check out can you have confidence and have publishable data.

Good luck!

Thomas

> --
> You received this message because you are subscribed to the Google Groups "rMATS User Group" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to rmats-user-gro...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/rmats-user-group/9dab8d3c-41a8-426c-b4eb-1f0cad6b147dn%40googlegroups.com.
>