[miso-users] eQTL advice requested

[Singh, Larry (NIH/NHGRI) [E]]

Dear Yarden (and miso-users),

I'm interested in using MISO to perform eQTL analyses and I would like some advice on what the best way to proceed is.  I've generated the miso_summary files for all my RNA-seq samples as well as pairwise comparisons and associated bayes factors for all pairs of samples.  The question is what should I use for the expression?  I'm considering using either the the miso_posterior_mean value for each isoform.  For example if the miso_posterior_mean was 0.1,0.2,0.3,0.4, I'd split this into 4 values.  Another possible option is to choose one of the samples (say sample_1) as my reference sample and then use the bayes factor relative to this sample.  Do either of these make sense and is one superior to the other?  Is there a better option?

Thanks very much for any help or advice.

Kind regards,
-Larry.