Hi everyone,

I want to do differential exon usage analysis with DEXseq. I have 3 affected (2 technical replicates for each individual, so 6 samples total) and 3 unaffected individuals  (also 2 technical reps for each sample, 6 samples overall). My question is how should I take into account technical replicates in this experiment with DEXseq? Should I add it in the formula somehow? And how my metadata table should look like in this case?

Thank you!

Hi g.e.novakowsky, 

The DEXSeq model considers two sources of variation, (1) biological, which is estimated from biological replicates and (2) technical, which is small compared to biological variability, follows a Poisson distribution and can be estimated directly from the mean (see for example: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2527709/). It is therefore recommended to collapse technical replicates in the DEXSeq analysis: this will give you more sequencing depth for each biological sample that translates in more power in the tests.

Alejandro