Hello,

I am using DEXSeq to test for differential transcript usage based on two different covariates.

My current full model is:

~ sample + exon + group:exon

where group = <covariate1>_<covariate2>.

The reduced model I use is:

~ sample + exon

This tests for any usage differences across groups, but I would like to test effects of each covariate separately (as well as potentially their interaction), rather than only testing the combined group factor. Is there a recommended way in DEXSeq to do this using contrasts or appropriate full/reduced models within a single fitted object?

Thank you for your guidance!