I am interested in answering such questions about fold change and significance of contrast of contrast: (drug at visit2 vs. drug at visit1) vs. (placebo at visit2 vs. placebo at visit1). I have played with interaction term but not sure that’s exactly what I want? Below is how currently I formulate and any advice would be much appreciated. Thank you.

formula = ~visit + treatment + visit:treatment

to get the results reflecting (drug at visit2 vs. drug at visit1) vs. (placebo at visit2 vs. placebo at visit1):

contrast = list(c(maintreatmenteffect, drugatvisit_2))

If you have set visit1 and placebo as your reference levels, then the following contrasts could be used:

the treatment effect for visit 1

 results(dds, contrast = c('treatment', 'drug', 'placebo'))

the treatment effect for visit 2

 results(dds, list(c('treatment_drug_vs_placebo', 'visit2.treatmentdrug')))

the interaction term

This answers the question: "is the effect of treatment different across visits?"

 results(dds, name = 'visit2.treatmentdrug')

You can decide which is more appropriate to take.

There are a lot of useful examples here:

• Analyzing RNA-seq data with DESeq2
• type ?DESeq2::results and scroll to the bottom of the manual page

Kevin