Hi everyone,

I have two assays : fraction and cell (for cell fraction and whole cell experiments, which have been made separately)

and two conditions tested in those two assays: treatment and control.

I don't know if it's relevant here, but I have 3 replicates per conditions, so a total of 12.

I followed this post , because I want to see the DE in the fraction when it's treated compared to the cell, but also compared to control (fraction and cell). We decided to take this design then:

(fraction_treated / cell_treated) / (fraction_control / cell_control)

So my colData have 2 columns : assay and condition

I then used this code (found in the post cited above):

dds <- DESeqDataSetFromMatrix(countData = countData,
                              colData = colData,
                              design = ~ assay + condition + assay:condition)
dds <- DESeq(dds, test="LRT", reduced= ~ assay + condition)
results(dds)

I am not sure if I am doing something wrong, because when I use resultsNames(dds), I obtain this:

#[1] "Intercept"                "assay_Fraction_vs_Cell"         "condition_treatment_vs_control" "assayFraction.conditionTreatment"

If it's the way it should be, does the results will correspond to my design? Or maybe should I use a different contrast?

Given what you want to do, to get that ratio of ratios, your design is correct. I'm not sure you need to do the LRT part, that's usually for when you have a whole bunch of different conditions, and you want to find something that is different between any two of those conditions, without having to call 10 different contrasts.

The "assayFraction.condition.control" contrast is what you want, though you might want to relevel to make sure that the references for treatment and condition are what you want.