I have cell type A (untreated vs treated) and cell type B (untreated vs treated), 3 replicates for each:
Within each cell type, to compare the treated samples to their corresponding untreated controls I'm using:
dds <- DESeqDataSet(se, design =~ combined) de <- DESeq(dds) results.A <- results(de, contrast=c("combined","A_treated","A_untreated")) results.B <- results(de, contrast=c("combined","B_treated","B_untreated"))
I think I'm doing this part correctly.
Now I want to know how/if the two cell types respond differently to treatment. I don't think this is right, because it ignores the untreated controls:
results.AB <- results(de, contrast=c("combined","B_treated","A_treated")
I think I need to first normalize (if that's the right term) each treatment to its corresponding control and then compare between cell types. What is the proper way to do this (what design formula and how to extract the results), is it simply this?:
dds <- DESeqDataSet(se, design =~ cell_type + treatment) de <- DESeq(dds) results <- results(de)