Hello,
I have RNA-seq data from 4 different conditions. Basically, it's wild type, knockout, wild type under heatshock and knockout under heatshock. I want to find whether there is a significant change between the wild type and the knockout regarding the up/down regulation of genes under heatshock conditions.
So far, I have analyzed the pairwise differential expressions using DESeq2. What I can now of course do, is to look whether the significantly differentially expressed genes upon heatshock in the wild type are also significantly up/down in the knockout. But it's not exactly what I want. Basically what I what to know is whether there's a significant change in the change upon heatshock between wild type and knockout.
I have so far not found a way to do this. One rather naive idea I had is to run DESeq2 with the difference in readcounts upon heatshock rather than the actual readcounts for each gene. Is this a good idea or there any other or better ways to do this?
How did you setup your
design
parameter? Please show this and the colData. One could do this probably with interactions but it depends on how you encoded your groups.I'm using DESeqDataSetFromHTSeqCount. For my pairwise comparisons, I had sampleTables like this:
(this one was for WT non-heat shock vs. knockout non-heat shock. My tables for the other comparisons look analogously). The design parameter is "~group". I'll have a closer look at the interactions function of DESeq2 then.
Have you read the vignette? Did you look at the section on interactions?
Yes, I did look at the vignette. I guess, I misunderstood the interactions section. Thanks for pointing me in that direction.