DEGseq - 3 conditions

Question

[pingu77]

Hello, I am fairly new to DEG analysis.

I have three conditions: A, B, C and I would like to get all the genes that are differentially expressed in A respect to B and C. I am testing DEseq2, should I run it twice? For example: A vs B A vs C

And take the genes that have a significant logFC and pvalue in both, or is there another way for doing this?

Thank you

[Shred]

Add more details about the design. Is something like disease-control-disease+drug? Do you expect that condition B would differ from condition C ?

[pingu77]

thank you for your help! A,B,C are three states of a disease. I would be interested to find the genes that are consistently differentially expressed in A compared to B and C. I expect that not all genes will be the same if I compare A vs B and A vs C but there should be some genes that are DE in both. Should I just take the overlap?

[Shred]

If samples are indipendent each other, it's fine to have a design column (named here condition) with the three factors (A,B,C) that could be used as design=~condition. So, you'll run:

resB_A <- results(dds, contrast=c("condition","B","A"))
resC_A <- results(dds, contrast=c("condition","C","A"))

[pingu77]

thank you! But after I should take the intersection of the genes that are DE in resB_A and resC_A, right? thank you!