Differential expression comparing two groups normalized to a third group

Question

[jon]

Beginner here - analyzing a RNA-seq dataset in Galaxy for differential expression. I've got a grip on everything except for analyzing the differential expression. I have three experimental groups of RNA from cultured cells:

1. normal cells
2. cells transfected with a negative control oligo
3. cells transfected with an siRNA

Ideally I would like to normalize the expression of groups 2 and 3 to group 1, and then compare the expression of group 2 vs 3.

I can't figure out how to set up this kind of analysis in DESeq2 - is edgeR more appropriate? Set up each group and then set the contrast of interest to be 2 vs 3?

Thanks in advance!

[caggtaagtat]

How exactly do you want to normalized the expression? If the normalization factor would be the same for a gene of group 2 and group 3 it wouldn't make a difference, would it? Would it may be sufficient to do differential gene expression analysis with e.g. DESeq2 between group 1 and group 2 and also between group 1 and group 3 and than compare the differntially expressed genes?

Answer 1

[swbarnes2]

Ideally I would like to normalize the expression of groups 2 and 3 to group 1, and then compare the expression of group 2 vs 3.

I don't think you want to do this. You usually want to normalize all the samples together.

In DESeq2, use the "contrast" option when running "results" to specify what samples you want to compare to what.