I have a RNA seq data set with an experiment design as below (Sorry for the ugly table I am doing this on my phone):
Cell_line Treatment
A DMSO
A X
A Y
A X_Y
B DMSO
B X
B Y
B X_Y
There are two cell lines (A and B) with two drugs (X and Y) and DMSO control. My goal is to look at how the genes expression changes within the cell line, between control and each of every drug treatment, and the combo vs single drug.
I am going to do this in DESeq2. I am just wondering if the interaction between X and Y (if it's present) would affect the structure of the design matrix.
My first thought would be a straightforward comparison in the form of either ~ Cell_line: Treatment
or creating new groups with for each condition (e.g. A_DMSO, B_X, etc.) and just do ~ group
. Will I be missing some information by doing that? Is it more appropriate to split the treatment column into two for X and Y?
Also, I have seen some posts suggesting the use of ~ Cell_line + Cell_line:Treatment
. Does it make a difference if I only want to look at changes within the cell line? And is there a rationale for not including Treatment as an individual block?
Thanks a lot!