Hello, I am trying to determine the best DESeq2 design formula for our study, but I'm unsure what would be the most appropriate for what we want to test.

Our study consists of primary cells from 3 Donors that are cultured in 2 conditions across time. For each cell line and each condition, we have several time points. Donors differ in Sex and Age.

\ We want to evaluate the differential expression of the treatment compared to the control, controlling for the cell line and for the time. That is, if I were to do this manually, I'd compare the the Donor 1-Treatment-Time point 1 to Donor 1-Control-Time point 1, Donor 1-Treatment-Time point 2 to Donor 1-Control-Time point 2, and so on, for each time point, for each donor. We also think Sex and Age of the donor will influence the results.

Should the design be: ~ Condition + Donor + Time + Sex + Age + Condition:Time?

Thank you in advance for any help!

In general, to compare a subset of samples to another subset of samples make a new column concatenating the elements you want, and put that in your design. So you'd make a Treatment_Time column. And have a design like Sex + Age + Treatment_Time, and specify wht subgroup to compare to what with contrasts.

Your sexes and ages are naturally nested in your donors, so Donor + Sex + Age will give the "Matrix not full rank" error unless you tweak some things.

You can do more complicated things with a Time + Treatment + Treatment:time design, but you need to look up how to do that, because just throwing that Treatment:Time element in to the design drastically changes what you are doing.