Hi all,

I am interested in analyzing timecourse data but would like some advice as to the experimental design in DESeq2. The experimental data include 4 ramets of a single sapling (tree) clone collected on Day 0 (before treatment) and 4 treated ramets every day for 14 days. We did not do controls every day due to sequencing expenses, but we do have controls for Day 7 and Day 14. We included spike ins in the library prep and found factors of unwanted variation (W_1) using RUVseq.

Because of the design, I am limited to running: Model 1: LRT where the full model = ~W_1 + time and reduced = ~W_1, where I can pool all my controls across days into Time 0 (individually testing control days yields very few DEGs), and I would miss any treatment:time interaction.

However, other studies repeat their control data for each time point so I could also run: Model 2. LRT where the full model = ~W_1 + Treatment + Time + Treatment:Time and reduced = W_1 + Treatment + Time. Although here I would appreciate any advice on the set up/pooling for the controls.

We are interested in DEGs between the time points and between the treatment groups, and DEG patterns/clusters throughout the time course. What model or combinations of models would you suggest can answer these questions provided my experimental data?

Thanks for any advice!

For discussion of statistical design and interpretation of results I recommend working with a local statistician or someone familiar with linear models in R. I have to save my time on the support site for software related questions.