Hello,

I am performing a differential gene expression analysis and would appreciate some advice regarding the appropriate design formula.

Our goal is to identify genes that are differentially expressed between two conditions: pre-disease and disease. We have samples from 6 patients who were followed longitudinally, with sequential biopsies/hysterectomies collected over time as disease progression was monitored. Each patient has approximately 3–5 biopsies in the pre-disease state and 3–5 biopsies in the disease state.

A simplified version of the metadata is shown below:

... and similarly for the remaining patients.

My question is whether it would be appropriate to test for differential expression between pre-disease and disease while accounting for patient-to-patient variation and sample type variation (PCA shows the biggest sample separation based on sample type)? In DESeq2, I was considering the design:

design = ~ patient + sample_type + condition

In addition, one patient only has pre-disease samples and no disease samples. Would it be advisable to include this patient in the analysis, or would it be better to exclude them?

Finally, I am wondering whether a single model including all patients is preferable, or whether it would be more appropriate to perform differential expression analysis separately for each patient (pre-disease vs. disease) and then look for overlapping genes across patients.

I would be very grateful for any guidance or suggestions.

Thank you very much for your help!!