Hello,

I find myself with RNA-seq data where we want to look at differential gene over time. I've read DESeq2 vignette and performed LRT test just as propose. My question is about my data set. I have 6 patients from which I have 3-5 different time-point.

For example : Patient-Time Point 1 - 0 1 - 1 1 - 2 2 - 0 2 - 2 2 - 4 3 - 1 3 - 3 3 - 4

etc ...

I am aware that this is clearly not the best but I still need to analyse these data. I input the time as a factor in my design and I look at a model combining patients and time vs a reduced model with patient only, since I want to know if there change in miRNA in response to time vs any change between patients.

However, is there a way to know how the «missing» values (ex : patient 2- time point 1) are handle in such a case ?

If you have any input about such analysis i will be more than happy to discuss it with you

Best regards

A-A

I don't have any particular advice for how to best analyze this data with missing time points across patients. With plenty of time points per patient, you could do semi-parametric modeling (e.g. splines) but I would not recommend that here because you have data with 3-5 observations per patient, so not enough data in my opinion.