Hello Biostars, I'm not sure about the correct way to go about this, and so I ask for your assistance. I have a control and disease state group and have counted cells across all of them to determine differences. I'm wondering what a limma model/matrix (if there is any) would look like for controlling for disease state stage (as there is no data for disease state stage in controls). Would it be common to fill 0 for disease state stage? I don't have data for stage on controls, because there is no stage associated with healthy individuals. I already know that the disease state stage affects BMI, the thing I'm most interested in. Please find an example sample sheet.
I was planning on having a limma matrix with model.matrix(~pheno + Stage + ID), which might work with 0's in Controls with stage. I just filled in random numbers for Counts for toy data. sample ID is just 'ID'. Any help would be appreciated!
- pheno ID Stage Counts
- Control 1 20
- Control 2 23
- Control 3 24
- Control 4 21
- Control 5 21
- Control 6 23
- Control 7 25
- Control 8 26
- Disease 9 1 13
- Disease 10 1 5
- Disease 11 1 12
- Disease 12 3 12
- Disease 13 3 31
- Disease 14 3 24
- Disease 15 1 1
- Disease 16 3 24
- Disease 17 3 1
- Disease 18 3 24
- Disease 19 2 12
- Disease 20 3 41