I have paired sample data set and I'm not too sure how to set up the design matrix in edgeR for differential expression.

I have paired samples from women who have cancer (cancer tissue and benign tissue were obtained), these women were also divided into overweight and not overweight (by BMI >30). I have 6 patient samples for overweight and another 6 for not overweight all of which are paired

The samples are as below:

Patient      BMI>30    Cancer tissue

       1     No         Yes
       1     No         No
       2     No         Yes
       2     No         No
       3     Yes        Yes
       3     Yes        No
       4     Yes        Yes
       4     Yes        No

In the edgeR manual under '3.5 Comparisons both between and within subjects' there is a multilevel design and I think that is the most appropriate. Or that I compare the BMI>30 patients. Then compare the BMI<30 patients.