Tumor normal microarray data analysis

Question

[vinaycm911]

How to do comparision for tumors vs normal from different patients without taking average expression values using limma. How to find differentially expressed genes from 10 tumors and 5 normal from different patients?

[Devon Ryan]

Err, it'd be helpful if you clarified your experiment. Are the tumor/normal samples from the same patients (e.g., 5 patients each with 2 tumour biopsies and one normal tissue sample) or are these from 15 different patients? What is your objection to calculating group differences?

[vinaycm911]

All 15 samples from different patients. Individual patient with one tumor.

[Devon Ryan]

Then you need to use an average, since the results are otherwise meaningless.