gene expression in disease and normal

Question

[rob.costa1234]

I am stuck with a problem that is more of a trouble shooting with basic understanding rather than using any software/ statistical package. I want to show that a signature of 10 genes can be differentially expressed in disease patients tissue samples while such genes are not affected in healthy individuals. This is a simple gene expression affymatrix data and I have 20 replicates (patients with disease) ad 10 healthy individuals.

[GenoMax]

Where did you come up with the number 10 (for genes of interest)?

[rob.costa1234]

these 10 genes are known biomarker genes which have been identified in various previous publications. so now what I would like to show that these gene show expression pattern in disease state but remain unaffected in non-disease state.

[GenoMax]

And you are not able to show them as DE in your dataset? Is there at least a subset of them that are showing up or none do?

Answer 1

[Devon Ryan]

The simplest method would be gene set enrichment with something like GSEA. If these are really biomarkers then you'd presumably like them to be individually differentially expressed as well, but if the heterogeneity is high enough or the controls aren't matched well enough then this could be problematic (of course, they shouldn't be considered biomarkers then...or the array just sucks at measuring them).