Too many significant genes?

Question

[ab123]

Hi there,

Quick question: I have run a DE analysis using limma (normalisation with oligo, dataset uses GeneChip Human Genome U133 Plus 2.0 Array, around 151 cases) and I end up with 10,000 (50%) of significant genes at an adjusted p value of 0.05. I've rerun this analysis at least 4 times using slightly different gene filtering processes and also reuploading everything etc. Suffice to say, I double checked the actual analysis was done correctly.

Is this very uncommon? Does anyone maybe know of an article where a similar number of significant genes were found. I can't seem to find much.

Thanks!

Answer 1

[mforde84]

It's extremely unlikely that 10k genes are being actively regulated in the first place, let alone differentially expressed. Something is wrong with your data, normalization, and or analysis pipeline.