Hello,

I have a dataset of microarray gene expression (1900 genes) from about 400 samples divided into 3 groups, and the data was normalized using scaled robust sigmoid. My question is regarding the fold change criteria which should be used for differential expression analysis. Using the common criteria of 1.5 or 2 for fold change in DEA gives me zero significant genes, which is kind of biologically unreasonable for the samples I have. Is it acceptable to go below 0.5? since normalization had a huge impact on the raw data and the range of normalized data was between 0 and 1?

Please advise!

Thanks