Perhaps you'll find this paper useful. --Dennis Hi I have one other question is there an article somewhere which explains in a manner a non statistician might understand the difference between an analysis by limma and paired t-tests and SAM. We have now done all three and are trying to decide which analysis to give most weight to (our experiment compared wild type to mutant, increased dosage to wild type and as the mutant was ts a temperature variable) in choosing which genes to verify and follow up on. thanks, Simon