Dear all I have a general question. I am working on an experiment where the gene-set of interest is spread out over 4 arrays. Hence I think I have a three-fold multiple testing problem, on gene level, on contrast level (contrasts between conditions) and on array level. If I just apply the same strategy as e.g. method "global" in decideTests, ( appending all p-values from all the contrasts over all 4 arrays and then apply an appropriate p.adjust method), I suppose it will be far too conservative. I would appreciate any comments or experiences, Best Tefina [[alternative HTML version deleted]]