Hi all, I am analyzing single color data from Agilent using limma. The experimental design include a common reference (4 biological replicate), after 1hr treatment( 4 biological replicates) and after 3 hour treatment (4 biological replicates). The adjusted P values are higher than 5% FDR in all the three types of analysis (reference vs Ihr treatment, reference vs 3hr treatment and 1hr treatment vs 3hr treatment) I this scenario I would like to do gene filtering, but what should be the pipeline for filtering? whether I should perform filtering in raw data or after normalization. And I am just wondering whether I could get a better FDR values without losing biologically important genes even after filtering? Appreciate the guidanc ein advance. -- Regards, Abhilash [[alternative HTML version deleted]]