OK. We have arrays for an experiment that has run over several years. Initial arrays were performed using MG-U74v2 arrays, but we no longer use those data, as we repeated and expanded the experiment later. Currently, we have much data on 430A/B sets: naive mice (3/strain), and two treatment conditions at days 3, 8, and 21 (3/condition/strain). Recently, we moved to the 430 2.0 array, and now have data for days 49 and 77 (3/condition/strain). I should like best to be able to generate a normalization across all of the arrays in order to use LIMMA to evaluate for effects of condition, time, and strain and their interactions with all of the data as a whole. One option would be if there's a way to pseudo-split a 430 2.0 array into 430A and 430B to allow analysis as if they were all 430A/B sets. The other would be a merge, but I know that's come up before and doesn't seem doable. Thoughts, suggestions? Thanks, Anand C. Patel, MD Washington University School of Medicine acpatel at usa.net