Hello, I've several experiments (all the same but from different laboratories) on Affy chips. The intensities are quite different, and an anova using the experiment as a factor (in addition to a factor) shows that >90% of the genes on the chip are different between the experiments (after normalization). A collegue suggested to calculate ratios for each experiment (within each experiment), using the treated replicates (3 reps) per experiment versus the control per experiment (the average of the 3 control reps). This results in 3 ratio replicates per experiment and dose. This way one would get rid of the grand differences between the experiments. Then run an anova on the ratios. Has anybody tried this with affy experiments? I'd be interested in your experience in using a ratio anova. thanks + kind regards, Arne -- Arne Muller, Ph.D. Toxicogenomics, Aventis Pharma arne dot muller domain=aventis com