Dear List I'm dealing with a very unbalanced experiment(affymetrix gene 1.0st) with 14 experiments in one biological condition and only 3 experiments in the other condition. The big variance is observed in the most populated condition...so in your opinion what kind of test could be more useful to retrieve a list of differentially expressed genes between the 2 conditions? At the moment i'm using anova but i have several doubts concerning the fixed effect assumed for each gene . This could be overcome using manova....or It better that i use some non parametric tests like sam o wilcoxon? thank you for any tips Best Guido [[alternative HTML version deleted]]