Dear mailing list, I try to get a final solution to the following problem: I have two color microarray data (common reference, dye-swap, multi-factor, multi- group) of which I want to extract differentially expressed genes (DEGs) using LIMMA. Those DEGs I want to analyse for the over-representation of GO-Terms using GOstats. I got as far as a list with GO-Terms and their respective P-values and even could construct a nice direct-acyclic graph out of this data. *However, the output I get does not account for the following: Are single genes/transcripts up- or down-regulated? What about the up- and down-regulation of whole over-represented GO- Terms? * *Could I get there using clustering approaches? * I'm sure somebody on the list has already tried such an approach, *finally yielding over-represented GO-Terms, which are either up- or down- regulated.* Thanks in advance, Andreas [[alternative HTML version deleted]]