Dear All, I have a question concerning the "design" of a design matrix. I don't use Affymetrix arrays and I don't use a common reference. My experiment is supposed to be a 2x2 factorial experiment with 2 phenotypes(control cells & patient cells) and 2 treatments(treatment & control(=no treatment)). I would like to know more about the effect of a) the treatment b) phenotype patient cells c) interaction of phenotype patient cells and the treatment (in other words: what kind of additional effect has my treatment if the phenotype is patient cells) on the gene expression in the cells. I conducted the following experiments: patient cells/control vs. patient cells/treatment 5 independent samples in dye swap and control cells/control vs. control cells/treatment 5 independent samples in dye swap Questions: Is it possible to get the effect (a,b,c) out of the experiments with the help of limma? If not, what kind of additional experiments do you propose ? (maybe some cross swaps like patient cells/control vs. control cells/treatment ?) How do I use model.matrix especially with regard to the dye swap replicates ? (I know there are several examples in the manual and in the mailing list, but none for 2x2 w/o Affy and w/o common ref) Any help would be greatly appreciated, regards, Martin