Dear limma users, I am trying to analyze a 2x2 factorial experiment with Agilent two- color arrays (with direct design). I noticed (from the user?s guide and previous posts) that this is not so straight forward as with single-color experiments or ones with reference designs. My experiment is as follows: 2 strains (F and FC), 2 treatments (Con and Cd), hence 4 RNA sources (F_Cd, F_Con, FC_Cd, and FC_Con). If I use one of the RNA sources as ref, my design looks like this: > design<-modelMatrix(targets, ref = "F_Con") Found unique target names: F_Cd F_Con FC_Cd FC_Con > design F_Cd FC_Cd FC_Con 1 1 0 0 2 -1 0 1 3 0 1 -1 4 0 -1 0 5 0 0 1 6 0 0 -1 7 -1 1 0 8 1 -1 0 It is a little bit comparable with the Weaver mutant data example in the user?s guide (section 11.5); accept for the ref pool of course. I want to make all possible contrasts, but now it gets tricky. > cor<-duplicateCorrelation(MALoesssorted,design, ndups=3, spacing=1) Loading required package: statmod > cor$consensus.correlation [1] 0.9172524 > > fit<-lmFit(MALoesssorted,design,ndups=3,correlation=cor$consensus) > > contrast.matrix <- + cbind("F_Con-FC_Con"=c(0,0,-1),"F_Cd-FC_Cd"=c(1,-1,0), + "F_Cd-F_Con"=c(1,0,0), "FC_Cd-FC_Con"=c(0,1,-1), "Int"=c(1,-1,1) ) > > contrast.matrix F_Con-FC_Con F_Cd-FC_Cd F_Cd-F_Con FC_Cd-FC_Con Int [1,] 0 1 1 0 1 [2,] 0 -1 0 1 -1 [3,] -1 0 0 -1 1 My questions are, is this the correct way to analyze this experiment? i.e. is it ok to use one of the RNA sources as a reference for fitting the linear model? And is the contrast.matrix correctly defined, especially the interaction seems a bit strange to me. I hope someone can tell me if this is correct, and if not please explain me how to do this the proper way. Many thanks in advance! Yours, Ben Benjamin Nota Vrije Universiteit Department of Animal Ecology De Boelelaan 1085 1081 HV AMSTERDAM, The Netherlands Tel: +31 (0)20-5987217 Fax: +31 (0)20-5987123