Dear Thanh Le, Please remember to post your questions to the entire Bioconductor list. You can always send a cc to the person you think can answer it, but if you post to the list a) someone else might find the answer sooner (and when I get busy, I sometimes do not answer Bioconductor questions for weeks) b) everyone on the list will see the response You can do the contrasts the way you have set it up here, but there are 2 problems: a) You will need to drop the intercept from the model. b) Without replication no statistical routine including limma can do a statistical test. So, if you have only 2 cel files, there is no point in running limma. Just look at the fold difference. Sincerely Naomi At 06:39 PM 6/4/2008, you wrote: >Dear Professor Naomi Altman, > >I am THANH LE, a student of the Univ. of Colorado. > >Could you please help me on a problem of gene expression analysis. >Assume I have two CEL files from an experiment: A and B where A is >the control and B is the treatment with a drug. I would like to >extract genes which are differentially expressed from this >experiment. I used limma from R software with the following two matices: > >Design matrix: D > >C T >1 0 >0 1 > >Contrasts matrix: > >c<-makeContrasts(C-T,levels=D) > >Could you please tell me if the two matrices are correct to my experiment? > >Thank you for reading and answering my email. > >Best regards, >Thanh Le Naomi S. Altman 814-865-3791 (voice) Associate Professor Dept. of Statistics 814-863-7114 (fax) Penn State University 814-865-1348 (Statistics) University Park, PA 16802-2111 [[alternative HTML version deleted]]