Hello all, I have a question about setting up blocks for paired samples in limma. In reading the user guide, limma book chapter, & mailing list, I have noticed most of the questions about paired samples are with Affy chips. However, I have a two-color experiment, so I am not sure how to deal with my design. I hope to be able to use the duplicateCorrelation function, as suggested by the list, but am really stumped about the block vector. I have two samples (of my three) in pairs from a developmental time course (2 time points). The third sample is a baseline REF sample that is not paired. There are 4 biological reps for each sample type. My hyb design is below: -----> Ref ----- | ---- <--- | | | | | | ? | ? ExpT1 -------> ExpT2 <------- Slide Cy3 Cy5 ----- --- --- 1 T1a T2a 2 T2b T1b 3 Ref T1c 4 T1d Ref 5 T2c Ref 6 Ref T2d Here the a, b, c, and d designate the pairs at each Time point. So, the problem here is that slides 1 & 2 are each pairs, by themselves. And, slide 3 could be paired with slide 5, based on the T1c & T2c - but the Ref samples they are hybed with are not in pairs. Would this be set up as something like: c(1, 2, 3, 4, 3, 4) ?? Any suggestions here are greatly appreciated! John John Fowler fowlerj at science.oregonstate.edu Assoc. Professor, Botany & Plant Pathology, Oregon State University