I had a question about how limma handles dye swap experiments. The way we typically run these experiments is to label an aliquot of a test set of (amplified) aRNA with Cy3 and another aliquot of the same aRNA with Cy5. We do the same with the control aRNA and hyb opposite dyes on two slides (i.e. slide 1: test Cy5/control Cy 3, slide 2: test Cy3/control Cy 5). In the past we would loess normalize the slides, calculate an average log2 ratio for the set, and use 1 number to represent the experimental M value from the two slides. Q: Is there any consideration in limma given to the fact that these two slides represent the same original RNA sample and therefore (partially) represent a technical replicate? It appears that limma lumps all of the slides together regardless of which slides are supposed to be paired with others. It seems to me that these pairs of slides must be considered independently from other biological dye-swap replicates. Can anyone shed some light on how limma handles this? Thanks in advance. Mike Schaffer Ph.D. Candidate Bioinformatics Program Boston University

Dear Mike, There have been several threads about this, or similar, issues in the BioC list. Two I recall were on april 2004 and august 2003, but I think this issue has come up several other times. You can download the BioC archives and search for "technical replicates" and "dye swap". Best, R. P.S. A more general question, though, is why you use tech. reps. If availability of biological material is not a problem, this might not be the best design for a fixed number of arrays (this issue is discussed in the review papers by Speed and Yang, Churchill, and the book by Simon et al.). On Wednesday 15 September 2004 21:41, Mike Schaffer wrote: > I had a question about how limma handles dye swap experiments. > > The way we typically run these experiments is to label an aliquot of a > test set of (amplified) aRNA with Cy3 and another aliquot of the same > aRNA with Cy5. We do the same with the control aRNA and hyb opposite > dyes on two slides (i.e. slide 1: test Cy5/control Cy 3, slide 2: test > Cy3/control Cy 5). In the past we would loess normalize the slides, > calculate an average log2 ratio for the set, and use 1 number to > represent the experimental M value from the two slides. > > Q: Is there any consideration in limma given to the fact that these two > slides represent the same original RNA sample and therefore (partially) > represent a technical replicate? > > It appears that limma lumps all of the slides together regardless of > which slides are supposed to be paired with others. It seems to me > that these pairs of slides must be considered independently from other > biological dye-swap replicates. Can anyone shed some light on how > limma handles this? > > Thanks in advance. > > Mike Schaffer > Ph.D. Candidate > Bioinformatics Program > Boston University > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor -- Ram?n D?az-Uriarte Bioinformatics Unit Centro Nacional de Investigaciones Oncol?gicas (CNIO) (Spanish National Cancer Center) Melchor Fern?ndez Almagro, 3 28029 Madrid (Spain) Fax: +-34-91-224-6972 Phone: +-34-91-224-6900 http://ligarto.org/rdiaz PGP KeyID: 0xE89B3462 (http://ligarto.org/rdiaz/0xE89B3462.asc)