Hi Andreas, Use the annotation db for your chip (e.g. hgu133aENSEMBL, hgu133aREFSEQ, hgu133aSYMBOL, hgu133aUNIGENE) to get a grouping variable: > map.x <- unlist(as.list(hgu133aUNIGENE)) > head(map.x) 1007_s_at 1053_at 117_at 121_at 1255_g_at 1294_at "Hs.631988" "Hs.647062" "Hs.654614" "Hs.469728" "Hs.92858" "Hs.16695" you can use something like: collapsed.m <- by(mydata, map.x, median) or if you want the mean limma::avereps Regards, Amos Folarin ---------- Forwarded message ---------- From: Mete Civelek <mcivelek@mednet.ucla.edu> To: <bioconductor@r-project.org> Date: Thu, 28 Apr 2011 10:18:09 -0700 Subject: Re: [BioC] collapsing redundant probe sets of Affymetrix Chips Hi Andreas, Try the function avereps in limma. Mete -----Original Message----- From: bioconductor-bounces@r-project.org [mailto:bioconductor-bounces@r-project.org] On Behalf Of Andreas Heider Sent: Thursday, April 28, 2011 2:13 AM To: bioconductor@r-project.org Subject: [BioC] collapsing redundant probe sets of Affymetrix Chips Hi bioconductor mailing list, I got a (another) problem with redundant probesets from Affymetrix chips: For my downstream analysis to work, I need only 1 expression value per gene (eg Entrez Id). So by now I have a table that looks something like this: N probe set gene id expression 1 *12345_at* *123* 1.5 2 123456_at 1234 2 3 *1234567_at* *123* 1.6 4 12345678_at 12345 4 5 *123456789_at* *123* 1.4 By collapsing I mean as a result something like this: N probe set gene id expression 1 *12345_at* *123* *1.5* <= eg median 2 123456_at 1234 2 4 12345678_at 12345 4 I hope one can help me out! I have seen this in several publications, so there must be a way. Thanks in advance, Andreas Heider [[alternative HTML version deleted]] [[alternative HTML version deleted]]