Hi again, Now I've used the new gcrma I've noticed something strange, which is probably an error somewhere, maybe not with gcrma itself. When I computed gcrma with my old R devel 2.0, BioC devel (gcrma 1.1.1, ATH1cdf 1.4.3, ATH1probe 1.01, matchprobes 1.0.9, affy 1.5.2) I got 'sensible' expression estimates. Using the current R2.0, BioC 1.5 (all others above look the same except matchprobes is now 1.0.12 and affy 1.5.8) I get almost all low values - mean ~3 (see graphs). This is the same if I use justGCRMA() or gcrma(). This isn't due to the gcrma.bg.transformation file I requested - gcrma(fast=FALSE) now 'works' (no error) but the values are also low. I can't think where I'm going wrong as I'm not really typing many commands. Affy and matchprobes maybe candidates as they are the only ones with a different version number? Any ideas? Thanks, Matt -------------- next part -------------- A non-text attachment was scrubbed... Name: old.gcrma.png Type: image/png Size: 5969 bytes Desc: old.gcrma.png Url : https://stat.ethz.ch/pipermail/bioconductor/attachments/20041103 /dd53b377/old.gcrma.png -------------- next part -------------- A non-text attachment was scrubbed... Name: new.gcrma.png Type: image/png Size: 5085 bytes Desc: new.gcrma.png Url : https://stat.ethz.ch/pipermail/bioconductor/attachments/20041103 /dd53b377/new.gcrma.png