Ido M. Tamir wrote: > On Thursday 04 October 2007 06:13, Ido M. Tamir wrote: > >> Dear list, >> >> is there anybody with experience in normalising Agilent >> 4x44 single color expression arrays? >> >> Any suggestions on preferred normalization methods, >> or at least reading it into limma in a way that I can >> separate the 4 different conditions (time course) that >> are on each array? >> I have three biological replicates split on 3 arrays and >> each array contains 4 time points. >> >> Maybe I should just pipe it through VSN, but with >> Affymetrix I was quite happy with GCRMA and would >> actually like to apply a probe model based normalization >> method. >> >> thank you very much for your suggestions, >> ido >> > Sorry to follow up on my own e-mail, > but reading it in was simpler than thought after searching > the list archive more carefully. > > Now the question remains, which normalization method > people with experience would suggest for expression > arrays or should I use the "gProcessedSignal" column > and be happy? > You can use the gProcessedSignal column, which will already have a background correction step done, but you will still need to do a between-array normalization of some kind. Quantile normalization seems to be a reasonable candidate. However, you will still want to do some QC by looking at density plots, pairs plots, etc. to make sure that this makes sense for your arrays. Sean