Agilent arrays

As far as I can tell, most people who use limma to evaluate Agilent arrays use the MedianSignal. This is also the default in the read.maimages function. However, Agilent uses sophisticated methods for signal deconvolution and background correction, which results in values that are available in the column called ProcessedSignal (gProcessedSignal for the green channel). They can be loaded with the following line (for single channel arrays)

rg <- read.maimages( t, path= "rawdata/", 
   source= "agilent", green.only= T, 
   columns= list( G= "gProcessedSignal", Gb= "gProcessedBackground" ) )

See also


Just discovered that the limma package is even better than I thought (and I thought it was awesome). It implements GSEA-style functions roast, romer and camera. They can be used to directly assess gene enrichment for a particular gene set without the need to set arbitrary thresholds to define a list of differentially expressed genes.