Hello, I am working on a dataset of nine Affymetrix chip hybridization data, consisting in three time-points analized in triplicates. I am using SAM (multiclass) to identify differentially expressed genes and I realized that some genes have been discarded due to one "bad" value out of three, in the replicate. I woukd like to know if there is an alternative analysis able to "tolerate" this situation and label these genes as differentially expressed. Thanks in advance, -- Stefania Pilati, PhD Istituto Agrario San Michele a/Adige Via E. Mach 1 I-38010 San Michele a/Adige (Trento) - ITALY Tel: 0039 0461 615132 Fax: 0039 0461 650956