Hello everybody, Sorry to ask this kind of questions, but just to be sure. I am just confused on how to calculate the expresion value of a gene in an Affymetrix chip, as we know we have perfect match and Mismatch probs, so after doing notmalization using mas5 or rma, shall we just take the mean between the probs that map to the same gene? I am not interested in finding the diffrently expressed genes, but I want to the see the behaviour of some genes between diffrent samples. I run the classical analysis code

library(affy)
data<-ReadAffy()
eset<-mas5(data) # or eset<-rma(data)
data.expr<-exprs(eset)
dataExpr<-log(data.expr,2)
ID <- featureNames(eset)
Symbol <- getSYMBOL(ID, "hgu133a.db")
pos<-match(rownames(dataExpr),names(Symbol))
rownames(dataExpr)<-as.character(Symbol)[pos]
Notna<-which(!is.na(rownames(dataExpr)))
dataExpr.filter<-dataExpr[Notna,]

So shall I just do :

exprDF<-as.data.frame(dataExpr.filter)
exprDF$gene<-rownames(dataExpr.filter)
library(reshape)
finalExpr<-ddply(exprDF,.(gene),summarize,Sample1=mean(Sample1),Sample2=mean(Sample2))

Thanks