Dear Sir I am a novice to limma. I meet some questions and can't solve these by myself. So please give me a hand. firstly, I wonder could I use two wt.fun to filter bad spot, for example: myfun1 <- function(x, threshold=50) { + okred <- abs(x[,"F635 Median"]-x[,"F635 Mean"]) < threshold + okgreen <- abs(x[,"F532 Median"]-x[,"F532 Mean"]) < threshold + as.numeric(okgreen & okred) +} myfun2 <- function(x) as.numeric(x$Flags > -99) I think "myfun1" can help me to filter outliers and "myfun2" can help to filter bad spots according to Genepix output file(.gpr file). Could I combind two filter ways to filter spots, and how? Secondly, question about "duplicateCorrelation", there is a example in your limma usersguide to solve all genes (ESTs and controls) are printed more than once on the array, in my microarray, every gene have duplication,the command is "cor <- duplicateCorrelation(MA,design,ndups=2)", but when I do some technical replication, how can I use this function? (when it comes to technical replication, ndups will equal to 1,how to compute correlation within array duplication? And I isolate RNA from bacterial not animal bodys, how to say biological replication?). Finally, how to extract information from the eBayes function, I mean could I get $coef,$lods,$genes,etc simultaneously, as .gal file input order. Thanks your help. Best regards Chen