I am curious about how to use se.exprs (R version 1.70, . My understanding is that it can be used with an exprSet to access the model-based standard errors. In particular the code I have used is: #rma.ctl is an exprSet generated by normalizing an AffyBatch object ("control") with rma. My understanding is that this object contains model-based #expression values for each gene and each subject on the chip >rma.ctl<-rma(control) >rma.ctl Expression Set (exprSet) with 22283 genes 8 samples phenoData object with 1 variables and 8 cases varLabels batch: batch > exprs.rma<-exprs(rma.ctl) > dim(exprs.rma) [1] 22283 8 > exprs.rma[1,] C:\\cappola/C1.CEL C:\\cappola/C2.CEL C:\\cappola/C3.CEL C:\\cappola/C4.CEL 6.231048 6.372506 6.115754 6.147328 C:\\cappola/R1.CEL C:\\cappola/R2.CEL C:\\cappola/R3.CEL C:\\cappola/R4.CEL 6.352380 6.283031 6.426585 6.207264 > se.exprs.rma<-se.exprs(rma.ctl) > dim(se.exprs) [1] 22283 8 > se.exprs.rma[1,] [1] NA NA NA NA NA NA NA NA I don't understand why the se.exprs.rma matrix is empty? Thanks in advance, Mary Putt -------------- next part -------------- A non-text attachment was scrubbed... Name: mputt.vcf Type: text/x-vcard Size: 173 bytes Desc: Card for mary putt Url : https://www.stat.math.ethz.ch/pipermail/bioconductor/attachments /20030624/588d52c7/mputt.bin

You are using se.exprs() correctly. The rma() function does not compute standard errors. You can get an adhoc standard error estimate by computing the RMA expression measure using the expresso() function. Thanks, Ben > > se.exprs.rma[1,] > [1] NA NA NA NA NA NA NA NA > > I don't understand why the se.exprs.rma matrix is empty? Thanks in > advance, Mary Putt -- Ben Bolstad <bolstad@stat.berkeley.edu> http://www.stat.berkeley.edu/~bolstad