Hi,

I am a beginner in R and have the following problems.

1) I have tried to RMA normalize my microarray data and then to make the translation from Affy IDs to custom CDF (ENSEMBLE ID). I could not however find the ENSEMBLE gene annotation file (.db)(symbol and description). Could you please share from where i could download it? On the brainarray website i could find only the file (HGU133Plus2_Hs_ENSG_desc.txt) with the gene descriptions but not symbols.

2) It is really a stupid question but how do you calculate a mean of log2 values (e.g. treatments versus control)?

Thanks in advance for your time!

The probeset IDs from the MBNI re-mapped arrays are a concatenation of the annotation source ID and _st. So for your data you will have probesets that are something like ENSG0000000001_st, and to get the Ensembl ID you just do

ens <- gsub("_st", "", featureNames(eset))

where 'eset' is the name of the ExpressionSet that resulted from your call to rma().

You can then annotate using the org.Hs.eg.db package, using the select() function

annot <- select(org.Hs.eg.db, ens, c("SYMBOL","GENENAME"), ENSEMBL)

Which will almost surely give you a warning that there are some multiply-mapped IDs. You will have to decide how to handle that, but in general I go with the most naive thing possible

annot <- annot[!duplicated(annot[,1]),]

You calculate the mean of logged data the same as unlogged. The interpretation, however, is different, as the mean of logged data is a geometric mean, whereas the mean of data on the natural scale is an arithmetic mean.