[BioC] Limma: can't coerce Rgmodel into exprSet

James Wettenhall wettenhall at wehi.edu.au
Fri Jun 4 15:46:42 CEST 2004


On Fri, 4 Jun 2004, michael watson (IAH-C) wrote:
> use the Biobase function esApply on my limma data, and esApply uses
> exprSet objects.  Can I also coerce MA objects into exprSet objects
> through this method?

Mick,

Yes, the convert package offers a coerce method from an MAList 
object to an exprSet.  See the last example in the help for the 
convert package:
library(convert)
?"coerce,MAList,exprSet-method"

Hopefully you are using a recent version of limma, which should 
produce MAList objects, but if you have any really old limma 
.RData files with MA values stored in a list, you can convert
to an MAList object with: 
MAobj <- new("MAList", .Data=oldMA)

I have never used the esApply method so I don't know whether it 
will work on your limma objects.  Having a quick look at the 
help, it seems to use the phenoData slot of the exprSet object, 
which corresponds to the concept of a "targets" file 
in limma.  

So if you've read a targets file into your RG object with
RG$targets <- readTargets("Targets.txt")
then it will automatically become an attribute of your MA object 
when you do 
MA <- normalizeWithinArrays(RG)

and then when you do
eset <- as(MA,"exprSet")

You should be able to see your targets/phenoData with:

pData(phenoData(MAeset))

I just mention this because I saw "phenoData" mentioned in the 
help for esApply, but I've never really used esApply myself, so 
I can't help much further at the moment.

Regards,
James



More information about the Bioconductor mailing list