[BioC] From eSet to AffyBatch Object

[Martin Morgan]

Hi Boris --

Boris Umylny <umylny at apbri.org> writes:

> Thank you very much for your help!
>
> I wound up using Martin's suggestion for density and boxplots.  As Jenny 
> pointed out it does not make much sense to run image on an ExpressionSet.  
> However, that does require users to utilize AffyRead even if they are 
> processing their data with just.rma.  However, thats not a major obtacle.
>
> I did find a strange error with just.rma when reading gzipped CEL files from 
> GSE10641 (samples GSM268225, GSM268224, GSM268223, GSM268222, GSM268221, 
> GSM268214, GSM268213, GSM268212, GSM268211, GSM268210, GSM268209, GSM268226, 
> GSM268233, GSM268234, GSM268235, GSM268236, GSM268237, GSM268238) with 
> normalize=FALSE:
>
> Error in dimnames(x) <- dn : 'dimnames' applied to non-array
>
> and ExpressionSet was not created.
>
> When I used AffyRead and rma (with normalize=FALSE), it worked fine.

Unfortunately, I don't see this problem when I try with R 2.8.0 or
R-devel. To track it down further it would really help to see the
output of sessionInfo() (after following the instructions to update
packages at http://bioconductor.org/docs/install/ and confirming that
the error still occurs) and the precise command that you used. Here's
mine...

> setwd("~/tmp/GEO")
> f <- c("GSM268225", "GSM268224", "GSM268223", "GSM268222", "GSM268221", "GSM268214", "GSM268213", "GSM268212", "GSM268211", "GSM268210", "GSM268209", "GSM268226", "GSM268233", "GSM268234", "GSM268235", "GSM268236", "GSM268237", "GSM268238")
> f <- paste(f, ".CEL.gz", sep="")
> all(f %in% list.files())
[1] TRUE
> res <- justRMA(filenames=f, normalize=FALSE)
> sessionInfo()
R version 2.8.0 Patched (2008-11-10 r46882)
x86_64-unknown-linux-gnu

locale:
LC_CTYPE=en_US.UTF-8;LC_NUMERIC=C;LC_TIME=en_US.UTF-8;LC_COLLATE=en_US.UTF-8;LC_MONETARY=C;LC_MESSAGES=en_US.UTF-8;LC_PAPER=en_US.UTF-8;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C;LC_MEASUREMENT=en_US.UTF-8;LC_IDENTIFICATION=C

attached base packages:
[1] tools     stats     graphics  utils     datasets
grDevices methods
[8] base

other attached packages:
[1] rat2302cdf_2.3.0 affy_1.20.2      Biobase_2.2.2

loaded via a namespace (and not attached):
[1] affyio_1.10.1        preprocessCore_1.4.0

Martin

> I am using R 2.8.1, Bioconductor 2.3 and including: 
>
> library(affy)
> library(geneplotter)
> library(lattice)
> library(Biobase)
>
>
> Thank you again for your help!
>
>
> Sincerely,
>
>
> Boris Umylny
>
>
>
> On Tuesday 03 March 2009 02:48:13 pm Martin Morgan wrote:
>> Hi Boris --
>>
>> Boris Umylny <umylny at apbri.org> writes:
>> > Hi,
>> >
>> > I am working with affy library from Bioconductor 2.3.
>> >
>> > The library contains functions (hist, image and boxplot) that work
>> > directly on AffyBatch object, but apparently not on ExpressionSet/eSet
>> > object.  After I use rma, expresso or mas5 to do background correction
>> > and normalization I get an eSet object, which appears to be incompatible
>> > with these functions.  If I call hist(exprs(e)), where e is an
>> > ExpressionSet object, I get a plot of the first array only.
>> >
>> > Is there a Bioconductor library (other than geneplotter which apparently
>> > does not) that provides these functions on eSet/ExpressionSet object?
>>
>> Others will probably point to specific packages implementing plotting
>> methods for expression data, but I find myself doing things more
>> directly, typically using the lattice package
>>
>>   library(lattice)
>>   library(Biobase)
>>   data(sample.ExpressionSet)
>>   eset <- sample.ExpressionSet   # easier to type!
>>   df <- data.frame(Expression=as.vector(exprs(eset)),
>>                    Sample=sampleNames(eset)[col(eset)])
>>   densityplot(~log2(Expression)|Sample, df, plot.points=FALSE)
>>
>> or
>>
>>   densityplot(~log2(Expression), group=Sample, df, plot.points=FALSE)
>>
>> > Is there a way to get an AffyBatch object from eSet object to make use of
>> > these functions?
>>
>> An AffyBatch and an ExpressionSet are conceptually different ('raw'
>> vs. preprocessed data) so trying to convert back and forth at will
>> doesn't really make sense.
>>
>> Martin
>>
>> > Thank you in advance for your help.
>> >
>> >
>> > Sincerely,
>> >
>> >
>> > Boris Umylny
>> >
>> > _______________________________________________
>> > Bioconductor mailing list
>> > Bioconductor at stat.math.ethz.ch
>> > https://stat.ethz.ch/mailman/listinfo/bioconductor
>> > Search the archives:
>> > http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>

-- 
Martin Morgan
Computational Biology / Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N.
PO Box 19024 Seattle, WA 98109

Location: Arnold Building M2 B169
Phone: (206) 667-2793