[BioC] GSMList problem

[Sean Davis]

On Sun, Sep 7, 2008 at 7:41 AM, hemant ritturaj
<ritturajhemant at gmail.com> wrote:
> Dear Sean,
>
> Thanks for prompt reply and it solved that issue.
>
> I am sorry to bother u again as i am stuck at second point, i wrote
>
> probesets <- Table(GPLList(gse6901)[[1]])$ID
>> data.matrix <- log2(do.call("cbind", lapply(GSMList(gse6901), function(x)
>> {
> + tab <- Table(x)
> + mymatch <- match(probesets, tab$ID_REF)
> + return(tab$VALUE[mymatch])
> + })))
>
> Error in log(c("27", "59.8", "34.1", "110.5", "130.3", "569.3", "210.5",  :
>   Non-numeric argument to mathematical function
>
> Can u please help me solving this error as I am not able to understad what
> that error means
> I shall be very thankful for ur kind help

Are you using GSEMatrix=TRUE?  In one of your emails, it suggests that
you are trying to make an ExpressionSet from the GSE.  Using
GSEMatrix=TRUE will do that for you, so there is no need to construct
one by hand.  Could you again post the code that you are using to get
the GSE?

Sean

> On Sun, Sep 7, 2008 at 5:06 PM, Sean Davis <sdavis2 at mail.nih.gov> wrote:
>>
>> On Sun, Sep 7, 2008 at 6:55 AM, hemant ritturaj
>> <ritturajhemant at gmail.com> wrote:
>> > Dear All,
>> >
>> > I was trying to retrieve GSE record from NCBI GEO using GEOquery
>> > package.
>> >
>> > The packge was downloaded using
>> >  > gse6901 <- getGEO("GSE6901")
>> >
>> >> show(gse6901)
>> >
>> > $GSE6901_series_matrix.txt.gz
>> > ExpressionSet (storageMode: lockedEnvironment)
>> > assayData: 57381 features, 12 samples
>> >  element names: exprs
>> > phenoData
>> >  sampleNames: GSM159259, GSM159260, ..., GSM159270  (12 total)
>> >  varLabels and varMetadata description:
>> >    title: NA
>> >    geo_accession: NA
>> >    ...: ...
>> >    data_row_count: NA
>> >    (35 total)
>> > featureData
>> >  featureNames: AFFX-BioB-3_at, AFFX-BioB-5_at, ...,
>> > RPTR-Os-XXU09476-1_at
>> > (57381 total)
>> >  fvarLabels and fvarMetadata description:
>> >    ID: NA
>> >    GB_ACC: NA
>> >    ...: ...
>> >    Gene.Ontology.Molecular.Function: NA
>> >    (16 total)
>> >  additional fvarMetadata: Column, Description
>> > experimentData: use 'experimentData(object)'
>> > Annotation: GPL2025
>> >
>> > When i used
>> >
>> >  >gsmplatforms <- lapply(GSMList(gse), function(x) {
>> > + Meta(x)$platform
>> > + })
>> >
>> > It gave me an error
>> >
>> > Error in function (classes, fdef, mtable)  :
>> >  unable to find an inherited method for function "GSMList", for
>> > signature
>> > "list"
>> >
>> > I checked my R and Bioconductor versions which is as follows, and it
>> > seems
>> > fine as they are latest
>> >
>> >
>> >> sessionInfo()
>> > R version 2.7.1 (2008-06-23)
>> > i386-redhat-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  grDevices utils     datasets  methods
>> > [8] base
>> >
>> > other attached packages:
>> > [1] limma_2.14.6   GEOquery_2.4.5 RCurl_0.9-4    Biobase_2.0.1
>> >
>> > I am really not able to understand what is the error and how to solve
>> > it. If
>> > anyone has the solution please let me know. I shall be very thankful
>>
>> Hello, Hemant.  Sorry for the confusion.  Starting with the latest
>> release (bioc 2.2) , getGEO uses a default of GSEMatrix=TRUE (see help
>> page for getGEO).  This results in much faster parsing of GSE
>> information than using the GSEMatrix=FALSE, which was the older
>> behavior.  Using GSEMatrix=TRUE returns a list of ExpressionSet
>> objects.  In your case, to get the first ExpressionSet, you can do:
>>
>> gse6901.eset <- gse6901[[1]]
>> # to get the sample names (GSM accessions)
>> sampleNames(gse6901)
>>
>> Since gse6901.eset is an ExpressionSet, it will generally work with
>> other Bioconductor tools very nicely (no need to do all that manual
>> conversion, etc., that was required without GSEMatrix=TRUE).  Also,
>> there is extensive documentation on using ExpressionSets via Biobase.
>>
>> If you do, for some reason, need the older behavior, simply specifying
>> GSEMatrix=FALSE in the getGEO call will get that behavior back.
>> However, I would suggest moving to the faster, cleaner GSEMatrix
>> parsing if possible.
>>
>> Sean
>
>
>
> --
> Hemant Ritturaj Kushwaha
> PhD Student
> Center for Computational Biology and Bioinformatics
> Jawaharlal Nehru University
> New Delhi-110067
> Phone: 09868801604
>