[BioC] Re: gcrma()/rma() error

Michael Seewald mseewald at gmx.de
Fri Jun 18 13:42:22 CEST 2004


It seems, I found the reason:

When justGCRMA() is called in a subdirectory that contains the substring
"GMS_AllLCM-133A_Foo" it does NOT work. When I rename it to
"GMS_AllLCM-U133A_Foo" it runs fine. The "-" followed by a number is
apparently interpreted in a special (numerical?) way when justGCRMA reads the
CEL files from that directory.

All of that applies only to the non-interactive session, it runs fine in the
interactive session.

Best regards,
Michael


On Fri, 18 Jun 2004, Michael Seewald wrote:

>
> Sorry, I have to correct myself. I just noted that the command fails *ONLY* in
> the non-interactive environemnt, e.g. when calling a script with "source". It
> works when I paste the command interactively into the R shell. This means it
> behaves exactly like Jose described..
>
> Best wishes,
> Michael
>
>
> On Fri, 18 Jun 2004, Michael Seewald wrote:
>
> >
> > Dear Bioconductor people, dear Jean and Rafael,
> >
> > I see the same problem as Jose described (see below) with justGCRMA().
> >
> > The following (while located in a directory with CEL files) does not work:
> >
> > > DATA.A.GCRMA<-justGCRMA(phenoData=pd)
> > Computing affinities..Done.
> > Adjusting for optical effect......................Done.
> > Adjusting for non-specific binding.....................Done.
> > Normalizing
> > Calculating Expression
> > Error in just.gcrma(filenames = filenames, phenoData = phenoData, description
> > = description,  :
> >         dimnames applied to non-array
> >
> > Splitting the task into two commands works:
> >
> > > DATA.A.GCRMA<-justGCRMA()
> > Computing affinities..Done.
> > Adjusting for optical effect......................Done.
> > Adjusting for non-specific binding.....................Done.
> > Normalizing
> > Calculating Expression
> > > phenoData(DATA.A.GCRMA) <- pd
> >
> > There seems to be a (non-critical) bug hidden in the code for the phenotypic
> > data. By the way, doing the same with justRMA in one step works fine.
> >
> > Do you need more information for debugging?
> >
> > Regards & best wishes,
> > Michael
> >
> >
> > On Tue, 6 Apr 2004, Jose Duarte wrote:
> >
> > > Hi all
> > >
> > > I've got the file test.R with the following code:
> > >
> > > filenames <- c("1M2W050304.CEL", "2M2W050304.CEL", "3M2W050304.CEL",
> > > "1M1W050304.CEL", "1M3W050304.CEL")
> > > filepath <- "/net/array/WormDystrophy"
> > > samples <- c("1M2W050304.CEL", "2M2W050304.CEL", "3M2W050304.CEL",
> > > "1M1W050304.CEL", "1M3W050304.CEL")
> > > custom <- c("rma", "quantiles", "pmonly", "medianpolish")
> > > log2trans <- TRUE
> > > library(affy)
> > > library(gcrma)
> > > library(vsn)
> > > bgcorrect.methods <- c(bgcorrect.methods, "gcrma")
> > > normalize.AffyBatch.methods <- c(normalize.AffyBatch.methods, "vsn")
> > > express.summary.stat.methods <- c(express.summary.stat.methods, "rlm")
> > > setwd(filepath)
> > > exprset <- rma(ReadAffy(filenames = filenames))
> > >
> > >
> > >
> > > I run it in batch mode from the linux command line with "R --no-save <
> > > test.R". Everything is ok, except when executing the last line. The
> > > rma() runs fine with the usual output:
> > >
> > > 	Background correcting
> > > 	Normalizing
> > > 	Calculating Expression
> > >
> > > But then it comes with the error:
> > >
> > > 	Error in rma(ReadAffy(filenames = filenames)) :
> > >         	dimnames applied to non-array
> > > 	Execution halted
> > >
> > >
> > >
> > > This is REALLY weird because:
> > >
> > > 1) It runs well in an interactive session just by copying and pasting
> > > the code from the test.R file
> > >
> > > 2) It works fine if I change the order in which the file names are
> > > given!!
> > >
> > >
> > > Any ideas what the error means or what might be happening here? (I'm on
> > > RedHat9, R 1.8.1, affy 1.3.28)
> > >
> > > Thanks
> > >
> > > Jose
> > >
> > > _______________________________________________
> > > Bioconductor mailing list
> > > Bioconductor at stat.math.ethz.ch
> > > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
> > >
> >
> >
>
>



More information about the Bioconductor mailing list