[BioC] gcrma

Zhijin Wu zwu at jhsph.edu
Mon Sep 15 10:05:00 MEST 2003


HI, there is a bug in matching the probe sequence and the intensities from
the pm function for certain types of chips. We will release the updated
version soon. For now, please source the attached .r file after you load
library(gcrma) for a quick fix.  

Jean
On Mon, 15 Sep 2003, Michael McDonald wrote:

> Hello,
> Does anyone know why gcrma wouldn't work with the moe chips when it 
> works for the mgu74av2 chipset? Any input is appreciated.  the output 
> is:
> WORKS!
>  > gcrmadata<-gcrma(gcdata)
> Loading required package: mgu74av2probe
> Loading required package: matchprobes
> background correction: gcrma
> normalization: quantiles
> PM/MM correction : pmonly
> expression values: medianpolish
> background correcting...done.
> normalizing...done.
> 12488 ids to be processed
> .........
> 
> 
> ????
>  > gcrmadata<-gcrma(data)
> Loading required package: moe430bprobe
> Loading required package: matchprobes
> background correction: gcrma
> normalization: quantiles
> PM/MM correction : pmonly
> expression values: medianpolish
> background correcting...done.
> normalizing...Error in as.vector(data) : NA/NaN/Inf in foreign function 
> call (arg 1)
> 
> best,
> michael
> 
> Michael J. McDonald Ph.D.
> Universite de Geneve c/o Duboule Lab
> Department of Zoologie and Biologie Animale
> Quai Ernest-Ansermet 30
> CH-1211 Geneve 4
> phone 41.22.379.67.94
> fax	   41.22.379.67.95
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
> 
-------------- next part --------------
getGroupInfo<- function(object){
  allprobes<-probeNames(object)
  matchaffyName=paste(cleancdfname(object at cdfName,addcdf=FALSE),"probe",sep="")
 if (identical(.find.package(matchaffyName, quiet=TRUE),character(0)))
     stop(paste("Probe sequence information not available. Please download and install the",matchaffyName,"library from:\n http://www.bioconductor.org/data/metaData.html\n See the vignette of matchprobes and gcrma for more details"))

 require(matchaffyName,character.only = TRUE)
  seqData <- get(matchaffyName)
  ##check seqData
  t1=table(allprobes);t2=table(seqData$Probe.Set)
  good.set=names(t1)[which(t2[names(t1)]==t1)]##probesets with complete probe seq info
  good.probe=which(seqData$Probe.Set.Name%in%good.set)
  mm.seq<- complementSeq(seqData$seq[good.probe],start=13,stop=13)
  ATCGmm=basecontent(mm.seq)
  CGadj=ATCGmm[,3:4]
  CGadj[CGadj<=4]=4
  CGadj[CGadj>=8]=8
  Index2=which(allprobes%in%good.set) #probes with seq 
  group2d=as.list(NULL)
  for (C in 4:8) {
    for(G in 4:8) {group2d=c(group2d,list(Index2[which(CGadj[,1]==C & CGadj[,2]==G)]))}}
  c(group2d,list(which(!allprobes%in%good.set)))
}


More information about the Bioconductor mailing list