[BioC] gcrma background

[Zhijin Wu]

The fast transformation uses an ad hoc procedure.  When the intensities
are very close to what appear to be background level, they are adjusted
towards a small constant. So it is possible for some probes to get a
bigger number after the transformation.

On Fri, 10 Sep 2004, Ariel Chernomoretz wrote:

> Hi all,
> I just started to work with gcrma and  I have a question regarding the 
> background substracted values obtained with 'gcrma.engine'. If I 
> compare the corrected pm values against the original ones I found that 
> some of them (~20%) get bigger values after the correction . I did not 
> expected that. I guess I am doing something wrong here but I do not 
> realize what. Any help would be appreciate
> 
> 
> I compare:   pm.Data.gcrmabkg[,1] vs. pm(Data)[,1]
> using:
> pm.Data.gcrmabkg<-myGCRMA.background(Data,stype="affinities")
> myGCRMA.background<-function(ddata,affin=NULL,
> 			gOptical = TRUE,
> 			sType    = "fullmodel",
> 			bFast    = TRUE,
> 			fK       = 6*bFast+0.5*(1-bFast),
> 			fStretch = 1.15*bFast+1*(1-bFast),
> 			bVerbose = TRUE,
> 			fRho     = 0.7)
> {
> 
>   require(affy,quietly=TRUE)
>   require(gcrma,quietly=TRUE)
> 
>   if(class(affin)=="NULL"){
>    affin<-compute.affinities(annotation(ddata))
>    affin.new<<-affin
>   }
>   pm.affinities <- pm(affin)
>   mm.affinities <- mm(affin)
> 
>   if(gOptical){
>    ddata = bg.adjust.optical(ddata)
>   }
> 
>   pms<-pm(ddata)
>   mms<-mm(ddata)
> 
>   ppm <- gcrma.engine(pms=pms,
>                      mms=mms,
> 		    pm.affinities=pm.affinities,
> 		    mm.affinities=mm.affinities,
> 		    type=sType,
> 		    k=fK,
> 		    stretch=fStretch,
> 		    correction=1,
> 		    rho=fRho,
> 		    verbose=bVerbose,
> 		    fast=bFast)
>   return(ppm)
> }
> 
> 
> 
> 
> 
> 
> Ariel Chernomoretz, Ph.D.
> Centre de recherche du CHUL
> 2705 Blv Laurier, bloc T-367
> Sainte-Foy, Qc
> G1V 4G2
> (418)-525-4444 ext 46339
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
>