[BioC] normexp

[Wolfgang Huber]

Dear Yolande,

are you sure that you really want to "make up" data in this way? Among
other things, censoring data in such a way messes up the noise
distribution and will distort any inference you might want to make (e.g.
statistical tests for differential expression). A better approach would
be to replace such "0" values with NA (if that is what they really
mean), and look for subsequent normalization and testing methods that
can deal with NA.

I do not know what causes your problem with the "normexp" background
correction. You could also try "vsn" (from the package of the same name
or by using the option in the "normalizeBetweenArrays" function). The
glog transformation in vsn, and the normexp background correction and
subsequent log transformation, attempt to address the same problem,
although the computations and the theoretical motivations are different.

  Best wishes
    Wolfgang


> Hi everyone,
>  
> Please, I need some help for the following.
> I have 11 output files where some of the intensities are 0 (foreground and background). I tried to adjust by writing
>  
> for (j in 1:k) {
> +  for (i in 1: n) {
> + if (RG$R[i,j]<=0) RG$R[i,j]<-150
> + if (RG$G[i,j] <=0) RG$G[i,j]<-200} }
>  
> (150 and 200) are arbitrary values.
>  
> Some of the background intensity values are larger than the foreground intensity so to have positive corrected intensity, I used
> 
>  RG <- backgroundCorrect(RG, method="normexp", offset=100)
> Corrected array 1 
> Corrected array 2 
> Corrected array 3 
> Corrected array 4 
> Corrected array 5 
> Corrected array 6 
> Corrected array 7 
> Corrected array 8 
> Corrected array 9 
> Corrected array 10 
> Corrected array 11 
> Warning message:
> NaNs produced in: log(x) 
>  
> It seems that it stills gets zero value. Why would log(x) appears at this stage? 
> I then normalize to see the print-tip loess plot.
> MA <- normalizeWithinArrays(RG, RG$printer,method="printtiploess",weights=RG$weights, span= 0.5, iterations=10)
> plotPrintTipLoess(MA[,3])
> Attached is the plot. The normalization was not successful.
>  
> Yolande
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> 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


-- 
------------------------------------------------------------------
Wolfgang Huber  EBI/EMBL  Cambridge UK  http://www.ebi.ac.uk/huber