[BioC] vsn + backgroundSubtract

[Kasper Daniel Hansen]

On Jul 4, 2008, at 1:36 AM, DAVID ARTETA GARCIA wrote:

> Dear list, please apologise for not providing enough information.
>
>> sessionInfo()
> R version 2.7.0 (2008-04-22)
> i386-pc-mingw32
>
> attached base packages:
> [1] grid      tools     stats     graphics  grDevices utils      
> datasets  methods   base
>
> other attached packages:
> [1] arrayQuality_1.18.0  RColorBrewer_1.0-2   gridBase_0.4-3        
> hexbin_1.14.0        convert_1.16.0       marray_1.18.0         
> vsn_3.6.0
> [8] limma_2.14.0         lattice_0.17-6       affy_1.18.0           
> preprocessCore_1.2.0 affyio_1.8.0         Biobase_2.0.0
>
>
> I have 12 one-color gpr files. These were loaded using limma´s  
> read.maimages()
>
>> data = read.maimages(filenames,source="genepix")
> Read 0023.gpr
> Read 0024.gpr
> Read 0025.gpr
> .
> .
>
> data
> An object of class "RGList"
>
> I then transformed R and Rg to NULL as suggested on this list (after  
> searching in the archives).
>
> I actually tried vsn normalisation of G values using the most recent  
> vsn2() function ( I guess I could have also used justvsn()). If I am  
> not wrong, vsn2 can take a matrix as argument:
>
> data.norm = vsn2(data2$G)
> vsn: 2352 x 12 matrix (1 stratum).   0% done.
> 14% done.
> 28% done.
> 42% done.
> 57% done.
> 71% done.
> 85% done.
> 100% done.
> Please use 'meanSdPlot' to verify the fit.
>
> but I also wanted to check wether there was an improvement  
> correcting for backgroung effects since I can see a spatial trend on  
> some of the chips. I read on the vsn vignette that the argument  
> "backgroundsubtract" to justvsn might take care of this trend, but  
> of course, the function could not find "R":
>
> data.vsn_b<-justvsn(data2,backgroundsubtract=TRUE)
> Error in eval(expr, envir, enclos) : object "R" not found
> Error in is(object, Cl) :
>  error in evaluating the argument 'x' in selecting a method for  
> function 'vsn2'
>
> Maybe I should leave R and Rg as they are (all 0´s) instead of NULL- 
> ing them?

The background subtraction vsn does is really just (in your case)

vsn2(data$G - data$Gb)

provided that the Gb slot holds the estimated local background  
(whether that is true depends on how you have read the files, I am not  
too familiar with GenePix).

Kasper


> I haven´t tried converting to an ExpressionSet which I thought was  
> mainly for affy data, but how can I handle both the foreground G and  
> background Gb data in the ExpressionSet? I haven´t confronted this  
> matter before.
>
> Thanks,
>
> D.
>
> Quoting Wolfgang Huber <huber at ebi.ac.uk>:
>>
>>
>> Dear David,
>>
>> please provide the output of sessionInfo() and a reproducible  
>> example,
>> otherwise it is difficult to give a concrete reply.
>>
>> Also, please use the function "vsn2" instead of "vsn" (both in the
>> package vsn).
>>
>> From your message it appears (sorry if my guess is wrong - see above)
>> as if you use an RGList for one color data - which, if true, would be
>> weird. Have you tried putting the data into an ExpressionSet and then
>> calling vsn2/justvsn on that?
>>
>> Best wishes
>> 	Wolfgang
>>
>>
>> -- 
>> ----------------------------------------------------
>> Wolfgang Huber, EMBL-EBI, http://www.ebi.ac.uk/huber
>>
>>
>> DAVID ARTETA GARCIA wrote:
>>> Hi Wolfgang and list,
>>>
>>> Is it possible to do vsn() or justvsn() with   
>>> backgroundSubtract=TRUE for a one-color genepix chip? I tried and   
>>> received an error saying that "object R was not found". Could I   
>>> fool the function duplicating G and Gb into R and Rb or will the   
>>> normalisation result then be modified and get erroneous output??
>>>
>>>
>>> Thanks for your help,
>>>
>>>
>>> David
>>>
>>> _______________________________________________
>>> 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
>
> _______________________________________________
> 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