[BioC] affyPara: background correction & normalization only?

[Markus Schmidberger]

Dear Leo,

good point.

I did some changes to preproPara(). No there is a summarization method 
'none' (summary.method='none') available. There will be no summarization 
step and the results object eset is NULL.
I submitted these changes and documentation to the svn. Version 1.7.1. 
should be available ad midnight: 
http://www.bioconductor.org/packages/devel/bioc/html/affyPara.html

Now at all nodes the bgc and normalized affyBatches are available in the 
GlobalEnvironment. It is very simple to run some code on that.

res <- clusterCall(cluster, FUN)
FUN<- function()
{
    require(affy)
    if (exists("AffyBatch", envir = .GlobalEnv))
           AffyBatch <- get("AffyBatch", envir = .GlobalEnv)

    # do anything you want on the AffyBatch.  

}

res will be a list of results from all nodes. You have to find a way to 
combine these results.

Best
Markus



Leo Lahti wrote:
> Dear affyPara package maintainers and BioC developer community,
>
>
> The parallelizations of Affy preprocessing in the affyPara package 
> provide essential tools to handle large array collections.
>
> Before starting to hack on this myself, I would like to ask if there 
> are workarounds in affyPara to obtain a preprocessed & normalized (but 
> not summarized) PM intensity matrix from CEL files? Alternatively, 
> having an access to virtual affybatch (i.e. keeping it in the nodes 
> without 'rebuild') would solve the problem. Is such functionality 
> available?
>
> I would need the probe-level values (for PM probes) after 
> preprocessing and normalization (but without probeset summarization 
> step)  - in an ideal case a probes x arrays matrix, and not 
> necessarily other information from the affybatch object. It seems that 
> both background correction and normalization can be done in affyPara 
> by reading in the CEL files directly. However, the output of these 
> methods in itself an affybatch which will cause the memory problems 
> the package is trying to solve.
>
> Thanks once more for relevant work.
>
> with kind regards
>
> Leo Lahti
> Department of Information and Computer Science
> Aalto University School of Science and Technology
> Finland
>


-- 
Dr. rer. nat. Markus Schmidberger

Ludwig-Maximilians-Universität München
IBE - Institut für medizinische Informationsverarbeitung,
Biometrie und Epidemiologie
Lehrstuhl für Biometrie und Bioinformatik
Marchioninistr. 15, D-81377 Muenchen
URL: http://www.ibe.med.uni-muenchen.de 
Mail: Markus.Schmidberger [at] ibe.med.uni-muenchen.de