[BioC] DESeq countDataSets and arrayQualityMetrics (was: Re: DESeq estimateDispersion options for lower depth miRNA-seq)

[Wolfgang Huber]

Dear Praful,

DESeq >=1.9.1 in the devel branch contains a new function 
varianceStabilizingTransformation that returns the variance-stabilised 
data in an ExpressionSet, which you can then directly feed to 
arrayQualityMetrics, e.g. like this:

library("pasilla")
library("vsn")
library("arrayQualityMetrics")

data("pasillaGenes")

cdsBlind = estimateDispersions(estimateSizeFactors(pasillaGenes), 
method="blind")
pasillaVst = varianceStabilizingTransformation(cdsBlind)

meanSdPlot(pasillaVst)

arrayQualityMetrics(pasillaVst, intgroup=c("condition", "type"), force=TRUE)


Let me know if this works for you. As usual, the code will need a day or 
two to percolate from the svn archive to the website.

Kind regards
	Wolfgang



Mar/29/12 9:53 PM, Wolfgang Huber scripsit::
>
> Dear Praful
>
> On 3/29/12 8:00 PM, Aggarwal, Praful wrote:
>> Hello Wolfgang,
>>
>> In your previous reply you mentioned that I "do the 6 MA -plots for
>> all pairs of samples and where the hits are in there." According to
>> the DESeq documentation, an MA plot was done after we get the
>> foldchange using the nbinomTest. However, do you suggest trying MA
>> plots using the normalized counts of all 4 samples or do you mean
>> something else when you mentioned looking at the MA plots.
>
> Yes, you can do several things:
> 1. A = average expression over all six samples and M=log fold change
> between sum of the counts within groups
> 2. Each pairwise MA-plot
>
> I was thinking of 2. I think what you refer to in the documentation is 1.
>
>> Also, I tried calling 'arrayQualityMetrics' on the
>> variance-stabilized version with both the 'maximum' and 'fit-only'
>> options. However, in both cases I get the following error :
>>
>> Error in function (classes, fdef, mtable) : unable to find an
>> inherited method for function "platformspecific", for signature
>> "matrix"
>>
>> I have never used arrayQualityMetrics, but even after looking at the
>> function I couldn't quite understand what might be causing this
>> error. Any help on this would be greatly appreciated.
>
> Sorry. Indeed this requires some programming on my end to make it
> convenient. I'll implement this asap and will announce it here.
>
> Until then, if you are suffciently familiar with Bioconductor data
> structures, what needs to be done is to create an ExpressionSet which
> contains the data matrix, and whose featureData and phenoData are taken
> from the countDataSet. This is explained in the vignette of the package
> Biobase "An introduction to Biobase and ExpressionSets".
>
> Best wishes
> Wolfgang
>
>
>
>
>> Kind Regards, Praful
>>
>> Praful Aggarwal Broeckel Lab Human and Molecular Genetics Center
>> Medical College of Wisconsin 8701 Watertown Plank Road CRI/TBRC 2nd
>> Floor Milwaukee, WI 53226
>>
>> 414-955-2567
>>
>>
>> [[alternative HTML version deleted]]
>>
>> _______________________________________________ Bioconductor mailing
>> list Bioconductor at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/bioconductor Search the
>> archives:
>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>


-- 
Best wishes
	Wolfgang

Wolfgang Huber
EMBL
http://www.embl.de/research/units/genome_biology/huber