[BioC] obtaining the counts after batch effect removal

[Gilgi Friedlander]

Thanks a lot!
Gilgi

-----Original Message-----
From: Wolfgang Huber [mailto:whuber at embl.de] 
Sent: Monday, January 20, 2014 10:12 AM
To: Gilgi Friedlander
Cc: bioconductor at r-project.org
Subject: Re: [BioC] obtaining the counts after batch effect removal

Gilgi

the 'sva' package from Bioconductor, or the PEER package (http://europepmc.org/articles/PMC3398141) would seem like good places to start.

	Wolfgang


Il giorno 20 Jan 2014, alle ore 08:18, Gilgi Friedlander <gilgi.friedlander at weizmann.ac.il> ha scritto:

> Hi Mike,
> 
> Thanks a lot.
> 
> Could you possibly refer to packages that can take as input the DESeq transformed values for removing the batch effect?
> 
> Thanks a lot,
> Gilgi
> 
> From: Michael Love [mailto:michaelisaiahlove at gmail.com]
> Sent: Sunday, January 19, 2014 10:10 PM
> To: Gilgi [guest]
> Cc: bioconductor at r-project.org; Gilgi Friedlander
> Subject: Re: obtaining the counts after batch effect removal
> 
> hi Gilgi,
> 
> Please keep all discussion on list, we like to keep a record of questions and answers for all Bioc users.
> 
> The short answer is no, we don't offer any batch removal functions, something analogous to residuals after fitting the batch effect for Gaussian distributed data.
> 
> We do offer the two transformations (described in the vignette) for producing log2 scale data with stable variance across the range of counts (i.e. homoskedastic). The transformed data could then be used with various packages for removing batch effect.
> 
> Mike
> 
> On Sun, Jan 19, 2014 at 11:05 AM, Michael Love <michaelisaiahlove at gmail.com<mailto:michaelisaiahlove at gmail.com>> wrote:
> hi Gilgi,
> 
> By incorporating batch variables into the generalized linear model, we control for these effects, but what is happening is not a two step process of removal and then testing.
> 
> Mike
> 
> On Sun, Jan 19, 2014 at 5:34 AM, Gilgi [guest] <guest at bioconductor.org<mailto:guest at bioconductor.org>> wrote:
> 
> Hello,
> 
> I have a question regarding a design, in which there is a factor for "batch effect". I understand how to get the p-values when taking in account the batch factor.
> 
> I wanted to ask whether it is possible to obtain also the actual counts after batch effect removal.
> 
> Thanks a lot,
> Gilgi
> 
> -- output of sessionInfo():
> 
> none
> 
> --
> Sent via the guest posting facility at bioconductor.org<http://bioconductor.org>.
> 
> 
> 
> 	[[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