[BioC] Should I compute RMA expression measures separately for different treament groups?

[Wolfgang Huber]

Dear Clémentine

the short answer is 'no'.

A slightly longer answer is that alternatives to RMA exist that are less 
aggressive in removing differences between arrays. There is a trade-off 
between removing unwanted technical variation and wanted biological 
variability. Different methods address the trade-off differently. 
Careful QA/QC on your data is needed, and you need to make sure that 
your experiment has the appropriate controls. In order to explore these 
issues (and in order to know how to defend your non-standard choice e.g. 
to reviewers), it would probably be best to contact a local statistician.

Best wishes
	Wolfgang

Btw, as always, the attachment (boxplot) did not make it through the 
mailing list.



On 12/9/11 7:50 PM, Clémentine Dressaire wrote:
> Dear BioC users,
>
> I was wondering wether it would not make sense to perform rma
> normalization on the 2 (or more) conditions we want to compare
> separately instead of alltogether as usually performed.
> This questioning comes from the three conditions I currently have to
> deal with. They clearly display different raw levels and I'm afraid that
> using global RMA will somehow hide an effect that could be meaningful.
> To illustrate I put in attachement the boxplots obtained before and
> after the two "types" of rma normalization.
>
> Did anyone already have to deal with such a problem? Do you have any
> suggestions?
>
> Many thanks for your help,
>
> Clémentine
>
>
>
> _______________________________________________
> 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