[rmaexpress_help] Use of Quantile normalization (was Re: Re: RMAExpress 1.0 beta 9)
- From: Ben Bolstad <bmb@xxxxxxxxxxxxx>
- To: rmaexpress_help@xxxxxxxxxxxxx
- Date: Thu, 03 Apr 2008 07:01:04 -0700
Well, most garden variety normalization methods (including standard
quantile normalization implementations) make the implicit assumption
that the number of genes changing is small or that at least
approximately the same number of genes are going up and going down.
These assumptions may not hold very well in the situation you mention
below.
But the real question is whether or not you intend to make direct
comparisons between tissues. If you do, then I would still recommend
using quantile normalization across all arrays. Although it might be
sub-optimal, my experience has been that it is less likely to create
artifactual differences then approaches where you normalize separately
within tissues and then try to compare between them.
Best,
Ben
On Thu, 2008-04-03 at 10:02 +0200, Olivier Schaad wrote:
> Dear Ben
> I have a question about the use of the quantile normalization in
> the context of an experiment that have two different tissues like
> endonorium and drg or kidney and liver etc.
> do you thing it is better to normalize liver alone then Kidney alone or
> everything in one set. May worry is to mixe pear and apple
>
> best regards
> Olivier
>
>
> Ben Bolstad wrote:
> > RMAExpress 1.0 beta 9 has been released. As usual it can be downloaded
> > at:
> >
> > http://rmaexpress.bmbolstad.com
> >
> > Changes since 1.0 beta 8
> > - Better corruption checking when parsing CEL files. This should prevent
> > crashes of the same manner discussed in a recent mailing list thread:
> >
> > //www.freelists.org/archives/rmaexpress_help/03-2008/msg00000.html
> >
> >
> >
> >
> >
>
>
Other related posts:
- » [rmaexpress_help] Use of Quantile normalization (was Re: Re: RMAExpress 1.0 beta 9)
