[BioC] cross experiments or geo series normalization

Robert Gentleman rgentlem at fhcrc.org
Thu Jun 30 15:07:17 CEST 2005

Hi Chuming,

  If what you are asking about is the joint normalization of multiple 
different experiments then I think the answer is that it is probably not 
a good idea. You will almost surely need some sort of random effects 
model to address between experiment differences regardless of whether 
the joint normalization "works", and you can use that on experiments 
normalized separately. I am aware of no study that compares joint versus 
separate normalization, but I doubt that it would be of much real value, 
even if done. There are examples where joint normalization of highly 
related experiments (same people, similar biological material and 
similar processing) that do not remove all experimental artifacts, and 
hnece it is unlikely that this procedure would have better results on 
unrelated experiments.
  Random effects models have provided the basis for such analyses for 
some years now. Cox and Solomon, Components of Variance, have a pretty 
good discussion of the issues involved and include explicit discussion 
of microarray experiments.


Chuming Chen wrote:
> Hi, all,
> Has anybody done cross experiments or GEO series normalization on 
> Affymetrix arrays? Any suggestions or related links will be highly 
> appreciated.
> Thanks,
> Chuming Chen
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor

More information about the Bioconductor mailing list