[BioC] is-normalisation-really-required

Adaikalavan Ramasamy ramasamy at cancer.org.uk
Wed Apr 13 12:26:41 CEST 2005


Short answer : It is unlikely that you will get a paper published whose
results were based on unnormalised data.

Long answer : There are many systematic and unsystematic variation in
microarrays and it may be that in your case that some of these cancel
out nicely before normalisation.

To elaborate further on Sean's comments, let me give you an example.
Suppose you did a batch of arrays and its dye-swaps on today with a
particular PMT setting for scanning. 3 months later, you do another
batch of arrays and dye-swaps but with a higher PMT settings.

Assuming that all other experimental conditions are equal and of high
quality, then the dye-swap pairs would be consistent regardless of
batch. But all the intensity values of the 2nd batch would be much
higher than the 1st batch because of the higher PMT settings and thus
the two batches would not be comparable. In other words, the "center" of
the two batches are different and normalisation would be useful.

Regards, Adai




On Tue, 2005-04-12 at 14:00 -0400, Sean Davis wrote:
> On Apr 12, 2005, at 1:19 PM, vijayaraj nagarajan wrote:
> 
> > dear friends
> > i have situation, where i thought its ok for me not to
> > do normalisation, i am afraid i may be wrong. i want
> > your advice in this regard.
> >
> > we performed a wild type - mutant, dye-swap
> > experiment.
> > when we analysed the intensity values, they were
> > consistant among the two experiment (dye-swap). ie.,
> > almost same values for mutants in both the experiments
> > of the dye-swap.
> > since the values are almost same, i thought there
> > might not be any dye-bias, so i just went ahead,
> > averaged the two values, found out their ratio and
> > filtered genes with 2 fold change.
> >
> 
> If you don't normalize, then the "center" of your ratios may be 
> significantly off, not to mention the fact that the two arrays may have 
> different centers.  So, at the very least, I think you have to do 
> within-arrays normalization of some kind.
> 
> Sean
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
>



More information about the Bioconductor mailing list