ramasamy at cancer.org.uk
Wed Apr 13 15:02:44 CEST 2005
On Wed, 2005-04-13 at 14:29 +0200, Gorjanc Gregor wrote:
> > So, if the two results agree, then the results with normalization are
> > correct; if not then the results with normalization are still correct.
> > Sounds like we are pretty much stuck with normalization....
> > Sean
> Why should one do normalization if the results aren't different. But, in
> that case it really does not matter and one can do it or not.
Perhaps for the following reasons :
1) Generalisability and comparison of results with other datasets
2) Ease of programming and automation
4) In case there were other biases that we were not aware of. How do you
measure the agreement between two arrays may only show some aspects of
the data. For example, the M-A plot can reveal some biases within an
array that correlation between pairs will not.
There are many clever people (most of them on this list) who have spent
their time improving preprocessing algorithms. There must be a reason
why they have spent their time and effort on normalisation.
More information about the Bioconductor