[BioC] Normalizing non-normal array data (experimental sample vs. reference)

[Oliver Homann]

Hello,

I'd like to solicit some advice regarding normalization of two-color 
arrays in which one sample is a reference pool.  These arrays may 
violate the assumption that M~1 for most spots, and thus I am unsure of 
how best to correct the spatial and intensity biases that crop up in 
some of our arrays.  While it may be possible to achieve some degree of 
normalization post-transformation, the process of transforming between 
arrays can mask intensity-effects (especially if the scan intensities or 
sample qualities differ) as well as spatial effects.

Any advice on how to approach this problem would be much appreciated.

-oliver-

P.S.  Thanks for the earlier advice on using OLIN to correct spatial 
biases.  The quality of my (non-reference-based) array data was much 
improved by this approach.