[BioC] Two way ANOVA with bioconductor.

[Susan Holmes]

Alex,
If I were you I would start by using Hubers'
vsn normalisation, it takes out the dependence
of the variance to the expression intensities
and then makes for an easier analysis.
In the multtest pacakge you have the option
of using the F or a nonparametric equivalent which
is what I always do as the data often are NOT normal,

Best
Susan Holmes
Statistics
Stanford