[BioC] large data set correlations?

[Park, Richard]

Hi Everyone, 
this is a post to get some ideas on ways to approach large data set analyses. For instance, I have over 100+ affymetrix chips organized into 23 different categories of cell types. So basically it is a matrix with 23 columns and 12500 rows. What methods would people suggest to correlate probes? to be able to create clusters of similar behaving probe sets? 

It is an idea similar to what is found at http://expression.gnf.org, but it just provides basic bar plots of the different cell types. I'd like to use some sort of algorithm to "mine" the data. 

I have tried dealing w/ correlation coefficients between sets of probes, which can be visualized using software such as vxinsight (which basically creates a 3d surface placing probes together based on a correlation coefficient matrix). 

Any ideas would be sincerely appreciated. 

- richard park