[BioPython] Spatial clustering

[Andrew Dalke]

Shu-Hsien Sheu wrote:
> Next step would be to cluster these small molecules.

I don't know much about spatial clustering like what
you're talking about -- are you attempting to classify
binding affinities for different parts of the pocket?

Have you considered fingerprints (Daylight-style or MACCS
keys) or shape fitting (as from OpenEye)?  There's some
fingerprint generation code in frowns, or if you have
Daylight or OEChem it's easy to create on your own.

Another option you might consider is pharmacophore
points instead of RMSD.

					Andrew
					dalke at dalkescientific.com