Has anyone attempted to perform 2-way hierarchical clustering in which one dimension might be hierarchical clustering
of molecules using ECFP6 (or some other) fingerprints as a basis for the distance metric, and the other dimension might be
hierarchical clustering using perhaps a profile of biological activities for a set of assays e.g., a set of pIC50
measurements. Does anyone have a PP or PP + R stats protocol that they can share?
Is this approach useful for elucidating chemotypes and possible polypharmacology / target / mechanism of action
analysis?
Please share your experiences or ideas if you have tried this or had some success.
Thank you.
Regards,
Jim Metz
