Pharmacophore discovery datasets

Befitting the notion that these are fully automated pharmacophore discovery methods, this is minimal, although DISCO, Hiphop, and Catalyst s HypoGen require considerable user intervention to sift through the multiple pharmacophores that emerge. Furthermore, considerable effort is advocated in selecting a representative dataset for HypoGen to avoid nonsensical results. 

A new pharmacophore discovery method should be initially tested on Class I datasets if this is a failure, one should discard that method. If it is a success, that is no great achievement one must progress to steadily more difficult datasets. When comparing different pharmacophore discovery methodologies, one must take into account the relative differences of difficult of datasets to which they have been applied. 

Fig. 14.4 Pharmacophore-based database searching for drug discovery. This example is based on a glucocorticoid receptor dataset (Morgan et al., 2002).

---