Postprocessing of Pharmacophore-Based Screening Hits

As mentioned above, vfith respect to model selectivity and database size, hit lists of pharmacophore-based screenings can comprise tens to thousands of small molecules and can be prioritized and filtered by conformer-to-model alignment scores. For further filtering of hit lists, postprocessing methods using criteria for lead- and druglikeness, as well as structural diversity analysis, are available. 

Hann et al. [97] defined criteria for leadlikeness of compounds as follows a molecular weight equal to or smaller than 460, a dog P between —4 and 4.2, a log Sw equal to or larger than —5, a number of rotatable bonds equal to or smaller than 10, a number of rings equal to or smaller than 4, a number of hydrogen bond donors equal to or smaller than 5, and a number of hydrogen bond acceptors equal to or smaller than 9. 

The most famous filter for druglikeness is Lipinski s rule of five [98]. According to Lipmski, a compound is most likely to have good absorption and permeability if it has equal to or less than 5 hydrogen bond donors, equal to or less than 10 hydrogen bond acceptors, a molecular weight equal to or less than 500, and a clog P equal to or less than 5. 

Using sdentific informatics sofiware such as Pipeline Pilot [99] or script languages such as Perl [100] and Python [101], hit lists can be filtered with respect to these lead-and druglikeness criteria. 

---