Novel Scaffold Fingerprint (SFP): Applications in Scaffold Hopping and Scaffold-Based Selection of Diverse Compounds

A novel
2D Scaffold FingerPrint (SFP) for mining ring fragments is presented. The
rings are described not only by their topology, shape, and pharmacophoric
features (hydrogen-bond acceptors and donors, their relative locations,
sp3 carbons, and chirality) but also by the position and nature of
their growing vectors because they play a critical role from the drug
discovery perspective. SFP can be used (i) to identify alternative
chemotypes to a reference ring either in a visual mode or by running
quantitative similarity searches and (ii) in chemotype-based diversity
selections. Two retrospective case studies focused on melanin concentrating
hormone 1-receptor antagonists (MCH-R1) and phosphodiesterase-5 inhibitors
(PDE5) demonstrate the capability of this method for identifying novel
structurally different and synthetically accessible chemotypes. Good
enrichment factor (155 and 219) and recall values (46% and 73%) are
found within the first 100 ranked hits (0.3% of screened database).
Our 2D SFP descriptor outperforms well-validated current gold-standard
2D fingerprints (ECFP_6) and 3D approaches based on shape and electrostatic
similarity. Scaffold-based selection of diverse compounds has a critical
impact on corporate library design and compound acquisitions; thus,
a novel strategy is introduced that uses diverse scaffold selections
using this SFP descriptor combined with R-group selection at the different
substitution sites. Both approaches are available as part of an interactive
web-based application that requires minimal input and no computational
knowledge by medicinal chemists.