Reaction-Based Enumeration, Active Learning, and Free Energy Calculations To Rapidly Explore Synthetically Tractable Chemical Space and Optimize Potency of Cyclin-Dependent Kinase 2 Inhibitors

• DOI: 10.1021/acs.jcim.9b00367
• Publication Date: 2019-08-12T21:14:22Z
• Authors: Kyle D. Konze; Pieter H. Bos; Markus K. Dahlgren; Karl Leswing; Ivan Tubert-Brohman; Andrea Bortolato; Braxton Robbason; Robert Abel; Sathesh Bhat
• Keywords: Chemical space; Cheminformatics; Profiling (computer programming); Lead compound; Docking (animal); Pathfinder

The hit-to-lead and lead optimization processes usually involve the design, synthesis, profiling of thousands analogs prior to clinical candidate nomination. A hit finding campaign may begin with a virtual screen that explores millions compounds, if not more. However, this scale computational is frequently performed in or phases drug discovery. This likely due lack appropriate tools generate synthetically tractable lead-like compounds silico, methods accurately profile prospectively on large scale. Recent advances power provide ability much larger libraries ligands than previously possible. Herein, we report new technique, referred as “PathFinder”, uses retrosynthetic analysis followed by combinatorial synthesis novel accessible chemical space. In work, integration PathFinder-driven compound generation, cloud-based FEP simulations, active learning are used rapidly optimize R-groups, cores for inhibitors cyclin-dependent kinase 2 (CDK2). Using approach, explored >300 000 ideas, >5000 identified >100 predicted IC50 < 100 nM, including four unique cores. To our knowledge, largest set calculations disclosed literature date. rapid turnaround time, exploration, suggests useful approach accelerate discovery matter campaigns.