ACS Medicinal Chemistry Letters Innovations Webinar:
AI-assisted scaffold hopping and generative design of synthetically feasible lead analog space
Speaker:
Greg Makara, ChemPass Ltd. (Hungary)
Derivatization Design of Synthetically Accessible Space for Optimization: In Silico Synthesis vs Deep Generative Design.
Published in ACS Med. Chem. Lett. 2021, 12, 185−194.
Moderator:
György M. Keserű, Research Centre for Natural Sciences (Hungary)
AI-assisted scaffold hopping and generative design of synthetically feasible lead analog space
Summary:
Synthesis is typically the most time-consuming step in the DMTA cycle of lead optimization and timelines can only be effectively controlled if the design methodology considers synthetic feasibility and reagent availability. On the other hand, the completeness of ideation impacts the number of optimization cycles required to reach a preclinical candidate. Thus, the ability to generate a comprehensive and synthetically accessible idea set for optimization cycles promises a step change in the lead optimization paradigm. The SynSpace software has been created as a user-friendly design software to assist medicinal and computational chemists in accessing relevant, synthetically feasible chemical space around their lead series. Scaffold hopping, side-chain optimization, different generative design methods, as well as retrosynthesis are all easily carried out with a few clicks of the mouse and without need for cheminformatic or synthetic expertise. Design results are tabulated including full synthetic schemes, reagent data and compound properties for easy ranking and further processing. In this webinar, we will introduce and demo SynSpace, compare its results to that of a deep learning generative design technique, and present a few lead optimization case studies.