Data-driven automatic synthesis planning: Synthesis routes of S-Zanubrutinib identified with CDI CASP

Advances in computer-assisted synthesis planning (CASP) are revolutionising how new functional molecules in many chemistry-using industries are being developed. CASP tools allow to assemble and analyse prior knowledge of a specified chemical system (a molecule, a reaction, a synthesis route), to generate hypotheses on experimental campaigns that could either be performed manually or using automated reaction systems. Advanced CASP tools are combining data science, chemoinformatics, machine learning and physical models-based predictive tools. Compared to expert-based synthesis planning, the power of CASP techniques allows for faster and more comprehensive planning, which could significantly improve the efficiencies of chemical process/product development. This White Paper describes a recent collaboration project between Shionogi & Co. Ltd. and Chemical Data Intelligence (CDI) Ltd. The CASP system developed by CDI (CDI-CASP) was tested in developing a new synthesis of S-Zanubrutinib, a drug for lymphoma treatment. Three types of search in CDI-CASP - “search synthesis routes”, “search analogue routes” and “search chiral reactions” - were iteratively applied for synthesis planning. Setting search criteria requires expert involvement. This ‘human in the middle’ interactive strategy leads to a shorter, greener, and more efficient synthesis route compared to the benchmark route filed in a patent.