IUCr2026

A robust CSP platform for predicting crystal polymorphs of pharmaceuticals, agrochemicals, petrochemicals and energetic materials

Speaker:
Shiva Sekharan, Global Portfolio Leader of Formulations/CSP at Schrödinger

Abstract:
The ability of an active ingredient (AI) in pharmaceuticals, agrochemicals, petrochemicals and energetic materials to exist in multiple crystalline forms, which significantly affects solubility, stability, downstream processing, bioavailability, and efficacy is called crystal polymorphism. The synthesis and experimental testing of polymorphism in small molecules can be time consuming, expensive and even hazardous, but their many industrial applications necessitate their constant research and development. Over 50% of AIs exhibit polymorphism, requiring careful control during manufacturing to avoid inconsistent performance during formulation. A less stable (metastable) form may dissolve faster, while a more stable form may have better shelf life. We have developed a robust crystal structure prediction (CSP) platform that leverages a novel, systematic crystal packing search and a hierarchical energy ranking protocol. By integrating machine learning force fields with molecular dynamics and quantum mechanics, the platform efficiently generates and identifies the most stable crystal polymorphs directly from molecular structure. Access to our CSP platform is available via the Schrödinger’s Materials Sciences and Life Sciences software suite as well as via service engagements.