P07-26
Development of a Platform for Crystal Structure Prediction of Drug Molecules
Okimasa OKADA *1, Yuya KINOSHITA2, Koki NISHIMURA2, Aaron NESSLER3, Hiroomi NAGATA1, Michael SCHNIEDERS3, Kaori FUKUZAWA4, Etsuo YONEMOCHI5
1Mitsubishi Tanabe Pharma Corporation
2Takeda Pharmaceutical Company Limited
3University of Iowa
4Osaka University
5International University of Health and Welfare
( * E-mail: okada.okimasa@mc.mt-pharma.co.jp )
In the design and quality control of drugs, crystalline polymorphism differs in physical properties such as solubility and physical/chemical stability, and affects the quality of drugs. Therefore, polymorphism is one of the important control items for quality control of drugs. It is generally reported that crystalline polymorphism is present in 80% of pharmaceutical compounds. Pharmaceutical companies perform screening experiments to select the optimal crystal form for each drug, but it is not possible to cover all crystallization conditions within a limited time period. Therefore, the crystal form is determined by screening experiments under certain crystallization conditions. However, many pharmaceutical companies have experienced troubles where a new stable crystal form suddenly appears in the manufacturing process or during long-term storage.
Therefore, reducing the possibility to have new stable crystal forms is an important issue for pharmaceutical companies with a mission to stably supply drugs to patients. One potential solution is computational crystal structure prediction. In this research, we aim to develop a platform for automatic prediction calculation of stable crystal structures. In this system, a stable crystal structure is output from a 2D molecular structure by creating a polarizable force field, generating a stable candidate crystal structure for each space group, narrowing down the crystal structure by lattice energy and density, removing duplicates, and high-precision lattice energy calculation by density functional theory.