Machine-learning model can develop high potency molecules much more quickly as compared to conventional methods
Drug development is time consuming procedure that is manual and prone to errors. MIT researchers developed new system to automate design process, which can speed up development process along with providing better results. Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Department of Electrical Engineering and Computer Science (EECS) have developed a model that selects lead molecule candidates based on desired properties. It also modifies the molecular structure required to achieve a higher potency, while ensuring the molecule is still chemically valid.
The model basically takes as input molecular structure data and directly creates molecular graphs — detailed representations of a molecular structure, with nodes representing atoms and edges representing bonds. It breaks those graphs down into smaller clusters of valid functional groups that it uses as building blocks that help it more accurately reconstruct and better modify molecules.
The researchers trained their model on 250,000 molecular graphs from the ZINC database, a collection of 3-D molecular structures available for public use. They tested the model on tasks to generate valid molecules, find the best lead molecules, and design novel molecules with increase potencies. Furthermore, the team aims to examine test models based on more properties, beyond solubility, which are more therapeutically relevant. The work was supported by the Defense Advanced Research Projects Agency ‘Make-It’ program. Paper describing the model was presented at the 2018 International Conference on Machine Learning.