IIT Madras, Ohio State university develops AI framework to aid discovery of next-generation drugs
Named PURE (Policy-guided Unbiased REpresentations for Structure-Constrained Molecular Generation), this system has the potential to dramatically reduce the early-stage timelines of drug development, a process that currently takes up to a decade and costs billions of dollars.
IIT Madras
CHENNAI: Researchers from the Indian Institute of Technology Madras’s Wadhwani School of Data Science and Artificial Intelligence (WSAI) and Ohio State University have developed an AI framework capable of rapidly generating drug-like molecules that are both novel and easy to synthesise in real-world laboratory conditions.
Named PURE (Policy-guided Unbiased REpresentations for Structure-Constrained Molecular Generation), this system has the potential to dramatically reduce the early-stage timelines of drug development, a process that currently takes up to a decade and costs billions of dollars.
Researchers believe it could play a transformative role in combating drug resistance in cancer and infectious diseases. “Unlike conventional molecule-generation tools that depend on rigid scoring systems or narrow optimisation goals, PURE employs a reinforcement learning approach that enables it to ‘think’ more like a chemist. It simulates step-by-step molecular transformations using templates derived from actual chemical reactions, ensuring that the molecules it proposes are synthetically feasible,” a release from the Institute read.
Published in the Journal of Cheminformatics, the study demonstrates that PURE significantly outperforms existing AI models on established molecular benchmarks such as drug-likeness, dopamine receptor activity, and solubility. “AI is reshaping how we approach discovery itself. PURE moves us closer to AI systems that can reason through synthesis steps much like a human chemist,” said B Ravindran, head, WSAI, IIT-M.
Professor Karthik Raman added, “By eliminating bias towards specific metrics, PURE maps chemical space more effectively and grounds it in synthesisability.”
Professor Srinivasan Parthasarathy of Ohio State University piped in with: “The framework could revolutionise pharmaceutical research by identifying alternative, more effective drug candidates and accelerating the discovery of new materials.”
