Why in news?
Scientists at Kolkata’s S.N. Bose National Centre for Basic Sciences announced PathGennie, an open‑source algorithm that drastically speeds up simulations of rare molecular events, such as the unbinding of drug molecules from protein targets. Released in December 2025, the method could accelerate computer‑aided drug discovery.
Background
In drug development, understanding how long a molecule remains bound to its target (the “residence time”) is crucial for efficacy and safety. Conventional molecular dynamics simulations struggle to capture unbinding events because they occur over micro‑ to millisecond timescales. Researchers typically apply artificial forces or raise temperatures to force unbinding, which can distort the physics.
PathGennie takes a different approach. It uses a direction‑guided adaptive sampling strategy inspired by natural selection:
- Thousands of ultrashort, unbiased trajectories (just a few femtoseconds) are launched from the starting structure.
- After each round, only those trajectories that move closer to the desired end state—such as ligand unbinding—are extended, while unproductive paths are discarded.
- This “survival of the fittest” selection quickly uncovers the pathway over the energy barrier without applying external bias.
- The algorithm can operate in any set of collective variables, including machine‑learned order parameters, making it flexible and general.
Applications and advantages
- Multiple pathways identified: Proof‑of‑concept studies showed that PathGennie could map various exit routes for benzene from the T4 lysozyme enzyme and reveal three distinct ways in which the anti‑cancer drug imatinib detaches from Abl kinase, all without forcing the process.
- Versatility: Beyond ligand unbinding, the framework can be applied to chemical reactions, catalytic processes, phase transitions and self‑assembly, wherever rare events hinder conventional simulations.
- Open access: The software is freely available, enabling researchers worldwide to incorporate it into their simulation workflows and combine it with machine‑learning tools.
Source: News On Air
