BPS2026 – Capturing rare events with new path sampling algorithms: Exact membrane permeation kinetics of 5-ALA from ∞RETIS simulations

Computational approaches for estimating thermodynamic properties such as free energies and binding affinities have seen significant progress. However, predicting kinetic properties for drug binding/unbinding rates and membrane permeation remains a greater challenge as the underlying events are often rare and slow. These processes are largely inaccessible to conventional molecular dynamics (MD) simulations, which are limited by timescale and sampling efficiency. This contribution shows how our new developments in path sampling methods can be used to create unbiased membrane permeation trajectories. ∞RETIS is the most advanced variant of the replica exchange transition interface sampling (RETIS) algorithm, featuring asynchronous swaps in the infinite swapping limit. This design enables efficient parallelization across CPUs and GPUs, dramatically accelerating convergence. ∞RETIS is applied to study the membrane permeation of 5-aminolevulinic acid (5-ALA), a hydrophilic drug used in photodynamic therapy and fluorescence-guided surgery. From unbiased trajectories, we compute key kinetic metrics including permeability and mean first passage times. Additionally, the unbiased trajectories elucidate the mechanistic details of 5-ALA permeation, highlighting how molecular orientation and hydration within the membrane influence successful translocation. By resolving the full kinetic pathway of 5-ALA crossing a phospholipid bilayer, this study demonstrates the capability of ∞RETIS to capture rare event dynamics in biologically and pharmaceutically relevant systems.