Biophysical Society Thematic Meeting | Tutzing 2026

Single-Molecule FRET: The Next 30 Years

Poster Abstracts

73-POS Board 37 BAYESIAN MODELING OF SINGLE-MOLECULE FRET DATA BASED ON LANGEVIN DYNAMICS FOR TRANSITION-PATH-TIME ANALYSIS Tomoaki Yagi 1 ; Kunihiko Ishii 2,3 ; Tahei Tahara 2,3 ; 1 RIKEN, Theoretical Molecular Science Laboratory, Wako, Japan 2 RIKEN, Molecular Spectroscopy Laboratory, Wako, Japan 3 RIKEN, Ultrafast Spectroscopy Research Team, Wako, Japan Biomolecular function is governed by continuous stochastic motions on rugged energy landscapes, but single-molecule Förster resonance energy transfer (smFRET) data are commonly interpreted using discrete-state models. The objective of this study is to develop a Bayesian framework that infers continuous molecular dynamics directly from photon-by-photon FRET measurements and applies the inferred dynamics to transition-path-time analysis.We model the hidden variable as the time-dependent inter-dye distance evolving under overdamped Langevin dynamics. The corresponding Fokker–Planck propagator defines a physically constrained state space model. Each detected photon is described by its inter-photon interval, donor or acceptor detection channel, and fluorescence delay time after pulsed excitation. The likelihood combines distance-dependent FRET efficiency, donor and acceptor lifetime distributions, and the instrumental response function. Bayesian filtering and smoothing are used to infer posterior distributions of continuous distance trajectories, and an expectation–maximization procedure estimates the free-energy profile and friction coefficient from the photon stream.The inferred posterior trajectories allow barrier-crossing events to be analyzed without assigning photons to predefined states. By propagating posterior-conditioned diffusion with absorbing boundaries, first-passage-time and transition-path-time distributions are computed as probability fluxes. This approach converts sparse and noisy photon records into dynamical quantities such as free-energy landscapes, friction coefficients, first-passage statistics, and transition-path times.Our results show that Bayesian Langevin modeling provides a direct route from single photons to continuous molecular trajectories. By moving beyond discrete-state representations, the framework expands smFRET analysis from state classification toward quantitative inference of transient biomolecular dynamics and provides a basis for testing energy-landscape descriptions of molecular function.

99