Unified Biomolecular Trajectory Generation via Pretrained Variational Bridge

2026-02-07

AI Summary

This work addresses the high computational cost of molecular dynamics simulations and the limitations of existing generative models in cross-system generalization and structural information utilization. The authors propose the Pretrained Variational Bridge (PVB) method, which employs an encoder–decoder architecture to map initial molecular structures into a noisy latent space and introduces a multi-stage bridge matching mechanism to unify the training paradigms for single-structure and trajectory data for the first time. Furthermore, they incorporate an adjoint-matching-based reinforcement learning optimization strategy that significantly enhances the efficiency and accuracy of generating full-binding conformations of protein–ligand complexes. Experiments demonstrate that PVB accurately reproduces thermodynamic and kinetic observables from molecular dynamics simulations on both protein and protein–ligand systems, enabling stable and efficient biomolecular trajectory generation.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Search and Optimization: Sampling/Simulation-based Search
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

Molecular Dynamics (MD) simulations provide a fundamental tool for characterizing molecular behavior at full atomic resolution, but their applicability is severely constrained by the computational cost. To address this, a surge of deep generative models has recently emerged to learn dynamics at coarsened timesteps for efficient trajectory generation, yet they either generalize poorly across systems or, due to limited molecular diversity of trajectory data, fail to fully exploit structural information to improve generative fidelity. Here, we present the Pretrained Variational Bridge (PVB) in an encoder-decoder fashion, which maps the initial structure into a noised latent space and transports it toward stage-specific targets through augmented bridge matching. This unifies training on both single-structure and paired trajectory data, enabling consistent use of cross-domain structural knowledge across training stages. Moreover, for protein-ligand complexes, we further introduce a reinforcement learning-based optimization via adjoint matching that speeds progression toward the holo state, which supports efficient post-optimization of docking poses. Experiments on proteins and protein-ligand complexes demonstrate that PVB faithfully reproduces thermodynamic and kinetic observables from MD while delivering stable and efficient generative dynamics.

Problem

• Molecular Dynamics
• Trajectory Generation
• Generative Models
• Structural Information
• Computational Cost

Innovation

• Pretrained Variational Bridge
• bridge matching
• generative molecular dynamics
• reinforcement learning optimization
• cross-domain structural knowledge