Estimating Solvation Free Energies with Boltzmann Generators

2025-12-19

AI Summary

Traditional solvation free energy calculations rely on extensive sampling and multiple intermediate states, resulting in low computational efficiency and poor phase-space overlap. To address this, we propose a Boltzmann generator based on reversible normalizing flows, enabling the first direct, invertible mapping of solvent configurations across solutes of differing sizes—thereby substantially improving phase-space overlap. Our method integrates Boltzmann generative modeling with Lennard-Jones molecular dynamics, achieving high-accuracy free energy estimates (error < 0.5 kcal/mol) for challenging transformations such as solute growth and decoupling. Radial distribution function (RDF) analysis confirms physically consistent solvent reorganization. Compared to conventional free energy perturbation (FEP) or thermodynamic integration (TI) approaches, our method reduces required simulation steps by one to two orders of magnitude. This work establishes a new paradigm for efficient, end-to-end solvation thermodynamics computation.

Technology Category

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

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Accurate calculations of solvation free energies remain a central challenge in molecular simulations, often requiring extensive sampling and numerous alchemical intermediates to ensure sufficient overlap between phase-space distributions of a solute in the gas phase and in solution. Here, we introduce a computational framework based on normalizing flows that directly maps solvent configurations between solutes of different sizes, and compare the accuracy and efficiency to conventional free energy estimates. For a Lennard-Jones solvent, we demonstrate that this approach yields acceptable accuracy in estimating free energy differences for challenging transformations, such as solute growth or increased solute-solute separation, which typically demand multiple intermediate simulation steps along the transformation. Analysis of radial distribution functions indicates that the flow generates physically meaningful solvent rearrangements, substantially enhancing configurational overlap between states in configuration space. These results suggest flow-based models as a promising alternative to traditional free energy estimation methods.

Problem

• Estimates solvation free energies using Boltzmann generators
• Maps solvent configurations between different solute sizes
• Enhances configurational overlap to reduce simulation steps

Innovation

• Normalizing flows map solvent configurations directly
• Enhances configurational overlap between solute states
• Alternative to traditional free energy estimation methods