Machine learning coarse-grained potentials of protein thermodynamics

2023/09/15 by Maciej Majewski, Adrià Pérez, Philipp Thölke +7 · 1 voice · 3 citations
Materials Science · Biochemistry, Genetics and Molecular Biology · #Machine Learning in Materials Science #Protein Structure and Dynamics #Enzyme Structure and Function

paper · pdf · doi:10.1038/s41467-023-41343-1

openalex publication_date 2023/09/15 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/31

Abstract

A generalized understanding of protein dynamics is an unsolved scientific problem, the solution of which is critical to the interpretation of the structure-function relationships that govern essential biological processes. Here, we approach this problem by constructing coarse-grained molecular potentials based on artificial neural networks and grounded in statistical mechanics. For training, we build a unique dataset of unbiased all-atom molecular dynamics simulations of approximately 9 ms for twelve different proteins with multiple secondary structure arrangements. The coarse-grained models are capable of accelerating the dynamics by more than three orders of magnitude while preserving the thermodynamics of the systems. Coarse-grained simulations identify relevant structural states in the ensemble with comparable energetics to the all-atom systems. Furthermore, we show that a single coarse-grained potential can integrate all twelve proteins and can capture experimental structural features of mutated proteins. These results indicate that machine learning coarse-grained potentials could provide a feasible approach to simulate and understand protein dynamics.

Citations

• The multiscale coarse-graining method. I. A rigorous bridge between atomistic and coarse-grained models
• Nonphysical sampling distributions in Monte Carlo free-energy estimation: Umbrella sampling
• Perspective: Coarse-grained models for biomolecular systems
• Coarse-grained models of protein folding: toy models or predictive tools?
• The relative entropy is fundamental to multiscale and inverse thermodynamic problems
• Markov models of molecular kinetics: Generation and validation
• Molecular dynamics simulations of biomolecules
• Atomistic Description of the Folding of a Dimeric Protein
• SchNet – A deep learning architecture for molecules and materials
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Coarse graining molecular dynamics with graph neural networks
• Anomalous collapses of Nares Strait ice arches leads to enhanced export of Arctic sea ice
• Multi-body effects in a coarse-grained protein force field
• Highly accurate protein structure prediction with AlphaFold
• Accurate prediction of protein structures and interactions using a three-track neural network

Cited by

• AI-based Methods for Simulating, Sampling, and Predicting Protein Ensembles
• TICA-Based Free Energy Matching for Machine-Learned Molecular Dynamics
• Accelerating Protein Molecular Dynamics Simulation with DeepJump

Discussions

• #CitizenScience: Neue #OpenAccess Arbeit von @gpugrid: 9 ms lange #MDSimulationen für 12 #Proteine, erstellt mit #MachineLearning Ansätzen: doi.org/10.1038/s414... Anschauliches #Video, wie man sich [bsky, 1 points, 0 comments]

Related

• Navigating protein landscapes with a machine-learned transferable coarse-grained model
• Top-down machine learning of coarse-grained protein force-fields
• Machine Learning of coarse-grained Molecular Dynamics Force Fields
• A unified framework for coarse grained molecular dynamics of proteins with high-fidelity reconstruction
• Coarse graining molecular dynamics with graph neural networks