VAMPnets for Deep Learning of Molecular Kinetics

Overview

Journal Nat Commun

Specialty Biology

Date 2018 Jan 4

PMID 29295994

Citations 112

Authors

Andreas Mardt

Luca Pasquali

Hao Wu

Frank Noe

Affiliations

Soon will be listed here.

Abstract

There is an increasing demand for computing the relevant structures, equilibria, and long-timescale kinetics of biomolecular processes, such as protein-drug binding, from high-throughput molecular dynamics simulations. Current methods employ transformation of simulated coordinates into structural features, dimension reduction, clustering the dimension-reduced data, and estimation of a Markov state model or related model of the interconversion rates between molecular structures. This handcrafted approach demands a substantial amount of modeling expertise, as poor decisions at any step will lead to large modeling errors. Here we employ the variational approach for Markov processes (VAMP) to develop a deep learning framework for molecular kinetics using neural networks, dubbed VAMPnets. A VAMPnet encodes the entire mapping from molecular coordinates to Markov states, thus combining the whole data processing pipeline in a single end-to-end framework. Our method performs equally or better than state-of-the-art Markov modeling methods and provides easily interpretable few-state kinetic models.

Citing Articles

DeepPath: Overcoming data scarcity for protein transition pathway prediction using physics-based deep learning.

Pang Y, Kuo K, Yang L, Gumbart J bioRxiv. 2025; .

PMID: 40060558 PMC: 11888466. DOI: 10.1101/2025.02.27.640693.

The QCML dataset, Quantum chemistry reference data from 33.5M DFT and 14.7B semi-empirical calculations.

Ganscha S, Unke O, Ahlin D, Maennel H, Kashubin S, Muller K Sci Data. 2025; 12(1):406.

PMID: 40057556 PMC: 11890765. DOI: 10.1038/s41597-025-04720-7.

Markov State Models with Weighted Ensemble Simulation: How to Eliminate the Trajectory Merging Bias.

Bose S, Kilinc C, Dickson A J Chem Theory Comput. 2025; 21(4):1805-1816.

PMID: 39933004 PMC: 11866749. DOI: 10.1021/acs.jctc.4c01141.

Using pretrained graph neural networks with token mixers as geometric featurizers for conformational dynamics.

Pengmei Z, Lorpaiboon C, Guo S, Weare J, Dinner A J Chem Phys. 2025; 162(4).

PMID: 39873278 PMC: 11779506. DOI: 10.1063/5.0244453.

Enhanced sampling of protein conformational changes via true reaction coordinates from energy relaxation.

Li H, Ma A Nat Commun. 2025; 16(1):786.

PMID: 39824807 PMC: 11742398. DOI: 10.1038/s41467-025-55983-y.

References

Trendelkamp-Schroer B, Wu H, Paul F, Noe F . Estimation and uncertainty of reversible Markov models. J Chem Phys. 2015; 143(17):174101. DOI: 10.1063/1.4934536. View

Doerr S, Harvey M, Noe F, De Fabritiis G . HTMD: High-Throughput Molecular Dynamics for Molecular Discovery. J Chem Theory Comput. 2016; 12(4):1845-52. DOI: 10.1021/acs.jctc.6b00049. View

Buchete N, Hummer G . Coarse master equations for peptide folding dynamics. J Phys Chem B. 2008; 112(19):6057-69. DOI: 10.1021/jp0761665. View

Kube S, Weber M . A coarse graining method for the identification of transition rates between molecular conformations. J Chem Phys. 2007; 126(2):024103. DOI: 10.1063/1.2404953. View

Lindorff-Larsen K, Piana S, Dror R, Shaw D . How fast-folding proteins fold. Science. 2011; 334(6055):517-20. DOI: 10.1126/science.1208351. View

Noe F . Probability distributions of molecular observables computed from Markov models. J Chem Phys. 2008; 128(24):244103. DOI: 10.1063/1.2916718. View

Horenko I, Hartmann C, Schutte C, Noe F . Data-based parameter estimation of generalized multidimensional Langevin processes. Phys Rev E Stat Nonlin Soft Matter Phys. 2007; 76(1 Pt 2):016706. DOI: 10.1103/PhysRevE.76.016706. View

Perez-Hernandez G, Paul F, Giorgino T, De Fabritiis G, Noe F . Identification of slow molecular order parameters for Markov model construction. J Chem Phys. 2013; 139(1):015102. DOI: 10.1063/1.4811489. View

Noe F, Wu H, Prinz J, Plattner N . Projected and hidden Markov models for calculating kinetics and metastable states of complex molecules. J Chem Phys. 2013; 139(18):184114. DOI: 10.1063/1.4828816. View

10.

Wang H, Schutte C . Building Markov State Models for Periodically Driven Non-Equilibrium Systems. J Chem Theory Comput. 2016; 11(4):1819-31. DOI: 10.1021/ct500997y. View

11.

Bowman G, Beauchamp K, Boxer G, Pande V . Progress and challenges in the automated construction of Markov state models for full protein systems. J Chem Phys. 2009; 131(12):124101. PMC: 2766407. DOI: 10.1063/1.3216567. View

12.

Prinz J, Chodera J, Pande V, Swope W, Smith J, Noe F . Optimal use of data in parallel tempering simulations for the construction of discrete-state Markov models of biomolecular dynamics. J Chem Phys. 2011; 134(24):244108. PMC: 3139503. DOI: 10.1063/1.3592153. View

13.

Chodera J, Swope W, Noe F, Prinz J, Shirts M, Pande V . Dynamical reweighting: improved estimates of dynamical properties from simulations at multiple temperatures. J Chem Phys. 2011; 134(24):244107. PMC: 3143679. DOI: 10.1063/1.3592152. View

14.

Noe F, Clementi C . Kinetic distance and kinetic maps from molecular dynamics simulation. J Chem Theory Comput. 2015; 11(10):5002-11. DOI: 10.1021/acs.jctc.5b00553. View

15.

Hummer G, Szabo A . Optimal Dimensionality Reduction of Multistate Kinetic and Markov-State Models. J Phys Chem B. 2014; 119(29):9029-37. PMC: 4516310. DOI: 10.1021/jp508375q. View

16.

Molgedey , SCHUSTER . Separation of a mixture of independent signals using time delayed correlations. Phys Rev Lett. 1994; 72(23):3634-3637. DOI: 10.1103/PhysRevLett.72.3634. View

17.

Weber M, Fackeldey K, Schutte C . Set-free Markov state model building. J Chem Phys. 2017; 146(12):124133. DOI: 10.1063/1.4978501. View

18.

Schwantes C, Pande V . Improvements in Markov State Model Construction Reveal Many Non-Native Interactions in the Folding of NTL9. J Chem Theory Comput. 2013; 9(4):2000-2009. PMC: 3673732. DOI: 10.1021/ct300878a. View

19.

Sheong F, Silva D, Meng L, Zhao Y, Huang X . Automatic state partitioning for multibody systems (APM): an efficient algorithm for constructing Markov state models to elucidate conformational dynamics of multibody systems. J Chem Theory Comput. 2015; 11(1):17-27. DOI: 10.1021/ct5007168. View

20.

Yao Y, Cui R, Bowman G, Silva D, Sun J, Huang X . Hierarchical Nyström methods for constructing Markov state models for conformational dynamics. J Chem Phys. 2013; 138(17):174106. DOI: 10.1063/1.4802007. View