Abstract:Quantifying the conformational ensembles of biomolecules is fundamental to describing mechanisms of processes such as ligand binding and allosteric regulation. Accurate quantification of these ensembles remains a challenge for all but the simplest molecules. One such challenge is insufficient sampling which enhanced sampling approaches, such as metadynamics, were designed to overcome; however, the non-uniform frame weights that result from many of these approaches present an additional challenge to ensemble quantification techniques such as Markov State Modeling or structural clustering. Here, we present rigorous inclusion of non-uniform frame weights into a structural clustering method entitled shapeGMM. The shapeGMM method fits a Gaussian mixture model to particle positions, and here we advance that approach by incorporating nonuniform frame weights in the estimates of all parameters of the model. The resulting models are high dimensional probability densities for the unbiased systems from which we can compute important thermodynamic properties such as relative free energies and configurational entropy. The accuracy of this approach is demonstrated by the quantitative agreement between GMMs computed by Hamiltonian reweighting and direct simulation of a coarse-grained helix model system. Furthermore, the relative free energy computed from a high dimensional probability density of alanine dipeptide reweighted from a metadynamics simulation quantitatively reproduces the metadynamics free energy in the basins. Finally, the method identifies hidden structures along the actin globular to filamentous-like structural transition from a metadynamics simulation on a linear discriminant analysis coordinate trained on GMM states, demonstrating the broad applicability of combining our prior and new methods, and illustrating how structural clustering of biased data can lead to biophysical insight. Combined, these results demonstrate that frame-weighted shapeGMM is a powerful approach to quantify biomolecular ensembles from biased simulations.

Discriminant Analysis Optimizes Progress Coordinate in Weighted Ensemble Simulations of Rare Event Kinetics

Unsupervised learning of progress coordinates during weighted ensemble simulations: Application to millisecond protein folding

Reaction Coordinates for Conformational Transitions using Linear Discriminant Analysis on Positions

Frequency Adaptive Metadynamics for the Calculation of Rare-Event Kinetics.

Structuring and sampling complex conformation space: Weighted ensemble dynamics simulations

A Mode Evolution Metric to Extract Converged Reaction Coordinates for Biomolecular Conformational Transitions

Good rates from bad coordinates: the exponential average time-dependent rate approach

Reweighted ensemble dynamics simulations: theory, improvement, and application

RiteWeight: Randomized Iterative Trajectory Reweighting for Steady-State Distributions Without Discretization Error

Simulating rare events using a Weighted Ensemble-based string method

Quantifying Unbiased Conformational Ensembles from Biased Simulations Using ShapeGMM

Weighted ensemble milestoning (WEM): A combined approach for rare event simulations

Learning reaction coordinates via cross-entropy minimization: Application to alanine dipeptide

Investigation of rare protein conformational transitions via dissipation-corrected targeted molecular dynamics

Best of Both Worlds: Enforcing Detailed Balance in Machine Learning Models of Transition Rates

Augmenting Human Expertise in Weighted Ensemble Simulations through Deep Learning based Information Bottleneck

Order parameter prediction from molecular dynamics simulations in proteins

Teaching old docks new tricks with machine learning enhanced ensemble docking

Machine Learning for Parameter Auto-tuning in Molecular Dynamics Simulations: Efficient Dynamics of Ions near Polarizable Nanoparticles

Automated Event Detection and Activity Monitoring in Long Molecular Dynamics Simulations