Statistical Measures to Quantify Similarity between Molecular Dynamics Simulation Trajectories
Open Access
Abstract
Molecular dynamics simulation is commonly employed to explore protein dynamics. Despite the disparate timescales between functional mechanisms and molecular dynamics (MD) trajectories, functional differences are often inferred from differences in conformational ensembles between two proteins in structure-function studies that investigate the effect of mutations. A common measure to quantify differences in dynamics is the root mean square fluctuation (RMSF) about the average position of residues defined by C -atoms. Using six MD trajectories describing three native/mutant pairs of beta-lactamase, we make comparisons with additional measures that include Jensen-Shannon, modifications of Kullback-Leibler divergence, and local p-values from 1-sample Kolmogorov-Smirnov tests. These additional measures require knowing a probability density function, which we estimate by using a nonparametric maximum entropy method that quantifies rare events well. The same measures are applied to distance fluctuations between C -atom pairs. Results from several implementations for quantitative comparison of a pair of MD trajectories are made based on fluctuations for on-residue and residue-residue local dynamics. We conclude that there is almost always a statistically significant difference between pairs of 100 ns all-atom simulations on moderate-sized proteins as evident from extraordinarily low p-values.
Keywords
Funding Information
- National Institutes of Health (GM101570)
This publication has 54 references indexed in Scilit:
- Quantitative Comparison of Conformational EnsemblesEntropy, 2012
- Is an Intuitive Convergence Definition of Molecular Dynamics Simulations Solely Based on the Root Mean Square Deviation Possible?Journal of Computational Biology, 2011
- Best Probability Density Function for Random Sampled DataEntropy, 2009
- Similarity Measures for Protein EnsemblesPLOS ONE, 2009
- Dynamic fluctuation of proteins watched in real timeHFSP Journal, 2008
- GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular SimulationJournal of Chemical Theory and Computation, 2008
- Ensemble-Based Convergence Analysis of Biomolecular TrajectoriesBiophysical Journal, 2006
- GROMACS: Fast, flexible, and freeJournal of Computational Chemistry, 2005
- Calcium-induced conformational transition revealed by the solution structure of apo calmodulinNature Structural & Molecular Biology, 1995
- Comparison of simple potential functions for simulating liquid waterThe Journal of Chemical Physics, 1983