Quality Threshold Clustering of Molecular Dynamics: A Word of Caution

Clustering Molecular Dynamics trajectories is a common analysis that allows grouping together similar conformations. Several algorithms have been designed and optimized to perform this routine task, and among them, Quality Threshold stands as a very attractive option. This algorithm guarantees that...

Full description

Saved in:
Bibliographic Details
Published in:Journal of chemical information and modeling 2020-02, Vol.60 (2), p.467-472
Main Authors: González-Alemán, Roy, Hernández-Castillo, David, Caballero, Julio, Montero-Cabrera, Luis A
Format: Article
Language:English
Subjects:
Algorithms
Cluster Analysis
Clustering
Computer simulation
Frames
Molecular dynamics
Molecular Dynamics Simulation
Quality
Quantum Theory
Trajectory analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Clustering Molecular Dynamics trajectories is a common analysis that allows grouping together similar conformations. Several algorithms have been designed and optimized to perform this routine task, and among them, Quality Threshold stands as a very attractive option. This algorithm guarantees that in retrieved clusters no pair of frames will have a similarity value greater than a specified threshold, and hence, a set of strongly correlated frames are obtained for each cluster. In this work, it is shown that various commonly used software implementations are flawed by confusing Quality Threshold with another simplistic well-known clustering algorithm published by Daura et al. (Daura, X.; van Gunsteren, W. F.; Jaun, B.; Mark, A. E.; Gademann, K.; Seebach, D. Peptide Folding: When Simulation Meets Experiment. Angew. Chemie Int. Ed. 1999, 38 (1/2), 236–240). Daura’s algorithm does not impose any quality threshold for the frames contained in retrieved clusters, bringing unrelated structural configurations altogether. The advantages of using Quality Threshold whenever possible to explore Molecular Dynamic trajectories is exemplified. An in-house implementation of the original Quality Threshold algorithm has been developed in order to illustrate our comments, and its code is freely available for further use by the scientific community.
ISSN:1549-9596
1549-960X
DOI:10.1021/acs.jcim.9b00558