Metric and Multidimensional Scaling:  Efficient Tools for Clustering Molecular Conformations

Abstract

The application of metric and multidimensional scaling to conformer ensembles was demonstrated in this work. An automated process was devised to cluster and assign group memberships and cluster representatives. The method allows rapid clustering, leading to intuitive results that can be visually inspected. Multidimensional scaling was found to be superior to metric scaling for clustering conformers. The performance of different hierarchical clustering algorithms was compared using multidimensional plots, and the group average method was found to perform best.

Fuzzy Clustering as a Means of Selecting Representative Conformers and Molecular Alignments

“…We have recently described an efficient method to cluster molecular conformations using metric and multidimensional scaling. 1 We have also demonstrated recently 2 that clustering flexible molecular overlays 3 is useful in identifying potential binding modes and explaining partitioning behavior. The clustering process was shown to be robust and automatic, and the results could be displayed visually, enabling rapid determination of the optimum number of clusters in many cases.…”

Visualization of the Intrinsic Reaction Coordinate and Global Reaction Route Map by Classical Multidimensional Scaling

“…; as standard methods, principal component analysis, − multidimensional scaling (MDS), − and self-organizing maps , are well-known. In the MDS approach, multidimensional data are placed in a reduced dimensional space so that similar data are located close to each other, while the nonsimilar data are located far from each other. , The MDS method that employs a linear distance (Euclidean distance) to measure the similarity of data is called a classical MDS (CMDS), or a principal coordinate analysis. − By employing the CMDS approach for a given data set, the principal coordinates to represent the largest dispersion in geometrical structures can be chosen from all the degrees of freedom, based on the distances between each data. Very recently, the CMDS approach is applied to the classification of protein conformers and the analysis of molecular dynamics simulation. , …”

Evaluating data mining algorithms using molecular dynamics trajectories

“…Recent increases in computer power and enhancements in algorithms have enabled the execution of simulations of proteins on a large scale, resulting in huge amounts of data. One very useful way to analyse these data and extract important patterns is to cluster or classify molecular conformations into groups, according to the similarity of their conformations (as measured by an appropriate metric, for example the Root-Mean-Squared Distance (RMSD)) (Feher and Schmidt, 2001;Yona and Kedem, 2005).…”