Topological characterization of statistically clustered networks for molecular similarity analysis

Statistical clustering technique is extensively employed in molecular similarity analysis where molecules can be clustered on the basis of Euclidean similarity distances derived from rules of similarity measures. Such clusters can be extremely useful in quantitative structure–property relations as any representative within a cluster would have properties similar to other members of the cluster. Consequently, topological techniques for the characterization of such statistical clusters can be useful in many areas including drug research. We compute distance-based topological indices of the statistical cluster networks such as Szeged, Padmakar–Ivan and Mostar indices which provide different measures such as centrality, peripherality, and other properties of the corresponding networks.