How Similar Are Similarity Searching Methods? A Principal Component Analysis of Molecular Descriptor Space

Abstract

Different molecular descriptors capture different aspects of molecular structures, but this effect has not yet been quantified systematically on a large scale. In this work, we calculate the similarity of 37 descriptors by repeatedly selecting query compounds and ranking the rest of the database. Euclidean distances between the rank-ordering of different descriptors are calculated to determine descriptor (as opposed to compound) similarity, followed by PCA for visualization. Four broad descriptor classes are identified, which are circular fingerprints; circular fingerprints considering counts; path-based and keyed fingerprints; and pharmacophoric descriptors. Descriptor behavior is much more defined by those four classes than the particular parametrization. Using counts instead of the presence/absence of fingerprints significantly changes descriptor behavior, which is crucial for performance of topological autocorrelation vectors, but not circular fingerprints. Four-point pharmacophores (piDAPH4) surprisingly lead to much higher retrieval rates than three-point pharmacophores (28.21% vs 19.15%) but still similar rank-ordering of compounds (retrieval of similar actives). Looking into individual rankings, circular fingerprints seem more appropriate than path-based fingerprints if complex ring systems or branching patterns are present; count-based fingerprints could be more suitable in databases with a large number of repeated subunits (amide bonds, sugar rings, terpenes). Information-based selection of diverse fingerprints for consensus scoring (ECFP4/TGD fingerprints) led only to marginal improvement over single fingerprint results. While it seems to be nontrivial to exploit orthogonal descriptor behavior to improve retrieval rates in consensus virtual screening, those descriptors still each retrieve different actives which corroborates the strategy of employing diverse descriptors individually in prospective virtual screening settings.

Clustering files of chemical structures using the Székely–Rizzo generalization of Ward's method

“…It is noteworthy that it is superior to both TAN, COS and TAN-B, all of which have been used in previous studies of similarity searching. We note also that TAN-B is superior to TAN, despite the fact that several previous studies have suggested that weighted-searching is superior to binary searching [10; 36; 37]; however, our results here are in accord with recent work by Bender et al [38].…”

Quantifying structure and performance diversity for sets of small molecules comprising small-molecule screening collections

“…Significant progress has been made quantifying and visualizing properties of compound sets (26), including methods that relate structure to intuitive notions of shape (27)(28)(29), and similarity fusion methods (30)(31)(32)(33) that describe relationships between sets. Moreover, chemical similarity and diversity analyses continue to progress (34)(35)(36)(37), including studies using Shannon entropy (38) as a measure of structure information among compounds (39)(40)(41), addressing reagent selection (42), database similarity searches (43), and scaffold diversity (44). Entropy-based methods have also been used on assay data to distinguish single-target compounds from those with multitarget effects (45), and to quantify relationships between targets based on K i profiles among sets of common inhibitors (46).…”

Small-Molecule Library Subset Screening as an Aid for Accelerating Lead Identification

“…In this article, we describe how we have used diversity-based library subsets to complement this "screen-all" approach, as our subsets have largely been used to supplement full library screens rather than to replace them. Numerous methods have been described for assessing molecular similarity (e.g., see Bender et al 28 and references therein). Reduced topological representations (Murcko assemblies), ECFPs, and atomic property descriptors (BCUT) have all shown promise in the selection of diverse subsets that retain high levels of hit identification.…”