Neighborhood Behavior: Validation of Two‐Dimensional Molecular Similarity as a Predictor of Similar Biological Activities and Docking Scores

Abstract

We have used four large datasets to determine the extent to which Two-Dimensional (2D) structural similarity of chemical compounds predicts how similarly they bind to proteins. Structures of 1750, 1853, 1377, and 407 inhibitors for trypsin, thrombin, factor Xa (fXa), and urokinase-type Plasminogen Activator (uPA), respectively, with their observed binding affinities were collected from various literature sources. We also obtained calculated binding affinities for the datasets using an in-house docking program. Plots of the differences in affinity for all pairs of compounds versus their 2D similarity values were generated. All datasets with both observed and calculated affinities clearly exhibit structure -activity relationships (neighborhood behavior), though notable differences among plots are also observed. Guidelines for application of 2D similarity in structure-based virtual screening are discussed in the context of the results obtained.

Integrating Structure‐ and Ligand‐Based Virtual Screening: Comparison of Individual, Parallel, and Fused Molecular Docking and Similarity Search Calculations on Multiple Targets

“…In a benchmark investigation, this integrated approach identified approximately 60 % of available hits by processing only about 7 % of a screening database,10 which presents a substantial enrichment of compound recall through sequential screening. Furthermore, ligand similarity has been related to docking scores13 and ligand‐based virtual screening and docking methods have also been extensively compared 14. Such comparisons have frequently suggested superior performance of ligand‐ over structure‐based screening methods 14…”

Similarity methods in chemoinformatics

“…Analysis of such plots provides a simple way of validating the performance of a molecular descriptor for similarity and diversity applications. Initial studies (Dixon & Merz, 2001; Patterson et al, 1996) used QSAR datasets, containing small numbers of structurally related molecules, for the comparison of descriptors, but the utility of the approach on a large scale has been demonstrated by Perekhodtsev (2007). This study involved combining multiple literature sources to produce four large sets of structurally heterogeneous molecules, and showed that excellent Neighborhood Behavior was exhibited by simple 2D fingerprints.…”

Analysis of Neighborhood Behavior in Lead Optimization and Array Design

“…However, Perekhodtsev used four large, structurally diverse sets of serine protease inhibitors in a neighborhood behavior study to validate the use of standard 2D similarities (Accelrys Accord fingerprints and the Tanimoto coefficient) for the prediction of binding free energies (both measured and computed). 22 Analyses based on Patterson plots and on the correlation coefficient r (following Dixon and Merz 16 ) indicated that all four data sets demonstrate neighborhood behavior, i.e., that structurally similar compounds tend to exhibit more similar binding affinities than do those that are less structurally similar. Perekhodtsev's focus on the data set, rather than on the descriptors used to characterize molecules within it, mirrors recent work that seeks to quantify and classify the nature of the underlying structure-activity relationships.…”

“…Most neighborhood behavior studies have used relatively small, and often homogeneous, data sets. However, Perekhodtsev used four large, structurally diverse sets of serine protease inhibitors in a neighborhood behavior study to validate the use of standard 2D similarities (Accelrys Accord fingerprints and the Tanimoto coefficient) for the prediction of binding free energies (both measured and computed) . Analyses based on Patterson plots and on the correlation coefficient r (following Dixon and Merz) indicated that all four data sets demonstrate neighborhood behavior, i.e., that structurally similar compounds tend to exhibit more similar binding affinities than do those that are less structurally similar.…”