Hierarchical Analysis of Bioactive Matched Molecular Pairs, Encoded Chemical Transformations, and Associated Substructures

Abstract

The matched molecular pair (MMP) concept has become popular to assess molecular property changes resulting from small chemical modifications and study structure-activity relationships (SARs). In this study, we further extend MMP analysis by introducing an MMP-based hierarchical analysis scheme. Specifically, we report a large-scale analysis of MMPs derived from bioactive compounds following a defined "MMP-transformation-substructure" hierarchy. This makes it also possible to categorize transformations and corresponding substructures on the basis of activity information. MMPs were systematically generated for compounds active against current pharmaceutical targets and stepwise decomposed into transformations and substructures. Surprisingly, most chemical transformations were only associated with single MMPs. Hence, the structural context of transformations was unexpectedly narrow. In addition, nearly half of all substructures were found to exclusively form single-target transformations. Taken together, the results of our analysis provide a detailed view of MMP-transformation-substructure hierarchy and further increase the knowledge base of the MMP approach.