Mechanism-based bipartite matching molecular series graphs to identify structural modifications of receptor ligands that lead to mechanism hopping

Abstract

The rationalization of structural features that distinguish between different mechanisms of action of ligands active against a given receptor is of high importance in medicinal chemistry and drug design. We have adapted a bipartite molecular network structure that organizes compound datasets on the basis of substructure relationships to incorporate mechanism of action information. The resulting data structure readily identifies subsets of ligands with different mechanisms of action that display well-defined structural relationships. From the structural subset organization of the graph, structural changes that lead to mechanism hopping (i.e., a transition from one mechanism of action to another) can be directly selected, as demonstrated for different classes of receptor ligands. For medicinal chemistry applications, the ability to immediately access structural modifications that distinguish ligands having different mechanisms of action is a key aspect of the methodology introduced herein. The knowledge of substituents in receptor ligands that trigger mechanistic changes can be utilized for compound design.