Integrating the Structure–Activity Relationship Matrix Method with Molecular Grid Maps and Activity Landscape Models for Medicinal Chemistry Applications

This is the authors' abstract. We don't add key points for this paper.

High Resolution Image Download MS PowerPoint Slide The structure–activity relationship (SAR) matrix (SARM) methodology was originally developed to systematically extract structurally related compound series from data sets of any composition, visualize SAR patterns, and generate virtual candidate compounds. The approach is based upon a dual fragmentation variant of the matched molecular pair formalism. Compound data sets typically yield multiple SARMs that contain unique subsets of structural analogs and virtual candidates complementing existing series. SARM-specific activity predictions make it possible to prioritize virtual analogs for synthesis. The SARM design is intuitive and reminiscent of conventional R-group tables, although the underlying data structure is more complex. Navigating multiple SARMs in parallel can be challenging, depending on the data sets under investigation. Therefore, in this work, we further extend the SARM approach through integration of matrices with newly designed molecular grid maps and activity landscape representations, which provide complementary views of compound relationships and SARs. Moreover, a grid map provides a global view of SARM information including existing compounds, virtual candidates, and associated properties. Grid maps preserve the origin of compounds such that corresponding SARMs can be concomitantly analyzed. In their current implementation, second-generation SARMs make it possible to comprehensively organize and explore large data sets, visualize SARs, and select candidate compounds for practical applications.