Systematic analysis of structural and activity relationships between conventional hierarchical and analog series-based scaffolds

摘要

The concept of molecular scaffolds is widely applied in medicinal and computational chemistry to represent core structures of compounds and series. A hierarchical organization of compounds has long dominated scaffold design and generation. Recently, so called ‘analog series-based’ (ASB) scaffolds have been introduced as an alternative category of scaffolds. ASB scaffolds are designed to represent analog series and take reaction information into account, and do not follow a molecular hierarchy. We report a large-scale comparison of ASB scaffolds representing more than 15?000 analog series with activity against more than 1200 targets and their corresponding hierarchical scaffolds. Most ASB and conventional hierarchical scaffolds were structurally distinct. However, many ASB scaffolds contained conventional scaffolds as substructures or shared smaller substructures with these scaffolds. Although ASB scaffolds and their corresponding hierarchical scaffolds often shared the same target annotations, ASB scaffolds further distinguished between closely related compound series with different activities that yielded the same conventional scaffolds. Taken together, the findings reported herein reveal that ASB scaffolds further extend current core structure representations for the analysis of structure–activity relationships.