Relational database driven two-dimensional chemical graph analysis.

This manuscript presents a method for pre-computing and storing molecular features or ''scaffolds'' that can be used for rapid clustering of diverse compound sets within the context of a relational database based on hierarchies of scaffold structures. In addition, a method for rapid structure-based profiling of a large compound collection is demonstrated. Pre-organizing compounds by shared structural features in this way facilitates the merger of chemical features and biological data within a relational database.