Scaffold explorer: an interactive tool for organizing and mining structure-activity data spanning multiple chemotypes.

We introduce Scaffold Explorer, an interactive tool that allows medicinal chemists to define hierarchies of chemical scaffolds and use them to explore their project data. Scaffold Explorer allows the user to construct a tree, where each node corresponds to a specific scaffold. Each node can have multiple children, each of which represents a more refined substructure relative to its parent node. Once the tree is defined, it can be mapped onto any collection of compounds and be used as a navigational tool to explore structure-activity relationships (SAR) across different chemotypes. The rich visual analytics of Scaffold Explorer afford the user a "bird's-eye" view of the chemical space spanned by a particular data set, map any physicochemical property or biological activity of interest onto the individual scaffold nodes, serve as an aggregator for the properties of the compounds represented by these nodes, and quickly distinguish promising chemotypes from less interesting or problematic ones. Unlike previous approaches, which focused on automated extraction and classification of scaffolds, the utility of the new tool rests on its interactivity and ability to accommodate the medicinal chemists' intuition by allowing the use of arbitrary substructures containing variable atoms, bonds, and/or substituents such as those employed in substructure search.