Construction and use of fragment-augmented molecular Hasse diagrams

Collections of molecules can be organized in many different ways based on substructures that are common to two or more of the molecules. The article describes a method that builds on the ideas of partial orders and Hasse diagrams and which organizes molecules in a particularly simple and natural way using only sub- and superstructure relations. The method outputs the original molecule collection together with common substructures and a set of relations between fragments and molecules. The result is a complete deconstruction of the original structures into those fragments or building blocks that are shared between two or more molecules. Scaffolds for the R-group analyses that can be performed on the data set are automatically detected. Cyclic and linear substituents are treated in the same way. No rules are incorporated that express any form of domain expertise or judgment. The method should be useful for library profiling, data set navigation, fragment-based screening, identification of activity cliffs, and identification of library subsets that are amenable to fragment-based QSAR.