Mining of Randomly Generated Molecular Fragment Populations Uncovers Activity-Specific Fragment Hierarchies

Abstract

We introduce a methodology to analyze random molecular fragment populations and determine conditional probability relationships between fragments. Random fragment profiles are generated for an arbitrary set of molecules, and each observed fragment is assigned a frequency vector. An algorithm is designed to compare frequency vectors and derive dependencies of fragment occurrence. Using calculated dependency values, random fragment populations can be organized in graphs that capture their relationships and make it possible to map fragment pathways of biologically active molecules. For sets of molecules having similar activity, unique fragment signatures are identified. The analysis reveals that random fragment profiles contain compound class-specific information and provides evidence for the existence of activity-specific fragment hierarchies.