Research Seminar –

Abstract. Synthetic methods to access compound collections based on “privileged scaffolds” for a diverse array of biologically relevant targets are highly desirable. Therefore, the ability to interconvert between (often heterocyclic) scaffolds in a functionalized molecule is a topic of intense interest: such transformations have been termed “skeletal editing.” Skeletal edits such as single-atom insertion and deletion reactions can access similar structures with drastically altered molecular properties. We propose to develop computational tools that will enable medicinal chemists to use skeletal edits to (i) explore structural hypotheses (i.e., of a protein binding site) and (ii) target optimized molecular properties. We will use data science approaches to evaluate substrates, enumerate all possible skeletal edits, and then prioritize the selection of transformations to explore structure and pharmacophore space or optimize for desired physical properties while suppressing others. Our approach will assist in selecting compounds with similar reactivity to target compounds with enhanced binding affinities during the drug discovery phase without resorting to a new synthetic route. Furthermore, as late-stage modification reactions continue to be developed, we aim to assist in method development for skeletal editing reactions for existing and new heterocyclic substrates based on the learned reactivity from physical-meaning molecular and atomistic properties from a pool of diverse, aromatic heterocyclic substrates.