Computer-aided pattern scoring – A multitarget dataset-driven workflow to predict ligands of orphan targets

Abstract

The identification of lead molecules and the exploration of novel pharmacological drug targets are major challenges of medical life sciences today. Genome‐wide association studies, multi-omics, and systems pharmacology steadily reveal new protein networks, extending the known and relevant disease-modifying proteome. Unfortunately, the vast majority of the disease-modifying proteome consists of ‘orphan targets’ of which intrinsic ligands/substrates, (patho)physiological roles, and/or modulators are unknown. Undruggability is a major challenge in drug development today, and medicinal chemistry efforts cannot keep up with hit identification and hit-to-lead optimization studies. New ‘thinking-outside-the-box’ approaches are necessary to identify structurally novel and functionally distinctive ligands for orphan targets. Here we present a unique dataset that includes critical information on the orphan target ABCA1, from which a novel cheminformatic workflow – computer-aided pattern scoring (C@PS) – for the identification of novel ligands was developed. Providing a hit rate of 95.5% and molecules with high potency and molecular-structural diversity, this dataset represents a suitable template for general deorphanization studies.