Dean G. Brown; Jonas Boström
J. Med. Chem., 2018, 61, 21, 9442-9468
https://doi.org/10.1021/acs.jmedchem.8b00675

Abstract

An analysis of 66 published clinical candidates from Journal of Medicinal Chemistry has been conducted to shed light on which lead generation strategies are most frequently employed in identifying drug candidates. The most frequent lead generation strategy (producing a drug candidate) was based on starting points derived from previously known compounds (43%) followed by random high throughput screening (29%). The remainder of approaches included focused screening, structure-based drug design (SBDD), fragment-based lead generation (FBLG) and DNA-encoded library screening (DEL). An analysis of physical chemical properties on the hit-to-clinical pairs show an average increase in molecular weight (ΔMW +85) but no change in lipophilicity (ΔclogP -0.2), although exceptions are noted. The majority (>50%) of clinical candidates were found to be structurally very different from their starting point, as well as being more complex. Finally, several reports of non-covalent scaffolds modified by a covalent warhead using SBDD approaches are discussed.