Over the past 20 years, the convergence of the exponential rise in the number of experimentally determined protein structures, Moore’s law and improvements in docking and scoring algorithms has established virtual screening as a key component of the hit identification toolbox in drug discovery. Given sufficient information, virtual screening has the potential – demonstrably – to deliver chemical starting points for hit-to-lead optimisation in significantly less time and at a radically reduced cost compared to the industry’s workhorse: high-throughput screening.<br/>After a brief introduction to the history and techniques of virtual screening, I will illustrate its utility by means of two examples drawn from my personal experience over nearly two decades. In both cases, the hit derived from virtual screening led to a clinical candidate, one of which is now in Phase II clinical trials. Moreover, in contrast to the current trend towards ultra-high-throughput, or “Giga” virtual screening, these compounds were discovered with modestly-sized databases and computational resources.<br/>I will then outline some of the limitations that still bedevil virtual screening and, finally, look to the future to see where answers to some of these obstacles may be found.

Virtual screening: Personal reflections and future directions

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.