High-throughput virtual screening (HTVS) is often employed at the early stages of a drug discovery campaign to evaluate millions of compounds with the expectation of identifying a chemically novel lead series. Scoring functions are utilized to predict the binding free energies between a small molecule ligand and protein and prioritize those compounds that may be potentially active against the target of interest. However, it is well known that there are limitations to the accuracy of scoring functions, and they are not always capable of providing a reliable relative ranking. We have previously proposed using a tiering approach based on predicted scores that allows for prioritization of groups of compounds to obtain a number suitable for subsequent experimental testing.1 In small to medium-sized databases, we have found success with this approach but with the introduction of ultra-high through-put virtual screening (uHTVS) against virtual libraries containing millions to billions of compounds such as the Enamine data set or ZINC, additional triaging is necessary for even the top tier of ranked compounds. It is well-known that human intervention and the use of “chemical intuition” in choosing compounds to move forward from uHTVSs can greatly improve hit rates. Unfortunately, when millions to billions of compounds are being screened, it is not possible to manually examine each compound pose. Here, we have examined a variety of filters including 1) properties from the Glide docking output, 2) interaction fingerprints, 3) binding pocket shape similarity, 4) exposed surface area, and 5) pose strain to determine the most appropriate triage funnel for top scoring compounds utilizing a range of targets. These guidelines can be applied to prospective uHTVS studies to enhance the likelihood of identifying active chemical matter for the project at hand.

Enriched lead identification through ultra high throughput virtual screening

ACS Spring 2022

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