Virtual docking algorithms are often evaluated on their ability to separate active ligands from decoy molecules. The current state-of-the-art benchmark, the Directory of Useful Decoys (DUD), minimizes bias by including decoys from a library of synthetically feasible molecules that are physically similar yet chemically dissimilar to the active ligands. Since decoys in the DUD are drawn from a library of synthetically feasible molecules the molecules only represent a small subset of chemical space. This restriction prevents the DUD from identifying decoys for several active ligands. In our work we computationally generate virtual decoy sets (VDS) that are chemically reasonable yet not constrained by synthetic feasibility. We demonstrate two advantages of VDS. First, by ignoring synthetic feasibility, we can generate docking benchmarks that are comparable to the DUD while better preserving the underlying physical similarity between the decoys and the binders. Second, by using the background distribution of property-matched VDS docking scores we can normalize docking results in virtual high-throughput screening.

Virtual decoy sets for molecular docking benchmarks and ranking

241st National Meeting (2011)

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