From the manuscript “DOCKSTRING: Easy Molecular Docking Yields Better Benchmarks for Ligand Design”, Miguel García-Ortegón, Sergio Bacallado, et al [J. Chem. Inf. Model. 2022, 62, 3486−3502]¹ have developed a robust Machine Learning process that consists of: (1) an open-source Python package for straightforward computation of docking scores, (2) an extensive dataset of docking scores and poses of more than 260,000 molecules for 58 medically relevant targets, and (3) a set of pharmaceutically relevant benchmark tasks such as virtual screening or de novo design of selective kinase inhibitors. As part of the publication, the authors have made the data set used to develop their algorithm available for download. With the gracious assistance of the lead author, the publicly available data has been repackaged to work with the StarDrop Pose Generation Interface (PGI) and the AutoDock Vina engine (https://vina.scripps.edu/.)² StarDrop’s universal Pose Generation Interface provides a seamless link between StarDrop and 3D docking, and alignment models run on most major computational chemistry platforms. This enables chemists to get feedback on their compound designs from 3D models in real-time, while computational chemists can easily publish their validated docking and alignment models to StarDrop users via the Pose Generation Server.

To learn more about the StarDrop Pose Generation Interface, please visit our Optibrium Community site to see a short video.

Version

This material is for any version of StarDrop (Windows® or Mac®) that has the Pose Generation Interface installed.

Installation

To use the docking models, you will first need to have the StarDrop Pose Generation Interface installed. To access the PGI, please contact StarDrop Support for assistance. The models themselves and the associated meta.txt file required can be downloaded from the following link: https://optibrium.com/downloads/StarDrop_PGI_targets.zip

A CSV file containing the co-crystalized ligand, the RSCB PDB ID,³ a link to the RSCB database, the DUD-E Target,⁴ and a model description for each of the targets can be downloaded here: https://optibrium.com/downloads/StarDrop_Dockstring_Targets.csv

The CSV file can be imported into StarDrop to be used as a basis for evaluating the docking models and learning how to use the Pose Generation Interface.

1) Miguel García-Ortegón, Gregor N. C. Simm, Austin J. Tripp, José Miguel Hernández-Lobato, Andreas Bender, and Sergio Bacallado, DOCKSTRING: Easy Molecular Docking Yields Better Benchmarks for Ligand Design, Journal of Chemical Information and Modeling (2022) 62(15), pp. 3486-3502. DOI: 10.1021/acs.jcim.1c01334
2) O. Trott, A. J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, Journal of Computational Chemistry 31(2) (2010) pp. 455-461. DOI: 10.1002/jcc.21334
3) H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig, I.N. Shindyalov, P.E. Bourne. The Protein Data Bank Nucleic Acids Research, (2000) 28. Pp. 235-242. DOI 10.1093/nar/28.1.235
4) a) Mysinger M.M., Carchia M, Irwin J.J., Shoichet B.K., Directory of Useful Decoys, Enhanced (DUD-E): Better Ligands and Decoys for Better Benchmarking, Journal of Medicinal Chemistry, (2012), 55(14), pp. 6582-6594. DOI 10.1021/jm300687e; b) Huang, N., Shoichet, B.K., and Irwin, J.J., Benchmarking Sets for Molecular Docking, Journal of Medicinal Chemistry, (2006), 49(23), pp. 6789-6801. doi 10.1021/jm0608356