[No change in original Project Summary] Project Summary Computational docking is an essential tool for drug discovery and design, and it has been essential to the development of many drugs currently in clinical use. Given the growing availability of experimentally determined structures with atomic resolution, its application has become among the most important structure-based methods. The key role played by docking is underscored by the number of publications about it (more than 110k citations in the last 4 years), and its key role in the development of many drugs currently in clinical use. Despite being used now routinely by many researchers for decades, there are still many open challenges that need to be addressed. One of the main ones is the increase of accuracy of the scoring function, which would lower the number of false positives from virtual screenings, and increase the success rates. The other is the need for more efficient computing performance to be able to virtually screen not only the very large virtual libraries that have been developed recently (which can now easily contain even billions of chemicals), but also the new large macromolecular complexes that are being characterized with cryoEM, as well as homology modeling and de novo modeling. The combination of these limitations hinders the applicability of docking to these problems. We are proposing to develop better tools to address these challenges, to improve speed, accuracy, and provide the capability of navigating through large libraries and macromolecular structures.