Leveraging non-structural data to predict structures of protein–ligand complexes

Over the past fifty years, tremendous effort has been devoted to computational methods for predicting properties of ligands that bind macromolecular targets, a problem critical to rational drug design. Such methods generally fall into two categories: physics-based methods, which directly model ligand interactions with the target given the target’s three-dimensional (3D) structure, and ligand-based methods, which predict ligand properties given experimental measurements for similar ligands. Here we present a rigorous statistical framework to combine these two sources of information. We develop a method to predict a ligand’s pose—the 3D structure of the ligand bound to its protein target—that leverages a widely available source of information: a list of other ligands that are known to bind the same target but for which no 3D structure is available. This combination of physics-based and ligand-based modeling improves upon state-of-the-art pose prediction accuracy across all major families of drug targets. As an illustrative application, we predict binding poses of antipsychotics and validate the results experimentally. Our statistical framework and results suggest broad opportunities to predict diverse ligand properties using machine learning methods that draw on physical modeling and ligand data simultaneously.