RangeFinder predicts chemical property ranges from protein pockets
RangeFinder: Predicting Ligand Property Ranges from Protein Pockets
Machine Learning
Summary
Proteins interact with small molecules called ligands, but finding which ligands fit a protein is hard because there are so many possible molecules. The authors present RangeFinder, a machine learning tool that looks at the protein's shape and chemistry to predict the possible ranges of important properties of ligands that might bind. This helps reduce the number of molecules to consider before trying more expensive steps like docking or making the ligands. RangeFinder works quickly and better than other methods tested on real protein data.
What this means in practice
- •For drug discovery teams: Filter chemical libraries to focus on ligands whose properties match a given protein pocket before running expensive docking simulations.
- •For biotech computational chemists: Prioritize the synthesis or computational generation of ligands by predicting the property ranges compatible with a target protein site.
Authors
Alan Nadelsticher Ruvalcaba, Chris Alvin, Supratik Mukhopadhyay, Adam Bess
Abstract
Identifying ligands for a protein target requires searching a vast chemical space for a comparatively small set of compatible molecules. Existing machine learning approaches primarily address this problem by generating candidate ligands or scoring individual protein-ligand pairs, but provide limited information about the broader chemical space compatible with a binding pocket. We introduce RangeFinder, an SE(3)-equivariant transformer pocket-level model, trained and evaluated on BioLiP2 data, that predicts ranges over nine physicochemical ligand properties directly from pocket structure, chemistry, and evolutionary context. Using an intersection-over-union objective and evaluation metric (range-IoU), RangeFinder achieves a range-IoU of $0.464 \pm 0.002$, outperforming all deployable baselines while requiring only $12.2$ ms per pocket. We also quantify our method's ability to reduce the ligand search space while retaining compatible binders; in this regard, RangeFinder achieves a Reduction-Recall F-Score of $0.543 \pm 0.011$ --- once again outperforming all deployable baselines. Ablation experiments indicate that RangeFinder's performance depends more strongly on the density and quality of ligand supervision than on any individual input modality or specialized training strategy. Together, these results position RangeFinder as an inexpensive and effective molecular filtering prior, with demonstrated potential in narrowing the ligand search space of target pockets before docking, generation, or synthesis. The code for this project can be found at: https://github.com/alannadels/RangeFinder.