Explainable Physicochemical Determinants of Protein–Ligand Binding via Non-Covalent Interactions

Zhaohan Meng^1† · Zhen Bai^2† · Ke Yuan^3,4 · Jaime H. Cheah⁵ · Wei Jiang⁵ · Adam Skepner⁵ · Kevin J. Leahy⁶ · Iadh Ounis¹
William Oldham^6,7,8* · Zaiqiao Meng^1,9* · Hao Xu^6* · Joseph Loscalzo^6,8

¹ School of Computing Science, University of Glasgow

² School of Life Science and Technology, Institute of Science Tokyo

³ School of Cancer Sciences, University of Glasgow

⁴ Cancer Research UK Scotland Institute

⁵ Center for the Development of Therapeutics, The Broad Institute of MIT and Harvard

⁶ Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Harvard University

⁷ Department of Medicine, The Warren Alpert Medical School of Brown University

⁸ The Broad Institute of MIT and Harvard

⁹ Language Technology Lab, University of Cambridge

† Equal contribution; * Corresponding authors
William Oldham (william_oldham@brown.edu), Zaiqiao Meng (zm324@cam.ac.uk), Hao Xu (haxu@bwh.harvard.edu),

Preprint Source Code Interactive UI

News

[March 2026] ⛳ Our preprint is now available on bioRxiv.
[Feb 2026] 🚀 ExplainBind interactive UI released on Hugging Face Space.

Overview

ExplainBind is an interaction-aware framework for protein–ligand binding (PLB) prediction. It supervises token-level cross-attention using non-covalent interaction maps derived from curated PDB protein–ligand complexes in InteractBind. By aligning model attention with physically grounded interaction signals, ExplainBind transforms PLB prediction from a black-box reasoning into a chemistry-grounded process suitable for large-scale screening.

Demo UI

Explore ExplainBind through an interactive web-based demo. The interface visualises token-level interaction attention maps and enables qualitative inspection of predicted binding mechanisms.

🚀 Launch on Hugging Face

If the embedded demo fails to load, please use the external link above.

Acknowledgement

This work was supported in part by the National Institutes of Health under grants HL155107 and HL166137, and by the American Heart Association through the MERIT Award AHA1185447 awarded to J.L. K.Y. acknowledges support from Cancer Research UK (EDDPGM-Nov21/100001; DRCMDP-Nov23/100010; and core funding to the CRUK Scotland Institute, A31287), the Biotechnology and Biological Sciences Research Council (BB/V016067/1), Prostate Cancer UK (MA-TIA22-001), and the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 101016851).