Accurate de novo design of high-affinity protein-binding macrocycles using deep learning
Problem
Designing macrocyclic binders to therapeutic proteins relies on resource-intensive screening with little control over binding mode, and no robust de novo design approaches existed.
Why it matters
Macrocyclic binders are potential therapeutics against protein targets that are hard to drug.
What the AI did
Introduced RFpeptides, a denoising diffusion-based pipeline that designs macrocyclic binders against protein targets, tested experimentally on four diverse proteins.
Results
Binders with medium-to-high affinity against all four targets; Kd < 10 nM for RbtA; X-ray structures match designs with Cα RMSD < 1.5 Å.
Impact
Demonstrated experimentally validated de novo design of high-affinity macrocycle binders, confirmed by X-ray structures.
Record
| Venue | Nature Chemical Biology |
|---|---|
| Published | 2025 |
| Work type | article |
| Authors | Stephen Rettie, David Juergens, Victor Adebomi, Yensi Flores Bueso, Qinqin Zhao, Alexandria N. Leveille, Andi Liu, Asim K. Bera et al. |
| Citations | 75 (OpenAlex, at import time) |
| References | 70 |
| Model or method | RFpeptides (denoising diffusion-based pipeline) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41589-025-01929-w |
| Open access | hybrid |
| Keywords | macrocycles, diffusion model, de novo protein design, binder, X-ray validation |
Importance rating
significant — thousands to millions affected
Abstract positions macrocyclic binders as potential therapeutics for hard-to-drug targets and claims first robust de novo design approach, a genuinely new capability; no disease or market scale stated.