Related Experiment Video
Updated: Jan 17, 2026
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.7K
Parametrically guided design of beta barrels and transmembrane nanopores using deep learning
David E Kim1,2,3, Joseph L Watson1,2,3, David Juergens1,2,3
1Department of Biochemistry, University of Washington, Seattle, WA 98195.
Summary
Parametric design methods, enhanced by deep learning, now enable precise control over beta barrel protein structures. This breakthrough simplifies the creation of novel protein designs, including functional transmembrane nanopores.
Area of Science:
- Protein engineering
- Computational biology
- Biophysics
Background:
- Global parameterization successfully guides coiled coil protein design.
- Beta barrel protein design using 2D blueprints is complex and requires expertise, limiting control over global shape.
Purpose of the Study:
- To generalize parametric design for beta barrel structures using deep learning.
- To achieve precise control over beta barrel protein shape and function.
Main Methods:
- Parametric generation of beta barrel backbones.
- Incorporation of backbone irregularities using RoseTTAFold-based methods (RFjoint inpainting and RFdiffusion).
- In silico and experimental validation of designed beta barrels.
Main Results:
- Successful design of beta barrels with high accuracy, confirmed by X-ray crystallography.
- De novo design of transmembrane nanopores with tunable conductances (200-500 pS).
- High success rates across various beta sheet parameterizations.
Conclusions:
- Parametric generation combined with deep learning simplifies and enhances beta barrel protein design.
- This approach offers precise control over protein shape for specific functions, like nanopore creation.
- The method makes complex protein design more accessible and specifiable.
