Related Experiment Video
Updated: Apr 3, 2026
09:32
Immunopeptidomics: Isolation of Mouse and Human MHC Class I- and II-Associated Peptides for Mass Spectrometry Analysis
Published on: October 15, 2021
16.0K
A high-speed attention network for MHC-bound peptide identification and 3D modeling
Coos A B Baakman1, Giulia Crocioni2, Cunliang Geng2
1Medical BioSciences Department, Radboud University Medical Center, 6525 GA Nijmegen, the Netherlands.
Cell Reports Methods
|April 2, 2026
Summary
SwiftMHC is a new computational framework that models peptide-MHC interactions and predicts binding affinity. It offers rapid, accurate predictions and 3D structure generation, accelerating cancer immunotherapy research.
Area of Science:
- Computational biology
- Immunoinformatics
- Structural biology
Background:
- Peptide-MHC (pMHC) complexes are crucial for T-cell recognition in adaptive immunity.
- Accurate modeling of pMHC binding is essential for epitope discovery in cancer immunotherapy.
- Existing tools often face limitations in speed or accuracy for pMHC modeling.
Purpose of the Study:
- To develop an ultra-fast and accurate structure-based framework for pMHC modeling and binding affinity prediction.
- To improve the speed and scalability of pMHC analysis for epitope discovery.
- To generate high-resolution 3D structures of pMHC complexes.
Main Methods:
- Developed SwiftMHC, a deep learning framework trained on physics-derived synthetic data.
- Utilized a structure-based approach for pMHC modeling and binding affinity prediction.
- Benchmarked against crystallographic data and leading computational tools (netMHCpan, MHCflurry, AlphaFold2-finetune).
Main Results:
- SwiftMHC predicts pMHC binding affinities in 0.009 seconds per case (batch mode, A100 GPU).
- Achieved competitive accuracy in binding affinity prediction compared to leading sequence-based tools.
- Generated all-atom 3D pMHC structures with a median Cα-RMSD of 1.32 Å.
- Demonstrated comparable or superior structural modeling accuracy to AlphaFold2-finetune at lower computational cost.
Conclusions:
- SwiftMHC provides a significant advancement in speed and accuracy for pMHC modeling and binding affinity prediction.
- The framework enables high-throughput scalability for accelerating epitope discovery in cancer immunotherapy.
- SwiftMHC integrates structural insights with computational efficiency for robust pMHC analysis.
