Ember predicts immune protein binding without complex structural docking
EMBER: Predicting TCR--pMHC Binding Without a Docked Complex
Machine Learning
Summary
Predicting how immune system proteins bind to targets usually requires building a combined 3D model, but mistakes in placing them can cause errors. The authors propose EMBER, which analyzes the two proteins separately and compares their shapes without relying on their combined structure. This means the prediction is less affected by errors and can be computed faster. EMBER performs better than existing methods on tests involving new target peptides and is more consistent even when different folding tools are used.
What this means in practice
- •For biotech software engineers: Integrate EMBER to accurately predict TCR-pMHC interactions without costly complex folding for improved immunotherapy design tools.
- •For computational immunologists: Use EMBER as a reliable method to study TCR-pMHC binding across diverse peptides without structural placement errors affecting results.
Authors
Lin Ha, Chen Zhou, Haozheng Luo, Han Liu, Xiaoyu Zhang, Yan Chen
Abstract
We propose EMBER, a structure-based framework for TCR--pMHC binding prediction that eliminates the need for a predicted docked complex, and with it the placement error that dominates such predictions. Traditional structure-based predictors utilize the co-folded complex structure of the receptor and the antigen and interpret binding from its contact surface. Consequently, every downstream feature inherits any deviation the structure-prediction tool introduced. Methodologically, EMBER folds the TCR and pMHC independently, encodes them as SE(3)-invariant residue graphs, and compares their learned representations without constructing a shared coordinate frame. A relative pose is therefore not an input of the model, so placement error cannot propagate into the prediction. Computationally, EMBER preprocesses $O(N{+}M)$ monomer folds where a complex-based predictor needs $O(NM)$ operations. Experimentally, under epitope-disjoint evaluation with six retrained baselines, EMBER is first on every metric under all four folding tools, by 0.9 to 14.4% in micro-AUROC, 18.5 to 25.3% in micro-AUPRC and 10.6 to 24.1% in MCC. Its ranking is also up to 85% more stable across folding tools.