TCR-EML: Explainable Model Layers for TCR-pMHC Prediction
By: Jiarui Li , Zixiang Yin , Zhengming Ding and more
Potential Business Impact:
Shows how immune cells recognize bad cells.
T cell receptor (TCR) recognition of peptide-MHC (pMHC) complexes is a central component of adaptive immunity, with implications for vaccine design, cancer immunotherapy, and autoimmune disease. While recent advances in machine learning have improved prediction of TCR-pMHC binding, the most effective approaches are black-box transformer models that cannot provide a rationale for predictions. Post-hoc explanation methods can provide insight with respect to the input but do not explicitly model biochemical mechanisms (e.g. known binding regions), as in TCR-pMHC binding. ``Explain-by-design'' models (i.e., with architectural components that can be examined directly after training) have been explored in other domains, but have not been used for TCR-pMHC binding. We propose explainable model layers (TCR-EML) that can be incorporated into protein-language model backbones for TCR-pMHC modeling. Our approach uses prototype layers for amino acid residue contacts drawn from known TCR-pMHC binding mechanisms, enabling high-quality explanations for predicted TCR-pMHC binding. Experiments of our proposed method on large-scale datasets demonstrate competitive predictive accuracy and generalization, and evaluation on the TCR-XAI benchmark demonstrates improved explainability compared with existing approaches.
Similar Papers
Rational Multi-Modal Transformers for TCR-pMHC Prediction
Computational Engineering, Finance, and Science
Helps doctors find better cancer treatments.
Enhancing TCR-Peptide Interaction Prediction with Pretrained Language Models and Molecular Representations
Quantitative Methods
Helps doctors create better cancer treatments.
T-cell receptor specificity landscape revealed through de novo peptide design
Quantitative Methods
Helps design better ways to fight sickness.