Newest papers

EMBER: Predicting TCR--pMHC Binding Without a Docked Complex

Lin Ha, Chen Zhou, Haozheng Luo, Han Liu, Xiaoyu Zhang, Yan Chen · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.27334 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

Proponemos EMBER, un marco basado en la estructura para la predicción de la vinculación TCR-pMHC que elimina la necesidad de un complejo predicado, y con ello el error de posicionamiento que domina tales predicciones. Los predictores tradicionales basados en la estructura utilizan la estructura compleja co-foldada del receptor y el antígeno e interpretan la vinculación de su superficie de contacto. Por lo tanto, cada característica a continuación hereda cualquier desviación de la herramienta de predicción de la estructura introducida. Metodológicamente, EMBER dobla el TCR y el pMHC de forma independiente, los codifica como gráficos de residuos SE(3) invariantes, y compara sus representaciones aprendidas sin construir un marco compartido de coordenadas. Por lo tanto, una posición relativa no es una entrada del modelo, por

Machine translation (Meta M2M100); the English original follows.

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.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Metric (unit), Set (abstract data type), Folding (DSP implementation), Algorithm, Artificial intelligence

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange