DecoderTCR: Compositional Pretraining and Entropy-Guided Decoding for TCR-pMHC Interactions

Boqiao Lai, Melissa Englund, Ramit Bharanikumar, Isabel Nocedal, Ali Davariashtiyani, Jason Perera, Aly A Khan
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:62206-62229, 2026.

Abstract

Modeling recognition between T-cell receptors (TCRs) and peptide-MHC (pMHC) complexes is a fundamental challenge in computational immunology, constrained by sparse paired interaction data relative to abundant unpaired sequences. We introduce DecoderTCR, a masked language model framework that addresses this through two contributions: (1) a compositional continual pre-training curriculum that learns component representations from marginal data before refining cross-chain dependencies, and (2) Iterative Entropy-Guided Refinement (IEGR), a non-autoregressive decoding algorithm that resolves high-confidence positions first to provide context for uncertain regions. On held-out benchmarks, DecoderTCR achieves 0.96 AUROC for zero-shot pMHC binding prediction and 0.76 AUROC for epitope-specific TCR recognition, approaching supervised baselines without epitope-specific training. Learned representations recover structural contacts without coordinate supervision, and generated sequences exhibit realistic recombination statistics. Experimental validation across two rounds of wet-lab screening reveals a prediction-generation gap that can be narrowed via a lab-in-the-loop paradigm for TCR design.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-lai26b, title = {{D}ecoder{TCR}: Compositional Pretraining and Entropy-Guided Decoding for {TCR}-p{MHC} Interactions}, author = {Lai, Boqiao and Englund, Melissa and Bharanikumar, Ramit and Nocedal, Isabel and Davariashtiyani, Ali and Perera, Jason and Khan, Aly A}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {62206--62229}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/lai26b/lai26b.pdf}, url = {https://proceedings.mlr.press/v306/lai26b.html}, abstract = {Modeling recognition between T-cell receptors (TCRs) and peptide-MHC (pMHC) complexes is a fundamental challenge in computational immunology, constrained by sparse paired interaction data relative to abundant unpaired sequences. We introduce DecoderTCR, a masked language model framework that addresses this through two contributions: (1) a compositional continual pre-training curriculum that learns component representations from marginal data before refining cross-chain dependencies, and (2) Iterative Entropy-Guided Refinement (IEGR), a non-autoregressive decoding algorithm that resolves high-confidence positions first to provide context for uncertain regions. On held-out benchmarks, DecoderTCR achieves 0.96 AUROC for zero-shot pMHC binding prediction and 0.76 AUROC for epitope-specific TCR recognition, approaching supervised baselines without epitope-specific training. Learned representations recover structural contacts without coordinate supervision, and generated sequences exhibit realistic recombination statistics. Experimental validation across two rounds of wet-lab screening reveals a prediction-generation gap that can be narrowed via a lab-in-the-loop paradigm for TCR design.} }
Endnote
%0 Conference Paper %T DecoderTCR: Compositional Pretraining and Entropy-Guided Decoding for TCR-pMHC Interactions %A Boqiao Lai %A Melissa Englund %A Ramit Bharanikumar %A Isabel Nocedal %A Ali Davariashtiyani %A Jason Perera %A Aly A Khan %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-lai26b %I PMLR %P 62206--62229 %U https://proceedings.mlr.press/v306/lai26b.html %V 306 %X Modeling recognition between T-cell receptors (TCRs) and peptide-MHC (pMHC) complexes is a fundamental challenge in computational immunology, constrained by sparse paired interaction data relative to abundant unpaired sequences. We introduce DecoderTCR, a masked language model framework that addresses this through two contributions: (1) a compositional continual pre-training curriculum that learns component representations from marginal data before refining cross-chain dependencies, and (2) Iterative Entropy-Guided Refinement (IEGR), a non-autoregressive decoding algorithm that resolves high-confidence positions first to provide context for uncertain regions. On held-out benchmarks, DecoderTCR achieves 0.96 AUROC for zero-shot pMHC binding prediction and 0.76 AUROC for epitope-specific TCR recognition, approaching supervised baselines without epitope-specific training. Learned representations recover structural contacts without coordinate supervision, and generated sequences exhibit realistic recombination statistics. Experimental validation across two rounds of wet-lab screening reveals a prediction-generation gap that can be narrowed via a lab-in-the-loop paradigm for TCR design.
APA
Lai, B., Englund, M., Bharanikumar, R., Nocedal, I., Davariashtiyani, A., Perera, J. & Khan, A.A.. (2026). DecoderTCR: Compositional Pretraining and Entropy-Guided Decoding for TCR-pMHC Interactions. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:62206-62229 Available from https://proceedings.mlr.press/v306/lai26b.html.

Related Material