[Paper Review] AI Developments for T and B Cell Receptor Modeling and Therapeutic Design

Linhui Xie, Aurelien Pelissier+2 more|arXiv (Cornell University)|Jan 23, 2026

vaccines and immunoinformatics approaches0 citations

TL;DR

This paper surveys AI advances for modeling TCRs and BCRs using protein language models, structure-aware methods, and multimodal data, highlighting applications to immunoepidemiology and therapeutic design.

ABSTRACT

Artificial intelligence (AI) is accelerating progress in modeling T and B cell receptors by enabling predictive and generative frameworks grounded in sequence data and immune context. This chapter surveys recent advances in the use of protein language models, machine learning, and multimodal integration for immune receptor modeling. We highlight emerging strategies to leverage single-cell and repertoire-scale datasets, and optimize immune receptor candidates for therapeutic design. These developments point toward a new generation of data-efficient, generalizable, and clinically relevant models that better capture the diversity and complexity of adaptive immunity.

Motivation & Objective

• Summarize data resources and computational challenges in immune receptor analysis.
• Review sequence- and structure-based AI models for TCRs and BCRs and their applications in immunoepidemiology and therapy.
• Discuss emergent generative AI approaches for immune receptor engineering and design.

Proposed method

• Describe protein language models (PLMs) trained on large repertoires to learn embeddings for antibodies and TCRs.
• Compare general-purpose PLMs with antibody- or TCR-specific PLMs and discuss fine-tuning and LoRA-based adaptation.
• Explain structure-aware and multimodal representations that integrate sequence and structure data.
• Outline generative AI approaches (diffusion/flow, autoregressive/MLMs) for sequence design of receptors.
• Summarize case studies in SARS-CoV-2, HIV, influenza illustrating predictive prioritization and design.
• Highlight data resources and benchmarking needs for model development.

Experimental results

Research questions

• RQ1 What data resources and benchmarks support AI modeling of TCRs and BCRs?
• RQ2 How do sequence-based PLMs and structure-aware methods perform for binding prediction, specificity, and developability?
• RQ3 What are the roles and limitations of generative models for designing immune receptors?
• RQ4 How can multimodal representations (sequence + structure) improve receptor design and prediction?
• RQ5 What practical outcomes have AI approaches achieved in immunoepidemiology and therapeutic discovery?

Key findings

• Antibody modeling has advanced rapidly due to large repertoire data and available structures, enabling robust language models, structure predictors, and design pipelines.
• TCR modeling is less mature due to limited data and structural coverage, but multimodal and structure-informed approaches are closing the gap.
• Generative AI enables controllable design of antibodies and TCRs, with sequence-space strategies including MLMs, conditional generation, flow-based models, and reinforcement learning.
• Hybrid representations that fuse sequence and structural information improve paratope localization, affinity prediction, and developability assessments.
• COVID-19 and other pathogen studies demonstrate AI-guided triage, prediction, and prioritization guiding experimental validation and design.