New White Paper: How Reliable Are Antigen-Specificity Annotations for T-Cell Receptors?

We are excited to announce the release of our latest white paper: "How Reliable Are Antigen-Specificity Annotations for T-Cell Receptors?"

As computational TCR annotation tools become integral to cell therapy development, immunomonitoring, and target discovery, the question of reliability has never been more important. This white paper addresses the concern head-on, providing a comprehensive, data-driven overview of the state of the art.

What's inside

The white paper covers three major categories of TCR-antigen annotation approaches and their respective strengths and limitations:

• Database matching: We explain the four fundamental challenges — match ambiguity, database false positives, cross-reactivity, and TCR publicity bias — that make simple lookup approaches unreliable at repertoire scale.
• Seen epitope algorithms: We detail how modern machine learning methods, led by ImmuneWatch DETECT, overcome these challenges. DETECT won the independent IMMREP23 international benchmark, outperforming approximately 50 competing methods on a fully blinded dataset.
• Unseen epitope prediction: We map the current frontier and ImmuneWatch's strategy for closing the gap through hybrid modelling and large-scale data generation.

Validated beyond the benchmark

The white paper presents four layers of validation for ImmuneWatch DETECT beyond IMMREP23:

• External experimental validation achieving 100% accuracy on selected annotations in a TIL product study.
• MLOps best practices with over 400 independent validation experiments run with every code or database change.
• Patient-level confirmation in a 160-individual COVID cohort, demonstrating diagnostic-grade annotation accuracy.
• 14+ peer-reviewed publications from independent research groups using DETECT or its predecessor across oncology, autoimmunity, infectious disease, and transplant biology.

Who should read this

This white paper is essential reading for teams developing cell therapies, running clinical trials with immune monitoring endpoints, or exploring TCR-based diagnostics. It equips you with the scientific evidence to confidently integrate computational TCR annotation into your pipeline.