T-cell and B-cell receptors are generally believed to have strong potential for diagnostic applications. However, unlocking this potential requires navigating a complex landscape of vast datasets and intricate biological signals. In a newly published review in Immunoinformatics, the AIRR Community and our CTO Pieter Meysman explore how computational immunology is reshaping the way we classify and analyse Adaptive Immune Receptor Repertoire (AIRR-seq) data.

From TCR Sequencing Data to Diagnostics‍

The paper, titled "Machine learning for the classification of AIRR-seq data for different diagnostic applications," serves as a roadmap for researchers and clinicians. It examines the current state of machine learning algorithms used to identify disease signatures in immune repertoires.

The authors provide a high-level division of current approaches, distinguishing between:

• Repertoire-level features: Analyzing the global properties of the immune system.
• Sequence-level features: Identifying specific motifs or sequences associated with disease states.

The review also tackles the critical issue of data availability, providing an overview of public AIRR datasets currently available for model training—a vital resource for advancing computational immunology.

Expert Perspective‍

ImmuneWatch is proud to contribute to this field-defining work. Our co-founder and CTO Pieter Meysman co-authored a paper on this topic with the fantastic AIRR community:

“In this review, we examine the application of machine learning algorithms for the classification and analysis of AIRR-seq data for different diagnostic applications. We provide a high-level division of current approaches based on their focus on repertoire-level or sequence-level features. We provide an overview of the current state of public AIRR data sets available for model training. Finally, we briefly highlight what lessons can be learned from successful AIRR diagnostic approaches and what hurdles still must be overcome.”

Overcoming the Hurdles

‍While the diagnostic potential is clear, the paper highlights that hurdles remain, particularly regarding the standardisation and interpretability of machine learning models. For research teams, the complexity of setting up these pipelines can often be a bottleneck.

If you are looking to leverage these insights without building a computational infrastructure from scratch, our AIRR-seq analysis service provides the expert support needed to interpret your repertoire data effectively.

Read the full paper here

‍