Disease diagnostics using machine learning of immune receptors.

Clinical diagnoses rely on a wide variety of laboratory tests and imaging studies, interpreted alongside physical examination and documentation of symptoms and patient history. However, the tools of diagnosis make little use of the immune system’s internal record of specific disease exposures encoded by the antigen-specific receptors of memory B cells and T cells. We have combined extensive receptor sequence datasets with three different machine learning representations of the contents of immune repertoires to develop an interpretive framework, MAchine Learning for Immunological Diagnosis (Mal-ID) , that screens for multiple illnesses simultaneously. This approach can already reliably distinguish a wide range of disease states, including specific acute or chronic infections, and autoimmune or immunodeficiency disorders, and could contribute to identifying new infectious diseases as they emerge. Importantly, many features of the model of immune receptor sequences are human-interpretable. They independently recapitulate known biology of the responses to infection by SARS-CoV-2 or HIV, and reveal common features of autoreactive immune receptor repertoires, indicating that machine learning on immune repertoires can yield new immunological knowledge.