Cyber-diagnostic
New AI can spot diseases by reading the immune system’s ‘memory’
The immune system harbors a lifetime’s worth of information about threats it has encountered. Stanford Medicine researchers devised an artificial intelligence algorithm to identify diseases by investigating this immunological Rolodex. Specifically, the tool analyzed the receptors on B and T cells, two immune cell types.
The researchers believe the AI algorithm, abbreviated Mal-ID for “machine learning for immunological diagnosis,” can help diagnose tricky diseases and guide treatments.
To develop this tool, the team assembled a dataset of over 16 million B cell receptor sequences and over 25 million T cell receptor sequences from 593 study participants. (B cell receptors recognize free-floating pathogens, whereas T cell receptors recognize infected cells.) Participants were healthy, representing the control group; diagnosed with COVID-19, HIV, lupus or Type 1 diabetes; or recently inoculated against the flu.
The algorithm was highly successful at identifying who had which disease or a recent flu vaccination.
The research, published Feb. 20, 2025, in Science, found that T cell receptor sequences provided the most relevant information about lupus and Type 1 diabetes. B cell receptor sequences were most informative in identifying HIV or SARS-CoV-2 infection or recent influenza vaccination. In every case, combining the T and B cell results increased the algorithm’s ability to accurately categorize the disease state.
“Mal-ID could help us identify subcategories of particular conditions that could give us clues to what sort of treatment would be most helpful for someone’s disease state,” said Scott Boyd, MD, PhD, the Stanford Professor in Food Allergy and Immunology and co-director of the Sean N. Parker Center for Allergy and Asthma Research.
Boyd shares senior authorship with Anshul Kundaje, PhD, associate professor of genetics and of computer science. Postdoctoral scholar Maxim Zaslavsky, PhD, and graduate student Erin Craig are the lead authors.
The researchers believe the algorithm could quickly be adapted to identify immunological signatures specific to many other diseases and conditions. They are particularly interested in autoimmune diseases such as lupus, which can be difficult to diagnose and treat effectively.
Read the full story here.
