Unraveling Long COVID: New Discoveries and the Promise of AI
In a recent ISB Research Roundtable, ISB President Dr. Jim Heath shared new findings connecting Long COVID to immune dysfunction and autoimmune diseases, and explained how artificial intelligence could accelerate the search for answers — and potentially new treatments — for millions of people living with the condition.
Long COVID remains one of the most perplexing health challenges to emerge from the COVID-19 pandemic. Millions of Americans experience lingering symptoms ranging from debilitating fatigue and brain fog to respiratory problems and cardiovascular complications. Yet scientists are still working to understand why some people develop the condition, why symptoms vary so dramatically, and how best to treat it.
During an October 13, 2026, Research Roundtable, ISB President Dr. Jim Heath explored two promising avenues of research: discoveries about the immune system that may help explain Long COVID’s underlying biology, and the use of artificial intelligence to accelerate the search for effective treatments.
Heath, who leads the Pacific Northwest hub of the National Institutes of Health’s RECOVER Initiative, has studied the long-term effects of COVID-19 since early in the pandemic. The national initiative has enrolled more than 15,000 participants, collecting extensive biological and clinical data to understand the condition better.
When the Immune System Gets Misdirected
One of Long COVID’s central mysteries is why the immune system sometimes fails to return to normal after an infection.
Heath described ISB research into autoantibodies — antibodies that recognize the body’s own proteins rather than targeting foreign invaders. In earlier studies, his team found that people who produced certain autoantibodies during acute COVID-19 infection were more likely to develop Long COVID.
More recent investigations have begun to reveal why some individuals produce these autoantibodies while others do not.
By closely examining immune cells from a small group of patients, researchers identified a population of B cells — immune cells responsible for producing antibodies — that appeared unusually sensitive to stimulation. Rather than following the immune system’s typical process for producing highly targeted antibodies, these cells can take a faster developmental route that produces less precise immune responses.
Heath described these cells as being on a “hair trigger,” producing autoantibodies that may divert immune resources away from more effective responses to infection.
The findings also revealed genetic connections between these immune responses and several autoimmune diseases, including lupus, multiple sclerosis, rheumatoid arthritis, and type 1 diabetes.
Another discovery involved the Epstein-Barr virus (EBV), a common virus that can leave lasting changes in certain immune cells. Researchers found evidence that EBV-associated reprogramming of B cells may contribute to the same heightened immune sensitivity, providing another potential connection between Long COVID and autoimmune disease.
Together, these findings suggest that some patients with Long COVID may share underlying biological mechanisms with people who have other autoimmune conditions.
That raises an intriguing possibility: Treatments developed for certain autoimmune diseases might eventually prove useful for specific groups of Long COVID patients. But Heath emphasized that more research is needed before those connections can be translated into clinical care.
Using AI to Accelerate Discovery
The second half of Heath’s presentation focused on a different challenge: how to extract meaningful answers from the enormous amount of biological data researchers have collected.
ISB’s Long COVID studies include measurements of proteins, metabolites, immune cells, genetic variations, and other biological indicators across multiple time points. The national RECOVER dataset is substantially larger, encompassing thousands of participants and years of clinical observations.
Analyzing such complex information traditionally requires specialized computational expertise and can take months or years.
To help overcome that barrier, ISB researchers have developed ChatISB, an AI-powered research environment that allows scientists to interrogate large biological datasets using everyday language.
Rather than writing computer code, researchers can ask questions, generate analyses and visualizations, and explore connections across different types of data. Tasks that once required extensive programming can sometimes produce initial results in minutes.
The goal is not to replace scientific expertise, but to make sophisticated data analysis accessible to more researchers — including physicians and specialists who understand particular diseases but may lack advanced computational training.
Building on this approach, ISB is working with NIH researchers to organize collaborative Long COVID research hackathons. These intensive sessions will bring together specialists from different medical disciplines, computational scientists, and patient advocates to investigate the RECOVER data.
By combining diverse expertise with AI-enabled analysis, the teams hope to identify distinct biological subtypes of Long COVID, uncover potential biomarkers, and generate hypotheses about treatments that might benefit specific groups of patients.
The effort also represents a broader experiment in how large-scale biomedical research might be conducted in the future.
Questions from the Community
The presentation concluded with an extensive audience Q&A, reflecting both the complexity of Long COVID and the urgent need for answers.
Participants asked about the relationship between Long COVID and other post-viral illnesses, the possible role of Epstein-Barr virus, COVID-19 vaccination, antiviral medications, and the challenges of identifying reliable biomarkers.
Other questions focused on whether emerging discoveries might lead to new treatments, including therapies that target particular immune cells.
Heath emphasized that Long COVID is not a single, uniform condition. Different patients may experience different biological disruptions, making it unlikely that one diagnostic test or treatment will work for everyone.
Understanding those differences — and matching patients with therapies that address their particular underlying biology — remains a central goal of the research.
While many questions remain unanswered, the combination of increasingly detailed biological measurements, large patient cohorts, and powerful new analytical tools offers promising paths forward.
Watch the full Research Roundtable presentation: Understanding Long COVID: New Discoveries and the Promise of AI
About Research Roundtable
ISB’s Research Roundtable is a public-facing series that brings leading scientists and the community together to explore important advances in biomedical research. Designed for nonscientists, the free virtual events offer accessible explanations of complex scientific topics and opportunities for participants to ask questions directly of researchers.
