Multi-omics Model Identifies Chronic Fatigue Syndrome Biomarkers

Researchers from The Jackson Laboratory developed an AI-driven multi-omics platform to identify biomarkers for myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), a disabling condition frequently misdiagnosed due to the absence of reliable laboratory markers. Findings enabled by the new model suggest a path toward more individualized care by detailing how ME/CFS disrupts the connections among gut microbes, immune functioning, and body metabolism.

ME/CFS impacts between 836,000 and 3.3 million people in the United States alone, though many remain undiagnosed. The condition leads to substantial losses in productivity and direct healthcare costs of up to $51 billion annually. Individuals with ME/CFS commonly experience debilitating fatigue, persistent sleep issues, dizziness, pain, and difficulty with daily activities. ME/CFS is often compared to long COVID, as both can emerge following viral infections such as Epstein-Barr virus.

The study drew on data from 249 individuals—153 patients with ME/CFS and 96 healthy controls—collected at the Bateman Horne Center, a recognized site for research into ME/CFS, Long-COVID, and fibromyalgia. Researchers developed a deep neural network models called BioMapAI, which integrated data from the gut microbiome, blood metabolites, immune system profiles, and clinical lab tests over a four-year period.

“Our study achieved 90% accuracy in distinguishing individuals with chronic fatigue syndrome, which is significant because doctors currently lack reliable biomarkers for diagnosis,” said Derya Unutmaz, co-senior author on the study published in Nature Medicine. “Some physicians doubt it as a real disease due to the absence of clear laboratory markers, sometimes attributing it to psychological factors.”

The analysis showed that disruptions in the immune system most accurately predicted how severe a patient’s symptoms would be, while changes in gut microbiota were linked with gastrointestinal, emotional, and sleep problems. By mapping patient-reported symptoms to biological changes, the team identified 12 classes of symptoms reflecting the complex and highly variable nature of ME/CFS. 

The research found that the longer patients had ME/CFS, the more established these biological disturbances became. Compared to healthy controls, ME/CFS subjects had lower levels of butyrate, a fatty acid important for gut health and metabolism, as well as higher levels of tryptophan, benzoate, and inflammatory immune cell activity. The study highlighted the importance of MAIT cells in connecting gut health to broader immune responses, suggesting a significant imbalance in these pathways.

Lead researcher Julia Oh explained, “The microbiome and metabolome are dynamic. That means we may be able to intervene—through diet, lifestyle, or targeted therapies—in ways that genomic data alone can’t offer.” 

The BioMapAI model’s accuracy on outside data sets reinforced the robustness of the discovered biomarkers, encouraging further research and collaboration. By sharing their dataset and tools, the researchers aim to promote more detailed studies and, ultimately, help develop targeted treatments for ME/CFS.