Key Questions Answered
Q: What did the study reveal about ME/CFS?
A: The study shows that myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) disrupts critical interactions among the gut microbiome, the immune system, and metabolic pathways. Researchers identified biological signatures that distinguish people with ME/CFS from healthy individuals with high accuracy.
Q: How does the AI platform BioMapAI contribute?
A: BioMapAI integrates thousands of measurements—including gut microbiome profiles, blood laboratory tests, immune-cell data, metabolite levels, and detailed symptom reports—to detect patterns and disruptions characteristic of ME/CFS, supporting more precise, individualized approaches.
Q: Why do these results matter to patients?
A: The findings provide stronger biological evidence for ME/CFS and reveal pathways that may explain specific symptoms. This knowledge could guide dietary, lifestyle, and therapeutic strategies, and may be particularly relevant for people with long COVID who experience similar symptoms.
Summary: An AI-driven multi-omics analysis has mapped how ME/CFS alters the connections between gut bacteria, immune responses, and metabolic products. Using stool, blood, and symptom data, the BioMapAI platform distinguished patients from controls with around 90% accuracy, identifying metabolic deficits, immune dysregulation, and microbiome changes that offer testable targets for future interventions.
Highlights and Key Findings
- AI Classification: BioMapAI classified ME/CFS with roughly 90% accuracy by combining microbiome, immune, and metabolic data.
- Distinct Biological Signatures: Patients exhibited altered tryptophan metabolism, reduced levels of butyrate and other beneficial fatty acids, and pro-inflammatory immune cell activity.
- Clinical Relevance: These multi-omics signatures link to specific symptoms and suggest routes for precision medicine approaches targeting ME/CFS and related post-viral conditions.
Source: Jackson Laboratory
Introduction
Millions of people worldwide live with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), a disabling condition long hampered by a lack of objective diagnostic tests. New research analyzes how ME/CFS disrupts the interplay between gut microbes, immune cells, and metabolic processes, offering a clearer biological picture that could inform future diagnostics and therapies.
Research Overview
The study analyzed data from 249 participants collected over four years, combining gut metagenomic sequencing, plasma metabolomics, immune-cell profiling, routine blood tests, and detailed clinical symptom inventories. The work was led by researchers at The Jackson Laboratory (JAX) in collaboration with clinicians at the Bateman Horne Center. The findings were published in Nature Medicine.
To analyze this complex dataset, the team developed BioMapAI, a supervised deep neural network designed to integrate diverse data types and predict clinical severity. The model connected thousands of data points to reconstruct symptom patterns across domains such as sleep, pain, gastrointestinal function, fatigue, and emotional regulation.
Key Biological Insights
Overall, ME/CFS patients showed widespread disruptions across microbiome, immune, and metabolic networks. Notable findings include lower levels of butyrate—an anti-inflammatory short-chain fatty acid produced by gut bacteria—imbalanced tryptophan and benzoate metabolites, and elevated inflammatory activity in certain T cell subsets, including MAIT and γδ T cells.
Immune-cell profiles were particularly informative in predicting symptom severity, while microbiome features best predicted gastrointestinal symptoms, sleep disturbances, and emotional symptoms. The model also revealed that network disruptions tend to be less extensive in patients within four years of illness onset and become more entrenched in those ill for more than ten years.
Implications for Diagnosis and Treatment
These results strengthen the biological basis of ME/CFS and suggest concrete, testable hypotheses for targeted interventions. Because the microbiome and metabolome are dynamic and responsive to diet, lifestyle, and therapeutic modulation, the study points to possible routes for personalized treatment—options that genomic analysis alone cannot provide.
BioMapAI reproduced many of the key biomarkers in independent external datasets with approximately 80% accuracy, demonstrating consistent signals in fatty acids, immune markers, and metabolite profiles despite differences in data collection methods. This reproducibility supports the robustness of the identified signatures.
An Actionable Dataset
The authors plan to make the dataset available through BioMapAI to support broader research into symptom-specific mechanisms and treatment targets. Because animal models do not capture the full complexity of human neurological, immune, and metabolic interactions, large human multi-omics datasets are essential for advancing therapeutic discovery.
By mapping the connections between gut microbes, the chemicals they produce, and immune responses, the team aims to build a systems-level framework that explains what drives ME/CFS and guides precision medicine approaches.
Authors and Funding
Lead investigators include Derya Unutmaz and Julia Oh, with clinical collaborators Lucinda Bateman and Suzanne Vernon. Additional contributors are Elizabeth Aiken, Ryan Caldwell, Lina Kozhaya, Courtney Gunter, and others. Funding was provided by NIH grant 1U54NS105539.
About this chronic fatigue and microbiome research news
Author: Cara McDonough, Jackson Laboratory
Source: Jackson Laboratory
Contact: Cara McDonough, Jackson Laboratory
Image credit: Neuroscience News
Original Research: “AI-driven multi-omics modeling of myalgic encephalomyelitis/chronic fatigue syndrome” by Derya Unutmaz et al.; Nature Medicine. DOI: 10.1038/s41591-025-03788-3 (closed access).
Abstract
This study introduces BioMapAI, a supervised deep neural network trained on a four-year longitudinal multi-omics dataset from 249 participants. By integrating gut metagenomics, plasma metabolomics, immune profiling, blood laboratory data, and detailed clinical symptoms, BioMapAI predicts clinical severity, identifies symptom-specific biomarkers, and classifies ME/CFS in held-out and external cohorts. An explainable-AI approach generates a connectivity map linking microbial metabolism (for example, short-chain fatty acids, branched-chain amino acids, tryptophan, benzoate), plasma lipids and bile acids, and heightened inflammatory responses in mucosal and inflammatory T cell subsets (MAIT, γδT) that secrete IFN-γ and granzyme A. The model offers systems-level insights, refines prior hypotheses, and proposes mechanisms by which multi-omics dynamics relate to ME/CFS symptoms.