HEALTH
For millions of people, the worst part is not the exhaustion. It is the appointment.
It is sitting across a desk from a doctor who has run the blood tests, found nothing, and has begun, gently, to talk about stress. It is the years — and for people with ME, it is very often years — spent trying to prove that a body that looks normal on paper has stopped working.
This week a team at the University of East Anglia's Norwich Medical School published a paper that lands directly on that wound. Their claim, in the words of lead researcher Professor Dmitry Pshezhetskiy, is startling: "What we discovered is something approaching a biological unifying theory of fatigue."
The paper, published in the Journal of Translational Medicine on September 3, argues that five conditions long treated as entirely separate — myalgic encephalomyelitis / chronic fatigue syndrome, long Covid, post-traumatic stress disorder, rheumatoid arthritis and multiple sclerosis — are converging on the same underlying biological circuitry.
And within 24 hours, some of Britain's most respected genomics researchers had lined up to say that the claim goes considerably further than the evidence.
The puzzle that started it
Start with what clinicians have quietly noticed for decades.
A woman gets a viral infection and never recovers; she is diagnosed with ME. A man survives a car crash and develops PTSD. Another patient's immune system starts attacking her joints — rheumatoid arthritis. A fourth develops multiple sclerosis, in which the immune system strips the insulation from nerve fibres. A fifth catches Covid in 2021 and is still, five years later, unable to climb a flight of stairs.
Five different triggers. Five different specialties. Five different sets of textbooks.
And yet, as Professor Pshezhetskiy points out, the patients describe strikingly similar lives: "overwhelming fatigue, brain fog, poor concentration, disturbed sleep, autonomic dysfunction and a dramatic reduction in everyday functioning."
"We wanted to find out why this is," he said.
Reading the folds, not the letters
The method is where this study gets genuinely interesting — and where the argument starts.
Most genetics reads DNA as a sequence: a very long string of letters, scanned for spelling differences that track with disease. That approach, genome-wide association study or GWAS, has been run on all five of these conditions, and it has not obviously unified them.
The UEA team, working with the Oxford-listed company Oxford BioDynamics, did something different. They used the firm's EpiSwitch and Orion platforms, which model the three-dimensional architecture of the genome — how the DNA is folded inside a living cell.
Dr Ewan Hunter, Oxford BioDynamics' chief data officer, explained the logic: "DNA is folded in our cells, so regions far apart in the linear sequence can touch, and those contact points are where genes get controlled. Orion predicts where these contact points are likely to be."
Crucially, no new patients were recruited. The analysis was computational: published GWAS summary statistics for long Covid, PTSD, rheumatoid arthritis and multiple sclerosis, combined with 3D genomic data from the team's earlier ME/CFS study.
The result that surprised them
What came back, the team says, was the opposite of what they expected.
"We expected to find at least some overlap in genes across the conditions," said Professor Pshezhetskiy. "But we actually found the opposite. At an individual gene level, there was surprisingly little direct overlap."
Then they mapped how those genes interact in networks — and, they report, the picture inverted. Genes that shared almost nothing individually were feeding into the same handful of major biological systems: immune and inflammatory signalling, interferon responses, mitochondrial energy production, metabolic regulation, stress-response mechanisms and neuroendocrine signalling.
"This is not something you can see by reading the genetic sequence alone," Pshezhetskiy said, "which is why these conditions may have looked unrelated for so long."
One gene in particular was flagged in the ME/CFS arm of the analysis: LAG3, a molecule associated with "T-cell exhaustion" — the state in which immune cells, after prolonged activation, effectively burn out. If that finding holds up, the authors suggest, it might help explain why so many patients remain ill long after the original trigger has vanished. The authors stress LAG3 is a candidate thrown up by the analysis, not a confirmed mechanism.
The pushback, and it is serious
The Science Media Centre in London gathered expert reaction, and it was not the chorus of applause that usually greets a press release like this.
Professor Chris Ponting, Chair of Medical Bioinformatics at the MRC Human Genetics Unit at the University of Edinburgh, went straight at the central problem: "the 2026 study's claims are impossible to evaluate because the computational methods they used are proprietary, locked within black boxes they term EpiSwitch Analytical and Data Portals, Knowledgebase and Orion v1 anchor detection. Researchers are thus disallowed from reproducing this study independently."
He raised two further technical objections. First, the statistical threshold: the team predicted "3D genomic anchors" for each disease using GWAS data filtered at p < 0.01, rather than the field's usual stringent standard of p < 5 × 10⁻⁸. Most of those anchors, he argues, will have only very weak genetic support. Second, a cell-type mismatch: the chromosome conformation data comes from circulating blood cells, while the GWAS signals implicate other tissues entirely — lung epithelial cells and alveolar macrophages for long Covid, cortical neurons for PTSD and probably ME/CFS. Different cell types fold their DNA differently.
Dr Sjoerd Beentjes, of the same institute, was more measured but no gentler on the headline: the "biological unifying theory of fatigue" line, he said, "seems an overstatement." He credited the authors for splitting their data properly and correcting for multiple testing — real methodological care — but noted that the paper lacks the metrics that would show these five conditions' networks are more similar to each other than to a network built from any five other conditions. In that sense, he said, "the results seem mostly descriptive."
Professor Carmine Pariante, Professor of Biological Psychiatry at King's College London, offered the most sympathetic reading — and it still stung. The work presents "largely confirmatory evidence" that immune mechanisms matter in stress-related conditions as well as autoimmune ones, using an innovative method on already-published datasets. It does "not really identify novel mechanisms or concepts."
But he added something worth sitting with: the strength of that confirmatory evidence "will be helpful for researchers in the field as well as people with lived experience of these disorders."
Why "confirmatory" still matters
That last phrase is the reason this story is not simply a cautionary tale about press releases.
For people with ME, the history of their illness is a history of being disbelieved. Around 100,000 people in England carry an ME/CFS diagnosis on their records — 0.16 per cent of registered patients — a figure researchers widely regard as a substantial undercount, reflecting who manages to get diagnosed rather than who is ill. Long Covid added millions more people worldwide to the same category of patient: profoundly disabled, and repeatedly told their tests are normal.
Evidence that persistent immune and metabolic dysregulation sits underneath that exhaustion — even confirmatory evidence, even from a study whose bold framing is contested — changes the texture of the conversation in the consulting room.
The prize: a blood test
The prospect the UEA team is really chasing is diagnostic.
"ME/CFS and long Covid are currently diagnosed largely through symptoms, with no universally accepted laboratory test available," said Professor Pshezhetskiy. "That has left many patients facing years of uncertainty."
Earlier work using the same EpiSwitch platform produced a blood-based ME/CFS test that the company reported as highly accurate, and which still requires the validation needed for clinical use. The new paper raises the possibility of biological signatures shared across several conditions rather than one.
"We hope our work could pave the way for objective blood tests capable of identifying underlying biological signatures rather than relying solely on patient-reported symptoms," Pshezhetskiy said.
That is a genuinely large prize — and it is also exactly the kind of claim that demands the independent scrutiny the critics say the proprietary platform currently prevents. Oxford BioDynamics is a commercial diagnostics company. A finding that supports the value of its platform is not thereby wrong, but it does mean the reproducibility question is not academic hair-splitting. It is the whole ballgame.
What this means for you
Be clear about what this study is and is not.
It does not, today, change any diagnosis, any treatment or any prescription. It is a computational reanalysis of existing datasets. No new patients were tested. No drug was trialled. No blood test has been approved. Nobody's care plan should change this week because of it.
If you live with ME, long Covid or another of these conditions, the practical position is unchanged: symptom management, pacing, treatment of specific problems such as orthostatic intolerance or sleep disruption, and — this is not a small thing — a clinician who takes you seriously. If you are being offered a private "biomarker" test off the back of headlines like these, ask directly whether it has been validated in independent studies and whether any clinical guideline recommends it.
If you are a carer or a colleague, the useful takeaway is the one nobody disputes: the exhaustion in these illnesses is not a mood, a motivation problem or a character flaw. Every scientist quoted in this article, including the sceptics, agrees that real immune and metabolic biology is involved.
The verdict
Somewhere between the press release and the pushback sits a reasonable reading.
The UEA and Oxford BioDynamics team have produced an interesting, methodologically careful reanalysis suggesting that five very different illnesses may disturb the same deep biological systems — immunity, energy production, stress response. That is a useful framework and a testable hypothesis, and it aligns with a growing body of work on immune dysfunction in post-viral illness.
It is not yet a unifying theory of anything. It cannot be independently reproduced. Its statistical thresholds are permissive, and the cells it read may not be the cells that matter.
Professor Pshezhetskiy's own closing formulation is the fairer one, and it is worth quoting in full because it is the sentence the field can actually work with: "This study offers a framework for understanding how different triggers can converge to cause the exact same profound clinical exhaustion."
A framework. Not an answer. For patients who have spent decades being told there is nothing there at all, that may still be worth something — as long as nobody mistakes the map for the territory.
