Development and validation of blood-based diagnostic biomarkers for [ME/CFS] using EpiSwitch®… 2025, Hunter et al. (Oxford Biodynamics)

They do raise some of these concerns in the paper and the need for a larger study etc. they say they have tests using this tech in use for prostate cancer and predicting response to checkpoint inhibitors. So I’d like to assume they have some idea of what they are doing. Wishful thinking of a non scientist perhaps!
 
They do raise some of these concerns in the paper and the need for a larger study etc. they say they have tests using this tech in use for prostate cancer and predicting response to checkpoint inhibitors. So I’d like to assume they have some idea of what they are doing. Wishful thinking of a non scientist perhaps!

But prostate cancer is a completely different clinical decision-making process. For prostate cancer we have a cast iron definition of what we want to identify but most of the time no clinical evidence that it is there until too late. So a test for prostate cancer is a prognostic or predictive test of serious disease in the future. For ME/CFS we have no idea what we want to identify or whether there is one of them or six. And it is clinically staring us in the face now. It is defined by its clinical presentation. We don't want a test for ME/CFS. We want a tests for the things that might cause ME/CFS.

The caveat to this would be that within the first three months of post viral fatigue it would be useful to have a test that indicated a longer term ME/CFS course, which is the point Charles Shepherd makes. But then they should have done a study on people in the first three months after a virus. If their test is positive for those who rapidly get better as well it is not a diagnostic test even in that situation.
 
The point is that "accuracy" has nothing to do with it. If the test picks out 60% of ME/CFS and 25% of controls it may be much more useful than one that picks out 95% and 2%, if it tells us what is going on.
Understood. My question is still if this can be replicated will it tell us something useful about what is going on? It seems so from the pathways section of the paper?

If you’re also not comfortable being drawn on that sort of speculation I get it. It just seems pretty relevant here as there are pathways mentioned which tie in with other discussions we’ve had here.
 
It seems so from the pathways section of the paper?

The pathways section looks to me just like a random bucketful of the usual suspects. They talk of overlap with RA and MS so it seems unlikely to tell us anything very specific. The cytokines they pick out - IL-2, IL-10 and TNF - have not figured much in any plausible story I have seen so far. They compare some sort of notional network pattern for ME/CFS and for rituximab (yes, a drug) on the basis that ME/CFS patients responded to rituximab - I found it hard find the motivation to go much further.
 
The pathways section looks to me just like a random bucketful of the usual suspects. They talk of overlap with RA and MS so it seems unlikely to tell us anything very specific. The cytokines they pick out - IL-2, IL-10 and TNF - have not figured much in any plausible story I have seen so far. They compare some sort of notional network pattern for ME/CFS and for rituximab (yes, a drug) on the basis that ME/CFS patients responded to rituximab - I found it hard find the motivation to go much further.
Thanks. Really useful to have your thoughts and look forward to others too.

There seems to be a lot of the usual framing throughout but wondered if there may be something there nevertheless. But am having difficulty getting my head around it all and am not sure if that is the paper, the proprietary tech or my knowledge gaps.
 
The point is that "accuracy" has nothing to do with it. If the test picks out 60% of ME/CFS and 25% of controls it may be much more useful than one that picks out 95% and 2%, if it tells us what is going on.
Something that strikes me again and again is how many scientists there are who don’t seem to understand – or chose to ignore – some very basic principles which can be perfectly understood by people like me with limited scientific knowledge and no scientific training.

If this company were to start selling tests before they have been validated that would be a red flag.
 
Last edited:
Press release from UEA:


Revolutionary blood test for ME / Chronic Fatigue unveiled​

By: Communications

Wednesday 8 October 2025
SHARE
X (formerly known as Twitter) ShareFacebook ShareLinkedIn Share
An arm with a heart shaped plaster at the inside elbow.

With 96 per cent accuracy, a new blood test offers hope for those living with chronic fatigue syndrome - which is often misunderstood and misdiagnosed.
Scientists at the University of East Anglia and Oxford Biodynamics have developed a high accuracy blood test to diagnose Chronic Fatigue Syndrome, also known as Myalgic Encephalomyelitis (ME/CFS).

The debilitating long-term illness affects millions worldwide - including over 400,000 sufferers in the UK - but is poorly understood and has long lacked reliable diagnostic tools.

With 96 per cent accuracy, the test offers new hope for those living with the condition - which is often misunderstood and misdiagnosed.

And it is hoped that the breakthrough could pave the way for a similar blood test to diagnose long Covid.

Lead researcher Prof Dmitry Pshezhetskiy, from UEA's Norwich Medical School, said: “ME/CFS is a serious and often disabling illness characterised by extreme fatigue that is not relieved by rest.

“We know that some patients report being ignored or even told that their illness is ‘all in their head’.

“With no definitive tests, many patients have gone undiagnosed or misdiagnosed for years.

“We wanted to see if we could develop a blood test to diagnose the condition – and we did!

“Our discovery offers the potential for a simple, accurate blood test to help confirm a diagnosis, which could lead to earlier support and more effective management.”

"Post-Covid syndrome, commonly referred to as long Covid, is one example of ME/CFS, where a similar cluster of symptoms is triggered by the Covid-19 virus, rather than by other known causes such as glandular fever. We therefore hope that our research will also help pave the way for a similar test to accurately diagnose long Covid.”

How the discovery was made​

The team used advanced EpiSwitch® 3D Genomics technology from Oxford BioDynamics (AIM:OBD) to see how DNA is folded in blood samples from 47 patients with severe ME/CFS and 61 healthy controls.

Each of our cells contains about two metres of DNA, packed tightly and folded in 3D. These folds aren’t random - millions of them are deliberate, forming a hidden code that helps turn genes on or off to keep us healthy.

OBD Chief Scientific Officer, Alexandre Akoulitchev, said: “Chronic Fatigue Syndrome is not a genetic disease you’re born with. That’s why using EpiSwitch ‘epigenetic’ markers - which can change during a person’s life, unlike fixed genetic code - was key to reaching this high level of accuracy.

“The EpiSwitch platform behind this test, together with OBD’s vast 3D Genomic knowledgebase, has already been proven to deliver practical, rapid blood diagnostics accessible at scale.

“With this breakthrough, we are proud to enable a first-in-class test that can address an unmet need for a quick and reliable diagnostic for a complex, challenging-to-identify illness.”

This approach using EpiSwitch has previously shown success in identifying disease-specific blood markers in highly complex inflammatory and neurological conditions such as fast ALS (amyotrophic lateral sclerosis), rheumatoid arthritis, and certain cancers. This includes the EpiSwitch PSE test, which is a blood test with world-leading accuracy for prostate cancer already used in the UK and US.

The team discovered a unique pattern that appears consistently in people with ME/CFS that is not seen in healthy people.

Using a different approach, this work looked beyond the linear DNA sequence investigated by a previously published DecodeME study, the largest genetic investigation of ME/CFS to date‡.

By examining 3D genomic folds, UEA and Oxford BioDynamics revealed hundreds of additional changes, including five of the eight sites identified by DecodeME, which can now provide a deeper understanding of the disease.

The analysis showed remarkable accuracy – with 92 per cent sensitivity in identifying ME/CFS, which indicates how well the test identifies those who have the disease (a show of true positives) and 98 per cent specificity, which indicates how well it identifies those who do not have the disease.

The researchers also found signs of immune system and inflammation pathways involved in the disease, which may help guide future treatments and identify patients more likely to respond to specific therapies.

A vital tool for diagnosis and treatment​

“This is a significant step forward,” said UEA’s Prof Pshezhetskiy. “For the first time, we have a simple blood test that can reliably identify ME/CFS - potentially transforming how we diagnose and manage this complex disease.”

“Additionally, understanding the biological pathways involved in ME/CFS opens the door to developing targeted treatments and identifying which patients might benefit most from specific therapies.

“We hope that the Episwitch® CFS test could become a vital tool in clinical settings, paving the way for more personalised and effective care.”

This research was led by UEA and Oxford BioDynamics in collaboration with The London School of Hygiene & Tropical Medicine and Royal Cornwall Hospitals NHS Trust.

Development and validation of blood-based diagnostic biomarkers for Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) using EpiSwitch® 3-dimensional genomic regulatory immuno-genetic profiling’ is published in the Journal of Translational Medicine.

 
Last edited:
There’s definitely lots of the usual narrative of inflammation, neuro-inflammation, dysregulated immune states, etc. It has the feel of one of those papers which is trying too hard in parts and perhaps doesn’t try hard enough in others leaving me questioning the findings the more I read.

However I’m not dismissing them, as mentioned this is apparently technology that has been used before and the pathway analysis seems to find things we’ve seen reported before. A few snippets

We mapped the top 200 predictive 3D-genomic markers from the EpiSwitch® whole-genome screen to nearest genes and then performed protein–protein/network and pathway enrichment analysis (STRING, Reactome/MSigDB)

Interleukin signalling (IL-2, IL-10 and related cascades). IL-2 emerged as a central hub in the STRING subnetwork: multiple top nodes connect to IL-2 signalling and T-cell regulation modules. IL-10 and other interleukins also feature prominently, suggesting dysregulated cytokine networks and altered T-cell/ regulatory-T cell balance.

TNF / NF-κB axis and innate inflammation. Several top markers map to genes participating in TNF signalling and NF-κB mediated innate responses, consistent with reports of elevated TNF and related inflammatory proteins in some ME/CFS cohorts.

Toll-like receptor (TLR) signalling and innate immune sensors. TLR and MyD88-dependent modules are enriched, aligning with therapeutic and mechanistic interest in TLR3 (rintatolimod) and innate immune triggers in CFS.

JAK/STAT signalling and cytokine-driven transcriptional programmes. JAK/STAT pathway genes are over-represented, consistent with the centrality of cytokine receptor signalling (IL-2, IL-6 families) in the network

FOXP3 seems to be a high hit in their STRING analysis too. Will share anything else interesting as I read it.

PrecisionLife haven’t always been great at explaining and their technology is a proprietary black box and Zhang et al used small and mixed samples and ML in ways we had trouble with at first. But both seem to be useful or becoming more useful with work. Maybe this can too.
 
George Monbiot on BlueSky:


Full thread on one page:

George Monbiot​

@georgemonbiot.bsky.social

985
Save
ME/CFS is a devastating condition that has long been denied, dismissed, psychologised and underdiagnosed. Research is at last starting to catch up with it, with glimmers of hope for those who have been left untreated for so long. There's a huge BUT coming ...www.theguardian.com

www.theguardian.com
Scientists develop first ‘accurate blood test’ to detect chronic fatigue syndrome
Just as we are beginning to understand the biochemical basis of this horrendous condition, the government appoints the man perhaps most responsible for the mischaracterisation of ME/CFS – Simon Wesseley - to its *overdiagnosis* commission. www.benefitsandwork.co.uk

www.benefitsandwork.co.uk
Controversial professor to investigate overdiagnosis of mental health and neurodivergence for Labour
Like the Tories, the Labour government is trying to drive down the number of people who qualify for disability benefits by insisting there has been an epidemic of “overdiagnosis”: a favourite theme of the BBC and the junktanks of Tufton Street. Never mind the science: what outcome do we want?
In the firing line, as ever, are the UK’s hundreds of thousands of ME/CFS patients. As a practitioner once remarked, “the bastards don’t want to get better”. If there is one characteristic all the ME/CFS patients I’ve come across have in common, it is a desperation to get better.
They desperately want to be able to work, to socialise, to experience all the other joys of life. But because the condition is so poorly understood, they have been repeatedly treated as if they were “malingerers” or “hysterics”. Which suits the government just fine.
Many patients have been forced into useless and sometimes dangerous “buck up” treatments such as Graded Exercise Therapy and CBT, in some cases long after these “treatments” have been debunked. The most eminent advocate of these treatments? Professor Sir Simon Wesseley.
To rub salt into their wounds, these patients have watched as the man many see as their tormentor-in-chief has been garlanded with honours, including a knighthood. His latest appointment is yet another blow.
This commission points to where the government is heading. If the problem is “overdiagnosis” – which appears to be its foregone conclusion – then large numbers of people with horrendous conditions will be put through an even more punishing regime of tests and obstacles to qualify for benefits.
On the one hand, medical science is beginning to get a handle on the very complex physiological basis of this disease. On the other hand, we can expect the greatest medical scandal (so far) of the 21st Century – the mass mistreatment and neglect of ME/CFS sufferers – to continue.

I have replied to George, urging caution and pointing him to this thread but I don’t know if he will see it.

[Update: replies read and liked by George]
 
Last edited:
But prostate cancer is a completely different clinical decision-making process. For prostate cancer we have a cast iron definition of what we want to identify but most of the time no clinical evidence that it is there until too late. So a test for prostate cancer is a prognostic or predictive test of serious disease in the future. For ME/CFS we have no idea what we want to identify or whether there is one of them or six. And it is clinically staring us in the face now. It is defined by its clinical presentation. We don't want a test for ME/CFS. We want a tests for the things that might cause ME/CFS.

The caveat to this would be that within the first three months of post viral fatigue it would be useful to have a test that indicated a longer term ME/CFS course, which is the point Charles Shepherd makes. But then they should have done a study on people in the first three months after a virus. If their test is positive for those who rapidly get better as well it is not a diagnostic test even in that situation.
Wouldn’t it possibly be useful if just for legitimising the illness and increasing chances of people getting proper support? (Probably would help with disability insurance).
 
Science Media Centre:
https://www.sciencemediacentre.org/...blood-based-diagnostic-biomarkers-for-me-cfs/

As well as Chris Ponting's comment there are these two:

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“I think this new research is an interesting development in the search for biomarkers for ME/CFS, and one that may possibly prove important later. But it’s crucial to note that the researchers describe it as a proof-of-concept study, that is, one that is aimed at finding out whether the overall approach of measuring epigenetic markers in blood samples can potentially be useful. Its aim is not to produce a working diagnostic tool, not yet.

“I think the study does establish that the approach is potentially useful. This matches the quote from the lead researcher in the press release, saying that “Our discovery offers the potential for a simple, accurate blood test to help confirm a diagnosis […]”. He is rightly careful not to claim that the research already offers such a test.

“The research paper points out a number of important limitations about the research and its findings, and mentions other research that will have to be done before this can be taken further. Some of that other research is likely to be time-consuming and expensive.

“I think there are several reasons to be sceptical about the exact figures given in the paper and the press release about how well the new test, developed by the researchers, performs. I’m not saying by any means that the procedure is useless – only that on the basis of what we know so far, it’s just not possible to say how well it would work, if and when it is further developed for real clinical practice.

“The new research is a case-control study. In the simplest terms, this means that the researchers used blood samples from some people who had confirmed diagnoses of ME/CFS, known as the cases, and blood samples from other people, the controls, who were as far as possible known not to have ME/CFS. Then they looked for differences in how the DNA (taken from the blood samples) was folded in these two groups, and they found some differences.

“I can’t personally comment on the details of the approach of looking at DNA folding, because that is beyond my expertise, though I have no reason to doubt that it can be useful. But I can and will comment on points about statistics and machine learning.

“One important limitation is that the blood test wouldn’t, as far as I can see, help in deciding which of the participants in the study had or did not have ME/CFS. That’s because it would already be clear who was a case and who wasn’t. The cases were already diagnosed as having ME/CFS, presumably using something like the four key ME/CFS symptoms listed on the NHS website*. The four are extreme fatigue; serious sleeping problems; problems with thinking, memory and concentration (‘brain fog’); and post-exertional malaise. Also, all of them were classified as having severe CFS and were housebound.

“The controls, however, were chosen so that they all had reasonable exercise tolerance, and had never had any of those four key symptoms in the past or at present.

“So all of these are people who would be reasonably easy to diagnose as having ME/CFS, or not having it, using existing diagnostic approaches. A blood test might well speed up their diagnoses, and possibly overcome the scepticism that still exists in some quarters about whether ME/CFS is a real condition.

“But the new results don’t tell us anything direct about how the test method would work in diagnosing people who don’t have such a severe form of ME/CFS, or whether it can distinguish between people who have ME/CFS and other conditions that involve severe fatigue.

“Also anyone who had any history of chronic illnesses (apart from ME/CFS in the cases), or any cancers (current or previous) or autoimmune conditions, or had had various DNA-altering therapies, was excluded from the research. So there’s no guarantee as to how the test, developed in this research, would work with people who have had some of those conditions.

“It can’t tell us about how the test might perform in telling people with ME/CFS from people with other chronic diseases than involve inflammation, for instance. The researchers point this out, and until we know whether or not the test can make this distinction, it can’t be used to decide what treatment such patients should have.

“Overall, it’s probably not too much of a simplification to say that the researchers have found a blood test that seems to work pretty well on the easier diagnosis decisions, so that the concept has potential, but that we just don’t yet know how well it might work in more difficult diagnostic situations.

“Also, as the researchers clearly point out in their paper, the research can’t directly say anything about how the differences in DNA folding patterns are caused. That means that the research doesn’t throw any direct light on what actually causes ME/CFS. (That was not its intention anyway.)

“What is potentially more widely useful is the analysis of the genetic pathways involved in the differences in DNA folding patterns between the cases and the controls, that the researchers carried out. Putting this possibly too simply, this means that they looked at the biological function of the genes that are near the places on chromosomes where the folding differences were. The details of all that are beyond my own expertise, but the findings should give researchers information about where to look for potential causes or treatments. That’s all in the future though.

“As well as all this, there are important reasons to be cautious about the exact numbers quoted for the performance of the diagnostic tool.

“There is actually rather a lot of statistical uncertainty about many of them. (Details are in Table 2 of the paper.) For instance, the sensitivity is quoted in the press release and the paper as 92%. This means that 92% of the people who were known to have ME/CFS had a positive test result.

“But this is a statistical estimate. Yes, 92% is the best estimate if one really insists on a single figure. However, numbers in Table 2 of the paper indicate that it could plausibly be much lower, down to 73%, or rather higher, up to 99%. If it were really somewhere around 73%, that would mean that about a quarter of people who really do have ME/CFS would not in fact have a positive test result.

“There are some details about the way the study was carried out that seem unusual to me, and that in my view may have introduced some biases and limit how far the findings can be generalised “beyond the groups of people directly involved.

“Some of those concerns relate to the complication that all the ME/CFS cases in the study (in the part of the data that was used to construct the diagnostic tool, and also in the validation data that was used to assess the performance of the tool and provided the data in Table 2 of the paper) came from the same source, the LSHTM biobank, whereas some but not all of the controls used in the validation (but not to construct the diagnostic tool) come from a different data set, the ‘OBD repository’.

“The research paper gives no detail or references for either of these data sources, though information about the LSHTM biobank is available online**. Indeed the paper does not say what the OBD repository actually is. I presume it is a data source held by Oxford BioDynamics, who funded this research and whose employees make up most of the research paper’s authors, but no detailed information seems to be available.”

* https://www.nhs.uk/conditions/chronic-fatigue-syndrome-cfs/symptoms/

** https://cureme.lshtm.ac.uk/



Dr Alastair Miller, Retired Consultant Physician in Infectious Disease and Internal Medicine, said:


“The authors are claiming a higher rate of sensitivity and specificity than in most other biomedical tests. 96% accuracy is almost unheard of for this sort of test, so this is quite a remarkable claim. My main concern with this study is the lack of appropriate controls. They are using healthy controls rather than those with other chronic conditions such as depression or fibromyalgia or even MS. My worry is that it will prove to be yet another false dawn, launched with a huge amount of hype and will raise patients’ expectations unrealistically.”
 
Back
Top Bottom