mariovitali
Senior Member (Voting Rights)
I am not trying to defend Amatica but could someone explain to me how come we are so certain that a personalized approach is not relevant as a cure to MECFS ?
neophyte32
Senior Member (Voting Rights)
My friend, I completely agree with you. But I'm not the one you need to convince.
I've invited Jack, whom I've spoken to several times, to come and discuss his work on Science for ME. I thought it would be interesting for him to explain his approach and engage with forum members, particularly those with a strong scientific background. But he has never accepted. I suspect he knows perfectly well that his approach would face serious criticism here, and that this could damage his startup's reputation. I'm almost certain he occasionally reads the forum anyway, if only to see what people are saying about his work.
But what worries me most isn't Amatica itself. It's the fact that people with ME/CFS have been left to fend for themselves, desperately searching for answers to their own illness. Amatica isn't the cause of the problem. It's a symptom of the situation we're in.
Let me give you my own example. If I've now been bedridden for almost two years, in an absolutely dreadful state of health, to the point where I wonder every day how much longer I can endure this existence, it's also because, for more than three years, doctors kept telling me that I could continue exercising, that I needed to push myself, that there was nothing to worry about because my heart was fine. And this was despite my repeated complaints and increasingly alarming symptoms.
I pushed myself so hard that I ended up lying in bed having seizure-like episodes. Doctors came to my home three or four times a year, sometimes completely baffled by what they were witnessing. One of them even asked whether I had considered the possibility that my wife was poisoning me. Can you imagine the level of medical confusion I was dealing with?
So of course, when you reach that point, you start desperately looking for answers elsewhere.
And that's precisely the vacuum in which companies like Amatica thrive. But they're far from alone. There are German laboratories offering GPCR autoantibody tests, cytokine panels, all kinds of immunological investigations, and countless other services targeting people with ME/CFS and Long Covid. An entire market has developed around our desperation.
Amatica has simply been more successful than others at building a compelling narrative. Jack communicates constantly, tells a story, offers hope. With artificial intelligence in the background, he explains that we'll soon be able to understand the mechanisms specific to each patient and develop personalised treatments. That's exactly what people want to hear. And I completely understand why.
The message is simple, attractive and effective. It draws in investors willing to put millions into the company, along with patients prepared to spend more than £1,000 on testing. I'm one of those patients myself, by the way.
Meanwhile, what do patients actually have? Research progressing slowly, often contradictory findings, one hypothesis after another, and no solid prospect of effective treatment in the near future. That's obviously much harder to sell than the promise that AI will solve everything within two years and that we might have a cure within five.
And this is where I'd really like to thank the members of this forum.
Science for ME has genuinely opened my eyes. I've discovered just how much of what passes for science in our field sometimes rests on shaky foundations, excessive interpretations or premature conclusions. Above all, I've learned to distinguish an interesting hypothesis from actual scientific evidence. And I'm sincerely grateful to all of you for that.
I've also tried to bring several French patients here, so they could discover these discussions and this level of scientific rigour. But many simply don't want to come. Perhaps because it's difficult to confront a reality that's far less encouraging than the promises they read every day on social media.
Take another example. I regularly talk about DecodeME and SequencEM on Twitter. And the responses are almost always the same: "What's the point of genetics? ME/CFS isn't a genetic disease!"
I then try to explain that this isn't the point, that genetics can help us identify biological mechanisms and understand some of the pathways involved in the illness. But people respond that the mechanisms aren't the same in everyone.
When I suggest that the central nervous system, perhaps even the brain itself, might play a central role, I get responses like: "But I have microclots", "I have craniocervical instability", or "I have neuropathy." As though these observations necessarily ruled out the existence of common underlying mechanisms.
And very quickly, we return to the idea that ME/CFS must have countless different causes, that it's essentially a different illness in every patient and therefore requires an individual treatment for each person. Some people argue this far more eloquently than I can, and with tremendous confidence. Yet at this stage, we simply don't know whether it's true.
So I completely understand the anger expressed by some members of this forum towards these companies and their promises. But what can we actually do about it? Call them out? Of course, when there's good reason to do so. But will that really change anything?
Startups have always exploited gaps and unmet needs that others have failed to address. That's how the market works. And in ME/CFS, the gap is enormous. We have patients who have been abandoned, some in absolutely horrific conditions, and a medical system that still has very little to offer them.
That's why I sometimes wonder whether certain researchers, or even some members of this forum, might benefit from being more active on social media. Not to promise miracles or turn research into a spectacle, but simply to explain what we actually know, what we don't know, and why some avenues of research deserve more attention than others.
I know perfectly well that science isn't done on Twitter. But today, that's also where a large part of the patient community forms its opinions about research.
And when serious researchers are absent from that space, they leave it to those who are best at communicating, not necessarily those with the strongest scientific arguments.
Amatica has understood this better than many others. And I find it a shame that scientific rigour is so difficult to hear over a message that is so effective at selling hope.
Mij
Senior Member (Voting Rights)
"My WASF3 is elevated, super interesting marker."
https://x.com/JackHadfield14/status/2116154586277048621/photo/1
https://x.com/JackHadfield14/status/2116154586277048621/photo/1
First of all, someone needs to define what a ‘personalised approach’ actually means – does it refer to n=1, different diseases, or subgroups? Often ‘personalised approach’ seems to be more marketing jargon than anything else.
But let’s assume we’re talking about subgroups in ME/CFS. Personally, I’m not at all sure to what extent any treatments will need to be personalised, but the following points bother me in this discussion
1. The claim that such subgroups definitely exist, usually coupled with the claim that this is proven by differing responses to certain off-label therapies. I would disagree with both.
2. When people conclude that subgroups must first be defined or identified in order to carry out relevant research and/or find treatments. I consider this to be wrong as well.
For example, I’d be willing to bet that there is some common mechanism underlying sensory hypersensitivity in all people with ME. The aim for me would be to identify this, and then, from there, you can a) investigate whether there might be various upstream causes for this mechanism and b) starting to look for treatments.
mariovitali
Senior Member (Voting Rights)
It means that given differences we all have and since ME/CFS is a polygenic disease, this most likely means that one intervention or a number of certain interventions will not benefit every single patient that will receive it and it actually may make him worse.
Which evidence are you using where this disagreement could be based on?
Again, as above. People could not and should not conclude anything. It's the evidence which concludes.
On the contrary, Moreau's 2025 FGF-21 study is an example. Stratifying patients by FGF-21 concentrations revealed differences in PEM, immune/autonomic symptoms and cognition that were not adequately captured by diagnostic category alone. PrecisionLife’s combinatorial analysis frames ME as highly polygenic and heterogeneous
Jonathan Edwards
Senior Member (Voting Rights)
I don't think this is true in reality @mariovitali.
I can think of a dozen common diseases with polygenic risk. None of them are treated by genetic subsets: rheumatoid, lupus, Paget's, osteoarthritis, schizophrenia, ... Treatments that work tend to work for ~70% of patients or more and bear no relation to genetic make up. We almost never test for the risk genes in the clinic.
The only realistic example of precision medicine is in cancer where the treatment is tailored to acquired somatic mutations, not germline genome.
Peter T
Senior Member (Voting Rights)
I while ago I listened to radio interview with a neurologist who works with MND (ALS) who is working on different medications in relation to specific genetic patterns. She suggested that some patterns have now been identified within MND where those with those specific patterns respond to medication aimed at slowing or even reversing the progression of the disease. I did not follow this up and can’t now remember the name of the neurologist, though she works in Sheffield. But this would seem to be an example of a disease where sub grouping on the basis of genetics indicates different treatment courses.
Whilst the enormous variation within ME/CFS makes it tempting to speculate on sub groups and it is theoretically possible that ME/CFS may have subgroups which in the future might be found to indicate different treatment courses, we are no where near being able to define relevant criteria for such divisions. Certainly the current research on popular medications, such as LDN, that has produced null results does not yet indicate the existence in experimental conditions of a reliable minority that do respond to the treatment(s).
Whilst the enormous variation within ME/CFS makes it tempting to speculate on sub groups and it is theoretically possible that ME/CFS may have subgroups which in the future might be found to indicate different treatment courses, we are no where near being able to define relevant criteria for such divisions. Certainly the current research on popular medications, such as LDN, that has produced null results does not yet indicate the existence in experimental conditions of a reliable minority that do respond to the treatment(s).
I would like to make the case for "subgroups", if that's the only term we have for classifying pwME/CFS with differences, based on metabolomics.
pmc.ncbi.nlm.nih.gov
Metabolomics, which is defined as the comprehensive analysis of metabolites in a biological specimen, is an emerging technology that holds promise to inform the practice of precision medicine.
Historically, small numbers of metabolites have been used to diagnose complex metabolic diseases as well as monogenic disorders such as inborn errors of metabolism.
Current metabolomic technologies go well beyond the scope of standard clinical chemistry techniques and are capable of precise analyses of hundreds to thousands of metabolites.
Consequently, metabolomics affords detailed characterization of metabolic phenotypes and can enable precision medicine at a number of levels, including the characterization of metabolic derangements that underlie disease, discovery of new therapeutic targets, and discovery of biomarkers that may be used to either diagnose disease or monitor activity of therapeutics.
I will not list the variables here, but if anyone wants realtime examples of people with opposite biomarkers and opposite reactions to current supplements and meds, I will slowly gather them and post them here.
Metabolomics: an emerging but powerful tool for precision medicine
Metabolomics: an emerging but powerful tool for precision medicine - PMC
Metabolomics, which is defined as the comprehensive analysis of metabolites in a biological specimen, is an emerging technology that holds promise to inform the practice of precision medicine. Historically, small numbers of metabolites have been ...
pmc.ncbi.nlm.nih.gov
Metabolomics, which is defined as the comprehensive analysis of metabolites in a biological specimen, is an emerging technology that holds promise to inform the practice of precision medicine.
Historically, small numbers of metabolites have been used to diagnose complex metabolic diseases as well as monogenic disorders such as inborn errors of metabolism.
Current metabolomic technologies go well beyond the scope of standard clinical chemistry techniques and are capable of precise analyses of hundreds to thousands of metabolites.
Consequently, metabolomics affords detailed characterization of metabolic phenotypes and can enable precision medicine at a number of levels, including the characterization of metabolic derangements that underlie disease, discovery of new therapeutic targets, and discovery of biomarkers that may be used to either diagnose disease or monitor activity of therapeutics.
I will not list the variables here, but if anyone wants realtime examples of people with opposite biomarkers and opposite reactions to current supplements and meds, I will slowly gather them and post them here.
Peter T
Senior Member (Voting Rights)
I may be missing the point, but I just glanced through the review article on metabolomics and did not see any studies relating to ME/CFS?
Jonathan Edwards
Senior Member (Voting Rights)
And it is from 10 years ago. I may be out of date but I doubt metabolomics has achieved anything in that time, in any disease.
The general goal is to use many markers for prediction. I guess the problem is the design of this project doesn't go anywhere.
You could take a while bunch of markers from 500 patients and use ML (or now genAI) to define subsets. But unless you get a second sample from those same people month or so down the track, you will never know if those subset markers stick. You'd have to prove that those same people stay within the subset you defined, otherwise it's not a disease subset, it's a transient physical state.
This is even without understanding a subset has any biological value, if it's even actually to do with disease or because these people would cluster independent of the disease.
So it is a messy concept.
Aren't many diseases polygenic? In Parkinson's <15% of cases can be linked to genetic risk factors, they still all have a lack of dopamine producing neurons, even Diabetes Type 2, while some risk gets inherited is a polygenic disease by definition.
Mostly the 'evidence' you state above, that there simply isn't enough evidence to confidently state that there definitely are distinct 'subgroups'.
I think it's difficult to make such a claim about heterogeneity until it has been proven that any of the markers they’re using are actually relevant to the disease. Otherwise, the ‘heterogeneity’ could also stem from factors that have absolutely nothing to do with it, or maybe FGF-21 is a marker correlating with disease severity. Furthermore, the condition may still have common causes, either upstream or downstream, even if this FGF-21 turns out to be directly relevant to the condition.
Again, I am not against the subgroup idea per se, but what I find pointless is the assumption that subgroups definitely exist, are the only way forward and that we can easily find meaningful subgroups on the basis of some random markers, without knowing which processes are actually involved in the disease. The latter seems also to be the promise that Amatica makes.
Jonathan Edwards
Senior Member (Voting Rights)
Has that ever worked?
It has for Kidney disease, MS, Diabetes. There are many more in the pipeline because of UK Biobank, they may take 5 years or so. Nightingale health is a platform that has developed or do this for heart diseases.
I know a proteomics/metabolomics group developing this for confirmation of mitochondrial disease.
I'd say none are at massive scale because of the dominance of pathology companies in standard healthcare even those these other omics tools outperform it.
But I will say, the idea that this might work for subsets is not clear. It remains exploratory. I think you are right to be sceptical. I work in this space, the main area of scepticism is CV%.
A lot of these omics measures used have very high CV%, which means if you sent the same exact sample twice to a service with these assays then the data you get back might be varied by ~20% on average. If a disease has a crazy high marker then it doesn't matter but if you are using multiple low lying markers then you are playing in quicksand.
Main thing people should ask is "how reliable is this data?".
Jonathan Edwards
Senior Member (Voting Rights)
Sufficient to be clinically useful, presumably prognostically?
I am surprised to be honest. There are plenty of diseases where you make a diagnosis or predict one based on three or four biochemical measures but I wouldn't call that metabolomics. The brain people I have been talking with talk in terms of single predictive molecules like synuclein.
ChronicallyOverIt
Senior Member (Voting Rights)
@MelbME those other disease they were able to screen on other biomarkers first, to identify what was wrong? Other than test variability, my main fear is that anyone can send a test in? It feels like amaticas data set is a lump of selection bias and potentially not even Me/CFS.
Is that a valid concern?
Is that a valid concern?
Yes mostly prognostic.
Renalytix - Renalytix
Transforming kidney disease care with one simple blood test. With our KidneyIntelX™ technology and data-driven insights, we help clinical teams understand patient level of risk for progressive decline in kidney function in the earliest stages of diabetic kidney disease. This enables clinicians...
renalytix.com
Octave Bioscience | Precision Neurology for MS Care
Octave Bioscience is advancing multiple sclerosis care through precision neurology. The Octave MSDA Test measures 18 blood biomarkers to inform MS care.
www.octavebio.com
The Galleri Test - GRAIL
The Galleri Test The Galleri® multi-cancer early detection test Cancer screening saves lives, but 70% of cancer deaths are caused by cancers without
grail.com
Building better healthcare for everyone.
The most advanced health check for the health industry.
nightingalehealth.com
Well the point of the omics screening is to capture thousands of markers. If 3 or 10 jump out together then you go and take those to a lab to define a specific assay with a much lower CV% for translation to clinical use.
Jonathan Edwards
Senior Member (Voting Rights)
Hmm. Those are adverts. I wonder just how useful they are. They seem to be giving risk rates over a period of years. There aren't many situations where that alters what you actually do. One of them seems to be a combination of tests for different cancers rather than a multiprotein indictor for an outcome. At least in the UK any value of these tests is likely to be completely drowned out by the chaos of the haealthcare provision system!!
I guess I used half a dozen cell and protein assays to manage people on rituximab but I didn't think it needed calling omics or personalised medicine. Just doing the relevant tests.
No, they are much more controlled studies. Amatica's plan seems very noisy. But noise can be dropped with scale. The question will be how many people do they need? It might need to be decode scale (tens of thousands).
I don't think true diagnosis of ME/CFS matters, the diagnosis itself is messy anyway. I don't think it's a problem for the Amatica plan. They seem to be trying to tie markers to treatment outcomes in individual patients. I think their concept is good and we do something similar.
Main concern I have is that they need to be getting multiple time points from the same patients to validate that a marker is specific. They also need to understand the variability of some of their data producing tools. They also rely on subjective treatment success which cold just be placebo.
Overall though, the data they produce will be valuable. But will they share the data? Or is the intention to hoard it for their company?
Last edited:
Those are the companies and products. These ones are used clinically. They'd be more widespread if the gigantic pathology companies didn't have a stranglehold on public and private health systems.