Uncovering the genetic architecture of ME/CFS: a precision approach reveals impact of rare monogenic variation, 2025, Birch, Younger et al

[jnmaciuch]

But if these numbers are representative, then - assuming my math is correct - one out of eight ME/CFS cases alone can be attributed to a single variation of channelopathy

One out of eight. Are these not-ME/CFS? Are there more of what we call ME/CFS that are not ME/CFS? Is that just the ratio for a single brand of ME/CFS? Or an indicator that true ME/CFS is rarer still? Or is it not even a thing??

I don't think we can really answer any of those questions from this data--they were specifically looking for variants that were common between their ME/CFS cases to highlight in the results, not doing an unbiased search. And they're only guessing it's relevant here because it's relevant in some other disease contexts when it's a homozygous, rather than a heterozygous, mutation.

What they're doing here is the same level of causal inference as if I went to the homes of a bunch of people that suffered a heart attack looking for something which would "explain" the connection between them and found that a handful always bought the same brand of deli meat, which I thought was important because processed meats have sodium and that's been linked to heart disease.

 

[alex3619]

I am not sure why 'genetic heterogeneity' is problematic.

Agree. High genetic heterogeneity is only a problem is there is high downstream causal heterogeneity. If a huge number of genes have potentially a small impact on the pathological mechanisms, but the mechanisms are conserved, the focus should be on those mechanisms. Now metabolomic, RNA and proteomic sequencing are still relevant until we have identified and understood the mechanisms.

 

[LizWorthey]

Hey all. This is my labs paper, so I'm totally not unbiased. That said; we tried to be careful in what we wrote and as we outlined; the goal of putting this paper out there is that we felt it was 100% important to consider that in some cases - as with some of the participants we enrolled - the patients have rare, well supported pathogenic variants that explain all ME/CFS symptoms.. I absolutely feel that - at the molecular level - ME/CFS is not one, but rather perhaps even hundreds of disease with convergent symptoms... Anyway. In the spirit of open dialogue, I thought I'd answer the questions in the thread; as best I can. Totally open to critiques, corrections, criticisms, and suggestions. The goal is only to better understand the disease .

• “This study just shows how badly we need a much bigger whole genome sample like SequenceME”
  • Yes! And in fact we mention the need for larger cohort and validation in them many times in the article . But as we also state; a huge study will be useless if we don't try to understand what we might be looking for...

• “I’m not sure if Carruthers refers to CCC or ICC, but neither are the Fukuda criteria that they mentioned earlier.”
  • Sorry about the criteria vagueness; it is vague due to edits in last minute revisions. I didn’t catch them (was out sick). It will be updated in the final version (I had already contacted the journal last week when noted ☹).

• “Do we know that these genes are pathogenic for ME/CFS though? I’m sure you’d find lots of variability in any group of patients.”
  • We are all individuals with variability; agreed, but, where listed, these are all rare or ultra-rare variants already identified as known pathogenic/likely pathogenic etc. Where less evidence is noted; it is truly because sometimes variants that are good leads with less evidence are never published.. Thought better to try to get things out there even when they need more work in some cases (and we note all ambiguity in the paper; at least we try to do that very carefully..).

• “I am unclear how you can draw any conclusions from rare gene variants in a small population. They may well be irrelevant. I have not read the whole paper but the abstract tells us nothing useful and seems confused.”
  • Happy to discuss more any confusion/lack of clarity.. In general I'd say that there are thousands of studies of rare variants in small populations that have changed everything for those patients. Just think about cystic fibrosis.. And many common diseases have genetic factors that are rare and only seen in one/certain individuals.

• “The cohort seems to have been recruited online, which is problematic because such cases are likely to be atypical.”
  • Sorry if it was not clear; this was a local recruitment, part of the screening was online – in part because of the importance of not exposing people to large medical center environs in ’20-’22, but later steps were always in person.

• “May be wrong but I didn't see any control group?”
  • Yes this was not a large case – control study. Rather, as noted; it was a small, deep, well phenotyped pilot study . As outlined in the paper; we did try to enroll family members as good "genetic controls", but that wasn’t always possible..

• “But looking up mutations for the first gene, KCNJ18, shows that are "seldom pathogenic" as this paper concludes.” “however, labels them as: "pathogenic: definitive"
  • In earlier version we had a whole discussion about pathogenic versus deleterious re mendelian versus risk factor.. For the KCNJ18 variants in particular this is a very nuanced discussion and I hate to think anything was unclear or perceived as over-stated. These are deleterious variants – but absolutely show up in disease in complex situations as we discuss in the paper… KCNJ18 disease manifestations require presence of the variant plus often a particular thyroid state, plus exertion or other stressor. These patients have other things noted.. and we may have missed something additional - as noted.. But I 100% believe that these variants are part of the picture.

• “For the other mutations listed in the table, the patients were heterozygous so they had only this mutation only once. Suspect that most diseases that they associate with these mutations are recessive: they only manifest if people have two of these mutations without normal back up copy.”
  • Some variants are seen in a het state, but this is not the case across all variants. Some variants are associated with dominant diseases or traits. We discuss where there is evidence for an effect from a dominant variant or not.

• “The problem @ME/CFS Science Blog is highlighting is that if these gene variants are pathogenic then presumably the individuals concerned do not have ME/CFS but have some other disorder like a channelopathy that has been misdiagnosed. If they are not pathogenic they may mean nothing.”
  • That is one of the major points made in the paper. We believe that these patients have diseases someone called ME/CFS, but they have hemolytic anemias, mitochondrial disease, cardiomyopathies etc.. ME/CFS is not a disease always caused by one variant or one gene. ME/CFS is the clinical diagnosis, but the cause; well that is where the complexity comes in . We calculated that there are more than a 100k combinations of symptoms that would get a label of ME/CFS according to criteria.

• “For example, the first mutation, p.Gln407Ter (also known as Q407X) was first thought to be pathogenic, but this 2016 paper showed it was likely not the case. The same is true for other KCNJ18 mutations mentioned. This ME/CFS paper by Birch et al., however, labels them as: "pathogenic: definitive", which seems quite misleading.
  • Yes we carefully reviewed all the info out there, including the original and this noted reference (it really is not clear cut as represented here) - the variants noted have been shown to be deleterious (we actually did have a long discussion on the diff between deleterious, pathogenic, risk factor in an earlier draft, but it was deemed to be too complex, so we reduced it (maybe too much - eek). In this specific case; the disorder is complex and multifactorial, with varied influences. The use of the term definitive is explained as carefully as we could. The intent was not to over interpret, but rather provide context.

• “Suspect that most diseases that they associate with these mutations are recessive: they only manifest if people have two of these mutations without normal back up copy.”
  • We actually clearly list the known mode of inheritance for each variant; you are right - some are recessive - but we note any evidence of dominant phenotypic effects - and others are dominant. We tried to be very very specific on the evidence and the limitations throughout the paper - sorry if still not well enough outlined.

• “Do these findings at least hint that the Australian school might have some merit?”
  • I confess that for a variety of reasons I’m not sure what the “Australian school" refers to; is it the results that discussed heterozygosity? If so then – yes – this is what we believe/have seen.

• “But if these numbers are representative, then - assuming my math is correct - one out of eight ME/CFS cases alone can be attributed to a single variation of channelopathy. “
  • We see only two instances of shared variants, but there are 4 cases of variants in the same gene. It is a small cohort as we discuss and we do have noted limitations - familial variants, moderate disease impact, and geographic location. That said; yes – channelopathies do seem to be an avenue showing up.

• “I don't think we can really answer any of those questions from this data--they were specifically looking for variants that were common between their ME/CFS cases to highlight in the results, not doing an unbiased search”.
  • Actually; this was an unbiased search; well, unbiased in that we set up analyses to try to find variants of interest in an n-of-1 approach for each patient one by one; with no preconceived notion of exactly what we were going to find. We still had to filter for 5/6 million variants based on known scientific associations between diseases, genes, and phenotypes - so I guess not totally unbiased .

• “What they're doing here is the same level of causal inference as if I went to the homes of a bunch of people that suffered a heart attack looking for something which would "explain" the connection between them and found that a handful always bought the same brand of deli meat, which I thought was important because processed meats have sodium and that's been linked to heart disease.”
  • This isn’t quite right, respectfully. A better parallel would be if we went in to houses of people who had suffered an earlier in life heart attack and looked for genes/variants that were associated with heart attacks, chest pain, palpitations, family history of sudden cardiac death, heart abnormalities, channel defects etc. All the variants listed in the higher level of evidence categories are clinically well supported to be associated with core ME/CFS symptoms.

• “High genetic heterogeneity is only a problem is there is high downstream causal heterogeneity. “
  • Yes – 100% - this is how we think of this too. There are tens of thousands of variants associated with increased risk of breast cancer, dementia, developmental delay (choosing easy examples, I acknowledge), but no doubt that they can all cause these complex diseases despite heterogeneity.. The individual variants can differ even when the mechanisms are conserved.

Again; truly just trying to be helpful and transparent here. We're very invested in trying to move the field ahead whilst being careful in what we talk about, but not constrained by rhetoric or limited to just what we thought before. Hopefully that comes across. I'm always happy to answer questions or address critiques!

 

[ChronicallyOverIt]

• “I’m not sure if Carruthers refers to CCC or ICC, but neither are the Fukuda criteria that they mentioned earlier.”
  • Sorry about the criteria vagueness; it is vague due to edits in last minute revisions. I didn’t catch them (was out sick). It will be updated in the final version (I had already contacted the journal last week when noted ☹)

Wow awesome to see these responses, I think an open dialogue with the authors is always awesome.

I don’t have much to add but still a bit confused on the criteria used, will you only announce once the final draft is updated? Maybe I’m not reading this correctly but I’m still unsure what diagnostic criteria was used to filter for patients even after this reply.

 

[Hoopoe]

I have a pathogenic variant on one allele of the ACADM gene. I've also had some issues with hypoglycemia which would fit.

The conventional wisdom says that, even if this would lead to a loss of about half of enzyme activity, one defective allele is not enough to cause clinically significant disease. Since ACADM deficiency has been studied extensively in the population this is probably true, but I'm wondering if this mutation could be causing problems when its effects combine with other things.

 

[SNT Gatchaman]

Hi Liz (@LizWorthey) and welcome. Thank you very much for joining (and for posting on Bluesky too).

 

[Utsikt]

Hi and welcome @LizWorthey

Can I bother you about adding line breaks between the different questions? I appreciate the effort to format with indentations which already helps, but walls of text are notoriously difficult to read when you’ve got brainfog. Thank you!

 

[LizWorthey]

Hi and welcome @LizWorthey

Can I bother you about adding line breaks between the different questions? I appreciate the effort to format with indentations which already helps, but walls of text are notoriously difficult to read when you’ve got brainfog. Thank you!

Absolutely. Will amend!

 

[LizWorthey]

Wow awesome to see these responses, I think an open dialogue with the authors is always awesome.

I don’t have much to add but still a bit confused on the criteria used, will you only announce once the final draft is updated? Maybe I’m not reading this correctly but I’m still unsure what diagnostic criteria was used to filter for patients even after this reply.

Hopefully this helps; sorry again about that. I was in a pretty big crash when the final revisions came in and that's why it was mis edited :

Inclusion criteria required a confirmed diagnosis of ME/CFS meeting the Fukuda 1994 definition, with additional cutoffs for moderate fatigue severity and frequency proposed per Jason and colleagues (2014) to reduce misclassification. Whilst not used for inclusion/exclusion criteria, all participants also met the Canadian Consensus Criteria (CCC) criteria. I'm happy to go in and check whether the participants met other criteria and report back.. Won't get to that until next week at the earliest .

 

[V.R.T.]

I absolutely feel that - at the molecular level - ME/CFS is not one, but rather perhaps even hundreds of disease with convergent symptoms...

What is your evidence for this belief? I find it a rather extraordinary and unlikely claim that so many different diseases could cause post exertional malaise and all of these different pathologies could go completely undetected.

ME/CFS is the clinical diagnosis, but the cause; well that is where the complexity comes in . We calculated that there are more than a 100k combinations of symptoms that would get a label of ME/CFS according to criteria.

What criteria are you referring to here? Surely most sensible people now agree that the presence of PEM is the central defining feature of ME/CFS, rather than any other symptoms.

Thank you for engaging here.

 

[Kitty]

Thanks for all your work on this and your explanations, @LizWorthey, and welcome!

I don't know enough about genetics to contribute, really—except, I guess, to say that the Fukuda criteria ought to have been binned 25 years ago. The sooner researchers make a positive decision to stop using it, the better defined their cohorts will be.

The problem is that it doesn't require PEM, so it doesn't describe ME/CFS. It also allows people to pick and choose from a menu of symptoms, meaning it's too inclusive to be of any practical use. It has led to a lot of research money being wasted—no one can rely on the results of studies on groups that could have had any one of a dozen fatiguing illnesses.

Look forward to hearing more about your work, though!


ETA: cross posted with @V.R.T.

 

[Utsikt]

Absolutely. Will amend!

Thank you, much appreciated!

Do we know that these genes are pathogenic for ME/CFS though? I’m sure you’d find lots of variability in any group of patients.”

• We are all individuals with variability; agreed, but, where listed, these are all rare or ultra-rare variants already identified as known pathogenic/likely pathogenic etc. Where less evidence is noted; it is truly because sometimes variants that are good leads with less evidence are never published.. Thought better to try to get things out there even when they need more work in some cases (and we note all ambiguity in the paper; at least we try to do that very carefully..).

I’m not sure I follow. ME/CFS is a syndrome that is centered around PEM, where there are no other diseases that can explain the symptoms.

Of course, that will not always be the case when we understand how ME/CFS actually works (then the disease ME/CFS will explain the symptoms), but if some of these cases had pathogenic genetic variants then they have been misdiagnosed.

We this a lot when the use of criteria varies, and different doctors have different understandings of what PEM is and isn’t (see e.g. our factsheet for details on PEM), and they often don’t do a proper job ruling out alternative explanations.

“The problem @ME/CFS Science Blog is highlighting is that if these gene variants are pathogenic then presumably the individuals concerned do not have ME/CFS but have some other disorder like a channelopathy that has been misdiagnosed. If they are not pathogenic they may mean nothing.”

• That is one of the major points made in the paper. We believe that these patients have diseases someone called ME/CFS, but they have hemolytic anemias, mitochondrial disease, cardiomyopathies etc.. ME/CFS is not a disease always caused by one variant or one gene. ME/CFS is the clinical diagnosis, but the cause; well that is where the complexity comes in . We calculated that there are more than a 100k combinations of symptoms that would get a label of ME/CFS according to criteria.

The purpose of a diagnosis is to predict future health and response to treatments. If the patients have hemolytic anemia they shouldn’t be given an ME/CFS diagnosis, and they will not have the pathology that one day will be called the disease ME/CFS.

It will no doubt be useful for the individual patients to know why they are actually sick, but that’s a different question than if this tells us something about ME/CFS in general.

 

[forestglip]

Of course, that will not always be the case when we understand how ME/CFS actually works (then the disease ME/CFS will explain the symptoms), but if some of these cases had pathogenic genetic variants then they have been misdiagnosed.

I don't agree with this. "When we understand how ME/CFS actually works" then presumably we'll have mechanisms and possibly pathogenic variants that explain the symptoms. If we identify a variant that, for example, causes ion channels to not work in half of people with ME/CFS, do we say, "well I guess it wasn't ME/CFS, it was a channelopathy in these people"?

In my opinion, ME/CFS should just be describing symptoms at this point. If someone has an explanation for their symptoms in the form of a variant already, great. It'd help to identify the common pathways in everyone with the syndrome if we don't dismiss the pathways we identify in a subset as being not ME/CFS.

 

[Utsikt]

I don't agree with this. "When we understand how ME/CFS actually works" then presumably we'll have mechanisms and possibly pathogenic variants that explain the symptoms. If we identify a variant that, for example, causes ion channels to not work in half of people with ME/CFS, do we say, "well I guess it wasn't ME/CFS, it was a channelopathy in these people"?

In my opinion, ME/CFS should just be describing symptoms at this point. If someone has an explanation for their symptoms in the form of a variant already, great. It'd help to identify the common pathways in everyone with the syndrome if we don't dismiss the pathways we identify in a subset as being not ME/CFS.

Lots of diseases will have lots of overlapping symptoms. So fulfilling the criteria isn’t enough to establish relevance.

Unless we think that ME/CFS is dozens or hundreds of diseases disguised in a trenchcoat, don’t we have to assume that clear one-off pathology is a misdiagnosis unless there is compelling evidence that we’ve stumbled upon something at the core of ME/CFS pathology? If that pathology is an ultra-rare genetic variant, how would you make that case?

 

[DHagen]

Welcome, @LizWorthey! Just wanted to chime in and express my gratitude to you for taking the time to join in the discussion here and to directly engage with the forum concerning your work. I am very much looking forward to reading more.

 

[forestglip]

Unless we think that ME/CFS is dozens or hundreds of diseases disguised in a trenchcoat, don’t we have to assume that clear one-off pathology is a misdiagnosis unless there is compelling evidence that we’ve stumbled upon something at the core of ME/CFS pathology? If that pathology is an ultra-rare genetic variant, how would you make that case?

The whole concept of ME/CFS seems less helpful to me as a whole than just the concept of PEM. ME/CFS seems like a bunch of random symptoms put together because they seem to somewhat occur together and which help put people in boxes at the doctor's office. And even with this ME/CFS diagnosis, there's a huge amount of variantion between patients, as evidenced by the variety within any criteria and between different criteria, so it's hard to consider it as one specific thing to figure out.

Instead, PEM is one very specific symptom, that more or less has no mechanistic explanation in any disease it crops up in, and I think a major goal of everyone here is identifying what that mechanism is.

If someone with PEM has an ultra-rare variant that might explain why physical and cognitive exertion is causing delayed and prolonged symptom exacerbation, then that's wonderful for helping put the puzzle together.

Just because someone has an ultra-rare variant doesn't mean they have a different disease. Diseases can be caused by a fault in a pathway consisting of a hundred different steps where every part along it needs to work properly. The rare variant might cause a break in one part of the pathway, but which leads to the same outcome as a break along any other part which might show up more often.

Researchers don't talk about rare variants in other diseases as indicating misdiagnosis. They use it to help identify common pathways.

 

[Utsikt]

The whole concept of ME/CFS seems less helpful to me as a whole than just the concept of PEM. ME/CFS seems like a bunch of random symptoms put together because they seem to somewhat occur together and which help put people in boxes at the doctor's office. And even with this ME/CFS diagnosis, there's a huge amount of variantion between patients, as evidenced by the variety within any criteria and between different criteria, so it's hard to consider it as one specific thing to figure out.

Instead, PEM is one very specific symptom, that more or less has no mechanistic explanation in any disease it crops up in, and I think a major goal of everyone here is identifying what that mechanism is.

If someone with PEM has an ultra-rare variant that might explain why physical and cognitive exertion is causing delayed and prolonged symptom exacerbation, then that's wonderful for helping put the puzzle together.

Just because someone has an ultra-rare variant doesn't mean they have a different disease. Diseases can be caused by a fault in a pathway consisting of a hundred different steps where every part along it needs to work properly. The rare variant might cause a break in one part of the pathway, but which leads to the same outcome as a break along any other part which might show up more often.

Researchers don't talk about rare variants in other diseases as indicating misdiagnosis. They use it to help identify common pathways.

Thank you for explaining. I can see your point. I guess two important questions would be:
Do they have PEM?
Are the rare variants co-morbidities or contributing to the PEM?

I don’t know how likely it would be to stumble upon a non-comobidity rare gene in a small study like this.

 

[jnmaciuch]

Thanks for your willingness to go back and forth. I see you're in the genetics and bioinformatics department! I'm doing my PhD in bioinformatics, glad to have someone with your background join the forum! Discussions on here tend to trend towards the critical side--I've found it really useful for my own research and thinking, though it can take some getting used to.

Actually; this was an unbiased search; well, unbiased in that we set up analyses to try to find variants of interest in an n-of-1 approach for each patient one by one; with no preconceived notion of exactly what we were going to find. We still had to filter for 5/6 million variants based on known scientific associations between diseases, genes, and phenotypes - so I guess not totally unbiased

I’ll try to explain what I mean better—the bias here comes from the assumption that if a person with ME/CFS has a variant and overlapping symptoms with the monogenic illness for that variant, that in itself is support for the idea that the variant is causally relevant for the specific symptoms in ME/CFS (especially when we don't have information on the frequency of symptoms in heterozygous individuals).

What we would need to know in order to accept that assumption here is whether an analysis like this is just as likely to yield a list of pathogenic variants to explain any laundry list of symptoms that are shared across many chronic diseases. If the cohort included people with known pathologies causing fatigue, myalgia, muscle weakness, exercise intolerance, brain fog etc. (lupus, MS, hypothyroidism...) and we blinded your analysis to disease-specific labels/symptoms/clinical results, would we get a similar list of high-confidence pathogenic variants that "explain" the already-explained symptoms?

There are plenty of methods that incorporate a-priori knowledge with genetic data to better differentiate between signal and noise, so I don't think that bias is necessarily a bad thing. But the really important part is the additional validation steps to make sure those assumptions hold up. That's why I made the (admittedly blithe-sounding) comparison to looking for sodium-high deli meat in a heart disease case--to point out the degree of remove between the finding and how sure you can be that its causally relevant in specific cases. Finding heart disease-related variants in the cases of people with heart disease is only comparable to this study if you can be reasonably sure that these heterozygous pathogenic variants are more likely to explain ME/CFS symptoms than an underlying non-congenital disease process. I'm not sure that's a justifiable assumption.

I absolutely feel that - at the molecular level - ME/CFS is not one, but rather perhaps even hundreds of disease with convergent symptoms..

I guess I am with @V.R.T. in being skeptical about this belief. Plenty of chronic illnesses with known pathology are similarly heterogenous as ME/CFS. And plenty of people harbor pathogenic variants and are fine. So neither heterogeneity of disease or presence of potentially pathogenic variants seems a good indication that there is no unifying disease process. Can you explain more why you think this is true?

 

[Trish]

I add my thanks to @LizWorthey for joining us to discuss your research and answer questions. I hope you will stick around and join in other research discussions here relevant to your field of research when you have time too.