Preprint Initial findings from the DecodeME genome-wide association study of myalgic encephalomyelitis/chronic fatigue syndrome, 2025, DecodeMe Collaboration

[ME/CFS Science Blog]

Are these in any particular order?

Yes ranked according to p-value of the locus where they were predicted to be the causal gene. The loci with the lowest p-value (strongest signal) are on top. The Excel file on the other thread gives more information.

 

[Jonathan Edwards]

Isn't "causally related to ME/CFS status and not anything else" a bit too strong? From what I understand these are possible genes involved in genetic risk for ME/CFS, not proven causes of the disease or of specific symptoms.

No it is not too strong. Genetic risk = causal.
But causation is extremely complex, so we are not looking for single causes of any particular symptom. No disease has a single cause. An uncountable number of factors always contribue to a given case. The task is to tease out significant risk factors. Genes are one, but in total it looks as if they contribute around 10-20% of total causation. That leaves stochastic internal factors and environmental factors, which may be infrequent and explaining why some people are affected and not others but may also be universal and permissive, in which case they will not show up on epidemiological studies at all.

 

[mariovitali]

No it is not too strong. Genetic risk = causal.
But causation is extremely complex, so we are not looking for single causes of any particular symptom. No disease has a single cause. An uncountable number of factors always contribue to a given case. The task is to tease out significant risk factors. Genes are one, but in total it looks as if they contribute around 10-20% of total causation. That leaves stochastic internal factors and environmental factors, which may be infrequent and explaining why some people are affected and not others but may also be universal and permissive, in which case they will not show up on epidemiological studies at all.

I see, Thank you and also to @EndME for the explanation.I do not wish to derail the thread.

As a patient, I think it would be important to make clear to other patients these details perhaps via a webinar ? For example, I still do not understand whether these results essentially tell us to focus on the brain for a causal mechanism and pay much less attention or even disregard everything else (e.g. immune/metabolic abnormalities) as contributing to the symptoms of this disease.

 

[Andy]

Selected questions and answers from the DecodeME FAQs, available here, that were created to accompany our preprint last year.

Have you found a gene for ME/CFS

ME/CFS is a complex condition, so from the beginning of DecodeME, we didn’t expect to find just one gene responsible. Like other conditions such as Parkinson’s or diabetes, ME/CFS is influenced by small differences in several genes rather than in one gene only. In DecodeME, we looked at over 8 million points across the genome and found 8 regions, where people with ME/CFS were more likely to have certain genetic differences. These regions give us important clues about the biological processes that may be involved in the illness.

What do these genes do?

We found that people with ME/CFS are more likely to carry certain DNA differences in eight regions of their genome, and these variants tell us about possible biological causes of ME/CFS. Most of these regions contain several genes. Our methods did not allow us to conclusively locate the ones most relevant to ME/CFS in each region, but public data allowed us to pick out the most likely ones. Three of the most likely genes produce proteins that respond to an infection. Another likely gene is related to chronic pain. Overall, our initial genetic results show that ME/CFS is partly caused by genes related to the immune and nervous systems.

Do these genes cause ME/CFS?

The genes we’ve identified don’t directly cause ME/CFS on their own, but their genetic differences do increase the likelihood of someone develops ME/CFS, much like how certain genes are linked to conditions such as diabetes or Parkinson’s. They are an important piece of the puzzle, but not the whole picture. Having one or more of these genetic differences does not fully determine whether someone will get ME/CFS. There are likely many contributing causes of ME/CFS, but these findings help us point towards more specific biological pathways involved in the development of the condition.

 

[Jonathan Edwards]

For example, I still do not understand whether these results essentially tell us to focus on the brain for a causal mechanism and pay much less attention or even disregard everything else (e.g. immune/metabolic abnormalities) as contributing to the symptoms of this disease.

I think we can be pretty sure that the causal mechanism goes way out into the peripheral tissues in terms of at least nociceptor nerves, even if we are dealing with disease persistence entirely due to a shift in brain microstructure/activity. And nothing in the DecodeME findings is against disease persistence being held in a signalling loop that involves peripheral immune signalling too. The BTN2A1 link gives a prima facie case for an immune component that might be at initiation but might also be a long term shift in T cell or NK cell populations.

Thresholds for signalling could also be affected by metabolic shifts - in lipids maybe. I don't at all discount that. It is just that I think it is harder to build a model at a chemical level when we are almsot certainly dealing with acquired shifts in signalling at a very high level in a complex regulatory system (like brain or immune cells).

I don't think anyone is suggesting disregarding anything. The focus is on the brain-related gene variants because they are by definition antecedent and causal at least to some degree. Pretty much all other findings can be downstream rather than upstream and so less informative.

 

[Robert 1973]

The focus is on the brain-related gene variants because they are by definition antecedent and causal at least to some degree.

And they, along with BTN2A1, are the variants that are implicated so far from DecodeME data. But I think it’s important to add that, as far as I understand, SequenceME will almost certainly identify other gene variants which are more or less common in people with ME/CFS – and it is likely (?) that some of them will not be brain-related.

It may also enable us to identify whether there are subgroups with different predisposing genes and different disease mechanisms.

Is that correct?

 

[mariovitali]

Pretty much all other findings can be downstream rather than upstream and so less informative.

I agree with everything apart from the text I quoted. I came across patients having hyperammonemia and this would definitely be an upstream cause as we know that elevated ammonia affects negatively the synapses.

https://pubmed.ncbi.nlm.nih.gov/1354386/
EDIT : I do not suggest that all ME/CFS patients have hyperammonemia.

Please, please do not dismiss these events that easily.

 

[Jonathan Edwards]

I agree with everything apart from the text I quoted. I came across patients having hyperammonemia and this would definitely be an upstream cause as we know that elevated ammonia affects negatively the synapses.

I don't think you can be sure. Let us say that ME/CFS induces shifts in neural or endocrine signals that alter metabolism in some way that can in some cases raise ammonia levels. The fact that very high ammonia levels impact on cognition in liver failure and can be shown to affect synapses in experimental systems does not ensure that any ME/CFS symptoms are downstream of ammonia levels.

Pathological mechanisms are often extremely complicated and confusing. Standard explanations for cliinical features taught to medical students are often wrong. For instance, every student is taught that oedema in nephrotic syndrome is due to low albumin levels. But it cannot be because oedema often occurs before there is any fall in albumin. Moreover, the same oedema occurs in nephritic syndrome where the albumin stays normal. Students are taught that you cannot get double vision when looking through one eye but this is not so. Astigmatism can lead to visual cortex interpreting a single image as two 'best fit' images. And so on and so on.