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

[Yann04]

I’m struggling to keep up so apologies if this question has already been asked but I’m wondering if the published data can tell us whether ME/CFS is likely to one disease which involves the immune system and nervous system, or whether it is possible/likely that there may be two diseases – one which is predominantly associated with the immune system and another which is predominately associated with the nervous system. Or even if there may be more than two diseases.

Theoretically that should be testable right? Seeing if the immune hits are coming from the same people as the nervous hits or it looks like two different populations?

Haven’t seen anything in the paper regarding it tho I’ve just skimmed a tiny bit.

 

[Ariel]

I think this is normal for GWAS, they provide clues or pointers to what's going wrong rather than the gene being the culprit behind ME/CFS.

Does anyone have any examples of other kinds of percentage differences for signals in other diseases to compare? It's hard to think about without this context.

 

[jnmaciuch]

Would it indicate that whatever we’re looking for, it’s something that doesn’t need these genes to go wrong, they just make it more likely that it goes wrong?

That would be my guess--it seems to be the case for many diseases where we do understand the mechanistic basis and have similarly weak associations.

 

[Kronos]

To get a feel of the effect size, I've made an overview of the prevalence of these SNPs in patients versus controls. I couldn't find this in the paper or supplementary material so I've tried to calculate them using R (trying out different guesses until the combined prevalence and odds ratio match - hopefully someone will check if they get the same result!).

It shows that the effect size comes down to a 1-2% difference in prevalence of SNPs between ME/CFS patients and controls.

SNPid​	Combined prevalence​	Odds ratio​	Prevalence ME/CFS​	Prevalence controls​
chr1q25.1​	0.325​	0.927​	0.3095​	0.3259​
chr6p22.2​	0.261​	1.086​	0.2763​	0.2601​
chr6q16.1​	0.546​	0.934​	0.5300​	0.5470​
chr12q24.23​	0.139​	1.100​	0.1501​	0.1383​
chr13q14.3​	0.287​	1.077​	0.3015​	0.2861​
chr15q21.3​	0.312​	1.082​	0.3282​	0.3110​
chr17q22​	0.330​	1.084​	0.3470​	0.3290​
chr20q13.13​	0.634​	1.095​	0.6536​	0.6328​

I only have basic knowledge about GWAS.

But due to the huge number of hypothesis being tested, isn't this always been accounted for in a statistical analysis? So if this is accounted for, and these genes still come up as somehow significant, The standard deviation for a gene should be a lot lower than the difference in prevalence presented here.
Otherwise those genes would simply be expected as random false positives?

I don't think this really tells us anything about if ME is multiple diseases or a singular disease (which I personally very much doubt).

 

[Amw66]

I don’t think these results can really answer the question either way. But plenty of other diseases, even ones with a more homogenous presentation and a clear biomarker, don’t have “smoking gun” associations.

Would stronger hits perhaps be found when the wider genetic areas around these genes are explored ?

 

[Yann04]

But due to the huge number of hypothesis being tested, isn't this always been accounted for in a statistical analysis? So if this is accounted for, and these genes still come up as somehow significant, The standard deviation for a gene should be a lot lower than the difference in prevalence presented here.
Otherwise those genes would simply be expected as random false positives?

Yeah. They did Multiple Test Correction.

 

[Barry]

An oversimplified explanation I sent to someone else asking the same question on another group:

"Basically they took 15k with Me/cfs and looked at what genes were more common/uncommon in pwME.

So those genes still exist in healthy people, just that theyre more common/uncommon in pwME. And that might give us clues, like do those genes contribute to factors that makes developing ME more likely"

Yes, I guess the point is that there is a statistically significant higher occurrence of this genetic anomaly within the ME/CFS community compared to the general population, but for any one individual with that genetic anomaly, that would not be sufficiently unique to say they had ME/CFS. But maybe as other anomalies become known, this might become a useful contributor for identifying someone with ME/CFS.

 

[Amw66]

Theoretically that should be testable right? Seeing if the immune hits are coming from the same people as the nervous hits or it looks like two different populations?

Haven’t seen anything in the paper regarding it tho I’ve just skimmed a tiny bit.

Or perhaps it links to severity ?

 

[wastwater]

I have come across rabgap genes in this glaucoma paper

FOXC1 modulates MYOC secretion through regulation of the exocytic proteins RAB3GAP1, RAB3GAP2 and SNAP25 - PMC

The neurodegenerative disease glaucoma is one of the leading causes of blindness in the world. Glaucoma is characterized by progressive visual field loss caused by retinal ganglion cell (RGC) death. Both surgical glaucoma treatments and medications ...

pmc.ncbi.nlm.nih.gov

Some of the transcription factors are interesting
Eg foxo3 pax2/3 sox5

 

[5M1TH]

Does this mean that everyone with ME/CFS has atleast a combo of these gene variants or they can still have people with ME/CFS who don’t have any of these ?

 

[JonBiz]

Does this mean that everyone with ME/CFS has atleast a combo of these gene variants or they can still have people with ME/CFS who don’t have any of these ?

As I understand it from this thread, the average difference in prevalence between controls and CFS patients was 1.2%. So it's just a slight difference in propensity.

 

[jnmaciuch]

Would stronger hits perhaps be found when the wider genetic areas around these genes are explored ?

I think it's unlikely--it would have to be an allele with <1% minor allele frequency and one that was missed in whole genome studies as well.

 

[jnmaciuch]

Does this mean that everyone with ME/CFS has atleast a combo of these gene variants or they can still have people with ME/CFS who don’t have any of these ?

Hi @5M1TH welcome to the forum! These gene variants to not discriminate between ME/CFS and controls, they're just sliiightly more common in the group of ME/CFS participants vs. controls in this study. Entirely possible to have ME/CFS and to have none of the alleles identified here.

 

[Utsikt]

Does this mean that everyone with ME/CFS has atleast a combo of these gene variants or they can still have people with ME/CFS who don’t have any of these ?

The genes are not required to develop ME/CFS, they just make it slightly more likely.

So it’s not like e.g. Huntington’s Disease where you either have a mutation in the gene and are guaranteed to get HD, or you don’t have it and will never develop HD. (Edit: HD is a bit more complicated than this binary example, but the point still stands - the gene has to be involved for HD to occur.)

That doesn’t mean that the genes here aren’t important clues. They probably point in the direction of what’s going wrong in ME/CFS. The job now is to figure out how a bit more of this and a bit less of that together result in ME/CFS being more likely.

 

[wigglethemouse]

Has there been any comment of the previous finding of SLC25A15 in women in the UK Biobank ME/CFS dataset highlighted in the Joshua Dibble thesis, a member of Chris Pontings team. I couldn't find a reference in the new paper. Would be interesting to know the p-value in this study.

 

[Hutan]

Would members find it useful if we set up separate discussion threads for the 8 genetic hits plus a few more from the list of 29 in the Candidate genes paper, so that we can poke into each one a bit more at leisure?

 

[Alinda]

Would members find it useful if we set up separate discussion threads for the 8 genetic hits plus a few more from the list of 29 in the Candidate genes paper, so that we can poke into each one a bit more at leisure?

I think that would be very helpful for looking into it further, maybe if more research happens that could be discussed in those threads as well (if its directly related)?

 

[Hutan]

The limited replication of the findings when tested against other databases is disappointing. There's been some public commentary about that. It's suggested that this indicates that there are a lot of people with CFS and ME/CFS labels in many database who don't actually have ME/CFS, and is a reflection of the care taken to characterise DecodeME participants well. I think that is likely to be true.

If so, then it means that research trying to find biomarkers, or talking about risk factors is going to fail unless there is really careful selection of participants. It's probably even worse with Long Covid.

 

[forestglip]

Would members find it useful if we set up separate discussion threads for the 8 genetic hits plus a few more from the list of 29 in the Candidate genes paper, so that we can poke into each one a bit more at leisure?

I think that's a good idea.

 

[Alinda]

The limited replication of the findings when tested against other databases is disappointing. There's been some public commentary about that. It's suggested that this indicates that there is a lot of people with CFS and ME/CFS labels in many database who don't actually have ME/CFS, and is a reflection of the care taken to characterise DecodeME participants well. I think that is likely to be true.

If so, then it means that research trying to find biomarkers, or talking about risk factors is going to fail unless there is really careful selection of participants. It's probably even worse with Long Covid.

I am hoping future research will more frequently follow proper criteria, otherwise many studies will continue to be basically useless.. especially with long covid!