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

[Andy]

Apologies if this has been answered already: is there any hope of looking for genetic clues in early onset cases of the DecodeME data? Was age-of-onset data collected? (The thinking being that early-onset cases might have a stronger genetic predisposition, or something else distinct going on.)

Age of onset wasn't specifically collected, but date of birth and length of illness was. This isn't something that the remaining DecodeME team would be able to look at but if a researcher wanted to apply for the necessary data access then the process can be found here.

 

[Jonathan Edwards]

Apologies if this has been answered already: is there any hope of looking for genetic clues in early onset cases of the DecodeME data? Was age-of-onset data collected? (The thinking being that early-onset cases might have a stronger genetic predisposition, or something else distinct going on.)

I have discussed this with Chris. His problem is that the more post-hoc analyses of the DecodeME dataset that are done, the less statistical validity you have.

On the other hand, I would personally like to see someone do the necessary math (that is all you need) on data accessed from DecodeME and jut see what turns up. I think researchers are entitled to play around with the data as much as they like if that leads to some sharply focused questions for a further study of another cohort. If you have much more focused questions then the massive p value corrections needed for GWAS, and the consequent need for huge samples, do not arise.

 

[Jonathan Edwards]

I guess that there is another potential way to argue around the statistical Catch22 that comes with more analyses. You might be able to argue that if the 8 DNA segments identified are robust genetic markers then they ought to show up at least as strongly for very early onset cases, if not more strongly. The analysis could be seen as quality control!

 

[jnmaciuch]

I guess that there is another potential way to argue around the statistical Catch22 that comes with more analyses. You might be able to argue that if the 8 DNA segments identified are robust genetic markers then they ought to show up at least as strongly for very early onset cases, if not more strongly. The analysis could be seen as quality control!

I just attended a lecture on this exact question in child vs. adult onset asthma. Different sets of genes were predictive in either case. There was some small amount of overlap and I believe that the risk loci from one cohort did have some association in the other cohort when collapsed into an overall score, but it was definitely a much weaker signal. Risk loci from GWAS generally tend to be heavily context specific

[Edit: pretty sure this is the paper that some of the discussed data came from:

Integration of functional genomics and statistical fine-mapping systematically characterizes adult-onset and childhood-onset asthma genetic associations - Genome Medicine

Background Genome-wide association studies (GWAS) have identified hundreds of loci underlying adult-onset asthma (AOA) and childhood-onset asthma (COA). However, the causal variants, regulatory elements, and effector genes at these loci are largely unknown. Methods We performed heritability...

link.springer.com

And a press release with summary:

Research fine tunes tools used to search for genetic causes of asthma

New study from UChicago combines genetic data and new computational tools to identify likely genetic variants that cause childhood- and adult-onset asthma.

biologicalsciences.uchicago.edu

]

 

[Jonathan Edwards]

I just attended a lecture on this exact question in child vs. adult onset asthma.

I think there are reasons for thinking that chldhood onset asthma may have a very different origin from adult onset. It is a long time since I heard anything about it but that was my impression. For ME/CFS we tend to assume the same mechanism I guess, even if there seem to be two age peaks.

But it adds another burning question maybe - would the 8 regions replicate at all in people with onset under 13? I am not sure why there isn't more interest in doing the analyses but I respect Chris's position that he cannot hope to publish analyses that are just scattershot.