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

[Jonathan Edwards]

I don't actually know where sex ratio figures in h2 estimates - whether it is hived off as a separate issue.

 

[forestglip]

However, my memory is that other data fit with ~10% and Chris gave the impression (I thought) that there were reasons to think this was most of the risk.

My impression was that we don't really have good total heritability data. Snippet from 2020 paper from Dibble, Ponting, and McGrath:

Genetic risk factors of ME/CFS: a critical review, 2020, Human Molecular Genetics

Of three studies that have estimated narrow-sense heritability (h2) using large cohorts, two reported non-zero h2-values that provide evidence for heritability of risk for CFS and, presumably, ME/CFS.

An analysis of US health insurance claimed a high narrow-sense heritability (⁠⁠h2 =0.48) of CFS (23),

whereas an analysis of the UK Biobank individuals self-reporting a CFS diagnosis reported a less striking heritability (single nucleotide polymorphism- [SNP-] based approximate h2 = 0.08 with low confidence) (24) (http://www.nealelab.is/uk-biobank).

The third, a large twin-based study of CFS-like cases, produced an inconclusive result, with the 95% confidence interval of h2 including zero [0.03 (0.00–0.65)] (25).

So only two of these were not inconclusive. One of these (UK Biobank, 8%) is only SNP-based heritability, like DecodeME's figure. I believe the 48% from the insurance study would include more forms of genetic heritability, though I don't know anything about the study or patient selection.

My guess is that if genes are really rare they don't make much difference and if they aren't rare then it is a bit unfair if none of them show up on the GWAS, but others will know more than I do.

Not saying I know much either, and I hope an expert can weigh in. But in the image I shared it says very large proportions of heritability are as yet unattributed in all three diseases, larger than the amount determined from the common variants in GWAS. (Though looking at the figure, the lengths of the bars are kind of wonky. In AD, the 32% bar for unattributed risk is longer than the 35% bar for non-APOE common variants.)

I'm not sure why it'd be unfair if it didn't show up on a GWAS. There might be recessive and dominant alleles that contribute to risk. There might be copy number variants. The GWAS is only testing some forms of genetic risk.

 

[Jonathan Edwards]

My impression was that we don't really have good total heritability data.

It's a bit of a mess, yes, but I read it like this:
DecodeME points to 9.5%+ .
The Biobank data at 8%+ is similar.
The insurance twin study doesn't look plausible. If h2 was 48% I think that would have been clinically obvious. It is a bit hard to know why this one should be so out of line but I am inclined to either ignore it or allow that the others may be missing some rare stuff. But then the other twin study is even lower at 0-6%. That also seems a bit implausible since we do see multicase families more often than by chance it seems. But if it was 6% it wouldn't be so very far out of line.

Maybe we have to allow for 10-25% being plausible but my memory from the diseases I worked with is that rare genes don't add that much. RA is mostly XX, DRbeta and PTPN22.

 

[forestglip]

But then the other twin study is even lower at 0-6%.

Noting that it was 0-65%.

 

[Jonathan Edwards]

Noting that it was 0-65%.

Ah, that's a confusing figure. So it is a best fit of 3% with very wide uncertainty, including 10-25%!
I still go for 10-25%.