The genetic architecture of fibromyalgia across 2.5 million individuals, 2026, Kerrebijn et al

[Chandelier]

Summary from Bluesky:

ME/CFS Science​

@mecfsscience.org

1) I looked at the big genetic study on fibromyalgia (Kerrebijn et al. 2026) and how well it correlated with findings for ME/CFS in DecodeME.

Using European samples, the correlation was quite big: rg = 0.75.

A brief discussion of the implications

2) A first point is that this probably shows that the DecodeME results are measuring what they were supposed to measure. There was a concern that they might reflect selection bias or other confounders such as ancestry, intelligence, etc.



3) That doesn't seem the case. The genetic correlation with measures of education/intelligence in the UK biobank was much smaller (rg < 0.15). Some of the strongest correlations were for self-reported irritable bowel syndrome (rg = 0.75) and self-reported CFS (rg = 0.7).



4) That can't be a coincidence.

In the DecodeME sample of ME/CFS, patients, 41% had comorbid irritable bowel syndrome and 29% fibromyalgia. So a high correlation with these other syndromes was expected. It's a like a test to check if DecodeME is on target.



5) These other genetic studies used very different inclusion strategies (often data from multiple sources) than DecodeME. So their high correlation couldn't have happened by chance IMHO.

A big question, however, is: does it mean that ME/CFS and fibro are related pathologies?



6) The correct way to test this is to calculate the genetic correlation between the fibro study and DecodeME - but with the 29% of ME/CFS patients with comorbid fibromyalgia excluded.

A correlation of 0.75 is very high though, so I suspect it does mean shared/similar pathology.



7) DecodeME and the fibro study also share some of their top loci.

The first graph, for example, shows the hits around the gene HTT. This was the most significant one in fibro but the same signal is present in DecodeME as well (just below the significance threshold).



8) I used the R packacke coloc to test for 'colocalization'. It tests if both signals are the same: whether they share the same causal SNPs or if whether they just look alike.

The the result suggested they are indeed the same (PP.H4 = 99.4%).



9) Another interesting region is above the genes DARS2 and RABGAP1L. The signal also looks similar but the colocalization analysis wasn't confident about it. I suspect there might be multiple separate signals at play here.

10) Interestingly I found no coloc match for the regions around the genes DCC (PP.H3 = 80%) and OLFM4 (PP.H3 = 85%).

I've posted the full analysis (including my attempt at SuSiE-coloc that allows for multiple causal variants) here:
s4me.info/threads/th...

 

[Colin]

Fibromyalgia is not in your head (9m)

ABC Radio National: Health Report

Guest: Frances Williams, Professor of Genomic Epidemiology, Kings College London

Presented by Norman Swan

 

[Nightsong]

Perhaps others can double-check but I calculated the genetic correlation with DecodeME and the European-only GWAS of fibromyalgia from this study using LDSC and I got a value of rg = 0.757. That's quite high and similar to the rg = 0.751 between DecodeME and self-reported irritable bowel syndrome in the UK Biobank.

Trying to catch up on these threads; just to confirm - about a month ago I ran LDSCs with a few different studies' datasets &, for some, a few additional stratified analyses; for this fibromyalgia one:

pair rg se overlap s DME x FMS-615 (no UKB) 0.7713 0.0336 -0.00508 0.00537 DME x FMS-603 (UKB in) 0.7587 0.0311 -0.00200 0.00545 female x female 0.7890 0.0346 -0.00203 0.00542

i.e. using the -615 accession (European, leaving UKB out) I got rg=0.7713. Using -603 (European with UKB in) I got rg=0.7587. Both studies also published female-only data, so I ran that as well: rg=0.7890. This sits above the result reported in the FMS paper for the symptom of malaise & fatigue (rg=0.60). (I ran regional colocalisations, too, but that's a longer post for another day.)

 

[voner]

A neurogenetic blueprint for fibromyalgia

Diatchenko, Luda; Jahangiri Esfahani, Sahel

Abstract
The largest genetic study of fibromyalgia thus far identifies 26 risk loci across 2.5 million
individuals, providing a powerful framework for investigating disease biology. The new
map offers experimentally tractable targets while sharpening questions about peripheral
neuronal contributions, immune involvement, sex differences and ancestry-related gaps in the
current knowledge of fibromyalgia.

paywall:
Web | DOI | PDF | Nature Reviews Rheumatology

if you have Research Gate access:

https://www.researchgate.net/publication/414535836_A_neurogenetic_blueprint_for_fibromyalgia

 

[voner]

from the paper:

Box 1 | Recurrent loci across fibromyalgia and related pain phenotypes

CAMKV (fibromyalgia, widespread pain, multisite pain and chronic overlapping pain)

DCC (fibromyalgia, widespread pain, multisite pain and chronic overlapping pain)

MAML3 (fibromyalgia, widespread pain and multisite pain)

OLFM4 (fibromyalgia, multisite pain and chronic overlapping pain)

SEMA3F (fibromyalgia, multisite pain and chronic overlapping pain)

ANAPC4 (fibromyalgia and multisite pain)