Skip to content

Omics and emerging science in fibromyalgia

TL;DR — Fibromyalgia (FM) acquired a genome in 2026. A multi-ancestry GWAS meta-analysis across 2,563,755 individuals (54,629 cases) identified the first 26 risk loci, with heritability enriched exclusively in brain tissues and neural cell types, the strongest signal at a coding variant in HTT, and genetic correlations above 0.7 with low back pain, PTSD and IBS — establishing FM as a central nervous system disorder on genetic evidence rather than inference (Kerrebijn 2026, PMID 42521817). This displaces a decade of underpowered candidate-gene and small-GWAS work, including an explicitly null FM GWAS from a hospital biobank (Moscati 2023, PMID 37334860). The second major line is the gut microbiome: composition differences with machine-learning classification at AUC 87.8% (Minerbi 2019, PMID 31219947), serum bile-acid alterations tracking symptom severity (Minerbi 2023, PMID 35587528), and — most consequentially — faecal transplant from FM patients into germ-free mice inducing pain, with an open-label human transplant trial reporting reduced pain (Cai 2025, PMID 40280127). That causal claim now sits against a strictly-controlled null study finding no gut or oral microbial signature at all (Celik 2026, PMID 42550534). Transcriptomics, proteomics and metabolomics remain at the single-study, discovery stage. Precision medicine in FM is, honestly assessed, at the stage of having its first credible biological hypotheses — not its first stratified trial.

1. Genetics: from candidate genes to a real GWAS

1.1 The 2026 fibromyalgia GWAS

Feature Value
Design Multi-ancestry GWAS meta-analysis, 11 cohorts
Sample 2,563,755 individuals (54,629 cases, 2,509,126 controls)
Loci 26 genome-wide significant risk loci — the first for FM
Top signal Coding variant in HTT (the Huntington's disease gene)
Prioritised genes GPR52 (an HTT regulator), DCC, DRD2/NCAM1, MDGA2, CELF4, CAMKV
Heritability enrichment Exclusively brain tissues and neural cell types
Genetic correlations > 0.7 with low back pain, post-traumatic stress disorder, irritable bowel syndrome
Sex Genetic architecture nearly identical in males and females despite large prevalence difference

Source: Kerrebijn 2026 (PMID 42521817); the preprint of the same analysis is PMID 41001472.

Four implications deserve emphasis:

  1. "Central nervous system disorder" is now a genetic finding. Heritability partitioning that is enriched in brain and neural cell types and nowhere else is a strong constraint. It does not exclude peripheral contributions to symptoms (see peripheral pathophysiology), but it locates the inherited liability centrally.
  2. The HTT signal is unexpected and unexplained. A coding variant in the Huntington's gene as the strongest FM association, with the HTT regulator GPR52 prioritised, is the single most striking result in the paper and has no established mechanistic account in pain biology. GPR52 is a druggable orphan GPCR, which makes this the most immediately actionable target the field has ever had.
  3. The sex paradox sharpens. FM prevalence differs several-fold between women and men, yet the genetic architecture is nearly identical. Whatever produces the sex ratio is therefore largely non-genetic — ascertainment, criteria performance, hormonal, or environmental (see epidemiology and diagnostic criteria).
  4. The comorbidity structure is genetic, not just clinical. rg > 0.7 with low back pain, PTSD and IBS means these conditions share most of their common-variant liability with FM — direct genetic support for the chronic overlapping pain conditions framework in comorbidities and overlap.

1.2 The pre-2026 record and why it failed

Before the large meta-analysis, biobank-scale attempts at an FM-specific GWAS were underpowered. A study leveraging Swedish register and hospital-biobank data found FM prevalence under 1% in women and roughly a tenth of that in men, confirmed comorbidity clustering into pain-related, autoimmune and psychiatric groups, and showed that polygenic scores for psychiatric, pain-sensitivity and autoimmune traits were associated with FM — but its own genome-wide analysis of FM yielded no genome-wide significant loci, with the authors concluding that larger samples were required (Moscati 2023, PMID 37334860). A health-system GWAS of post-treatment Lyme disease syndrome identified 5,041 FM cases and 2,268 ME/CFS cases in the same biobank and found that neither of its two suggestive PTLDS loci was associated with FM or ME/CFS (Hirsch 2025, PMID 39994562) — a useful negative on shared post-infectious genetics.

1.3 Chronic widespread and multisite pain GWAS

FM-adjacent pain phenotypes were genotyped years before FM itself, and they anticipated the 2026 result.

Study Phenotype Sample Key result
Johnston 2019 (PMID 31194737) Multisite chronic pain (MCP), UK Biobank ~380,000 SNP-heritability 10.2%; 76 independent lead SNPs at 39 loci; gene-set enrichment for neurogenesis, synaptic plasticity, nervous-system development, cell-cycle and apoptosis; genetic correlations with MDD, asthma, BMI; Mendelian randomisation supporting a causal effect of MCP on MDD; MCP polygenic score predicted chronic widespread pain
Johnston 2021 (PMID 33830993) MCP, sex-stratified 178,556 men, 209,093 women 5 independent loci in men, 10 in women, 87 further SNPs on meta-analysis; 37 genes in men, 31 in women, only DCC shared; male–female rg = 0.92 (significantly < 1); nearly all associated genes expressed in dorsal root ganglion
Rahman 2021 (PMID 33926923) Chronic widespread musculoskeletal pain, UK Biobank + 6 replication cohorts 6,914 cases / 242,929 controls; replication n = 43,080 3 loci; RNF123 replicated (meta-analysis p = 0.0002), a second locus suggestive (p = 0.0227), third not replicated; partial genetic correlation with depressive symptoms, BMI, age at first birth, education; tissue-specificity analysis highlighted skeletal muscle; the most-studied candidate gene in the chronic pain field failed replication
Pan 2025 (PMID 40509746) "Pain all over the body" past month, UK Biobank 172,230 One novel locus at 5q13.2 (rs34691025); a female-specific locus on chromosome 10; genetic correlations with joint disorders and spondylosis

Two points of tension are worth carrying forward. DCC appears in both the MCP sex-stratified analysis (the only gene shared between sexes) and in the FM gene-prioritisation list — a rare cross-phenotype convergence. But the Rahman CWP analysis pointed at skeletal muscle tissue specificity, which sits awkwardly against the FM GWAS's brain-exclusive heritability enrichment; whether this reflects genuinely different phenotypes ("pain all over the body" vs. diagnosed FM) or methodological difference is unresolved.

1.4 FM in other conditions' genetics

FM now appears as a genetic correlate in GWAS of neighbouring conditions. A multi-ancestry GWAS of dry eye disease (132,657 cases) reported substantial genetic correlations with fibromyalgia, PTSD and Sjögren's disease, and derived a latent factor underlying dry eye and other chronic pain traits accounting for 51% of dry eye's genetic variance (Gorman 2025, PMID 40166553). A GWAS meta-analysis of hypermobile Ehlers–Danlos syndrome (1,815 cases) found significant genetic correlations with joint hypermobility, ME/CFS, fibromyalgia, depression, anxiety, autism spectrum disorder, migraine and gastrointestinal disease, with a lead regulatory signal near the atypical chemokine receptor ACKR3 colocalising with tibial-nerve eQTLs (Petrucci-Nelson 2025, PMID 41001447; preprint).

The pattern across all of these is one latent liability to widespread symptom burden, shared across nominally separate diagnoses — which is a genetic restatement of the nosological problem in history and nosology.

2. Gut microbiome: the field's most developed causal story, and its counterexample

2.1 The discovery line

The originating study compared microbiomes of 77 women with FM and 79 controls using 16S rRNA amplification and whole-genome sequencing. Several taxa differed in abundance; variance in microbiome composition was explained by FM-related variables more than by any other innate or environmental variable and correlated with clinical indices; targeted serum metabolite analysis confirmed altered butyrate and propionate concentrations, consistent with the observed changes in butyrate-metabolising species; and machine learning on microbiome composition alone classified patients versus controls at ROC AUC 87.8% (Minerbi 2019, PMID 31219947). This was the first demonstration of gut microbiome alteration in a non-visceral pain condition.

The same group then closed the most obvious confound. Dietary intake was assessed by 24-hour recall and three-day dietary records in 56 women with FM and 68 controls: energy and macronutrient intake, diet-quality score, vitamins, minerals, fatty acids, alcohol, caffeine, sugar and fibre did not differ between groups, and micronutrient intake correlated with neither symptom severity nor the relative abundance of almost any of the differentially abundant taxa — so diet is unlikely to explain the FM microbiome signature (Minerbi 2022, PMID 35328942).

Next came a metabolite bridge. In 42 women with FM and 42 controls, 16S sequencing plus targeted metabolomics found altered abundance of bile-acid-metabolising bacteria alongside significant changes in serum secondary bile acids, with marked depletion of α-muricholic acid; statistical learning on serum bile acid concentrations accurately detected FM cases; and serum α-muricholic acid was highly correlated with symptom severity including pain intensity and fatigue (Minerbi 2023, PMID 35587528).

2.2 The causal experiment

The line's decisive experiment: faecal microbiota transplantation from FM patients — but not from healthy controls — into germ-free mice induced pain and a set of molecular phenotypes paralleling those seen in patients, including immune activation and metabolomic alterations. Replacing the FM microbiota with a healthy microbiota substantially alleviated pain in mice. An accompanying open-label trial in women with FM (registry MOH_2021-11-04_010374) reported that transplantation of a healthy microbiota was associated with reduced pain and improved quality of life (Cai 2025, PMID 40280127).

This is the strongest causal evidence for any peripheral mechanism in FM. Its weight should be assessed carefully:

  • The germ-free mouse arm is a genuine causal demonstration in mice, with a rescue condition — a design that rules out reverse causation within the model.
  • The human arm is open-label and uncontrolled. FM is the condition with the largest documented placebo response in analgesic trials (see clinical trials landscape); an unblinded, high-ritual, high-expectation intervention reporting symptom improvement carries essentially no efficacy information. It is a safety and feasibility signal.
  • The trial is registered with an Israeli Ministry of Health registry, not on ClinicalTrials.gov; no NCT identifier was located for it in this session.

2.3 The counterexample and the state of replication

Replication has been inconsistent, and one recent study is a direct null.

Study Design Result
Minerbi 2019 (PMID 31219947) 77 FM / 79 controls, 16S + WGS Taxa differences; ML AUC 87.8%
Kim 2023 (PMID 36833885) 19 FM / 21 controls, 16S + faecal SCFA Lower observed OTUs, Shannon index and evenness in FM; significant beta-diversity differences on four metrics; propionate lower but only marginally (p = 0.069)
Lopez de Coca 2025 (PMID 41390899) 37 FM / 46 controls, 16S Reduced α-diversity (Shannon p = 0.009), distinct β-diversity (p < 0.001), SIBO in 59.5% vs 32.6%, increased Akkermansia, Oscillospira, Methanobrevibacter, Christensenellaceae
Durán-González 2025 (PMID 41113645) 199 FM / 43 controls, plasma proteome + faecal metagenome; parent study NCT05921409 (FIBROKIT, Pronacera Therapeutics, planned n=250, women 40–59, healthy-volunteer comparator with inflammatory/autoimmune/GI/metabolic disease excluded) 30 proteins and 19 taxa differential; integrated algorithm discriminates cases from controls
Celik 2026 (PMID 42550534) 16 FM / 15 controls, paired oral and faecal 16S, strictly standardised for diet, metabolic disease, medication, hormonal phase, recent infection No significant difference in α-diversity (all p > 0.05) or β-diversity (faecal R² = 0.032, p = 0.529; oral R² = 0.032, p = 0.464); no genus survived FDR correction; powered only to detect R² ≥ 0.11

The null study is small and its own authors state it could not detect subtle effects. But its design point is the important one: it controlled aggressively for comorbidity, medication, diet and cycle phase, and under those conditions the signature disappeared. Their proposed interpretation — that previously reported dysbiosis may reflect comorbidity-driven or phenotype-specific variation rather than a universal microbial hallmark — is the central open question of this line, and it is entirely compatible with the germ-free mouse result (a real causal effect in a subgroup would still be diluted to nothing in an unstratified 16-person comparison).

An interventional dietary crossover trial adds a different kind of evidence: 20 FM patients received ancient Khorasan wheat or control wheat products for 8 weeks each with an 8-week washout, and the Khorasan arm produced greater increases in butyric acid and shifts in specific taxa, with correlations between taxon abundance and tiredness, sleep, widespread pain index and symptom severity scores (Baldi 2022, PMID 35664958). Small, but it demonstrates that the FM microbiome is modifiable and that modification tracks symptoms.

A concise mechanistic synthesis of the whole line — short-chain fatty acids, bile acids, neurotransmitters and bacterial antigens entering host circulation as candidate routes — is available in review form (Minerbi 2020, PMID 32116215; Minerbi 2024, PMID 39726854).

3. Transcriptomics

This is the thinnest of the omics layers in FM. What exists is intriguing rather than established.

Whole RNA sequencing of peripheral blood mononuclear cells from participants in a manual-therapy trial (NCT04174300 — eight 25-minute pressure-controlled sessions over four weeks, planned n = 40, single centre, completed 2020) found that the salt-induced kinase gene SIK1 was consistently downregulated after treatment, and that the change correlated with symptom improvement; critically, the same protocol applied to a non-FM control cohort produced no change in SIK1 expression, making the change FM-specific rather than a generic response to touch (Bonastre-Férez 2024, PMID 39273470). The sample is small and the finding awaits replication, but a treatment-responsive, disease-specific transcript is exactly the shape of a pharmacodynamic biomarker.

From an adjacent field: gene co-expression network analysis of whole blood in systemic lupus erythematosus defined "type 2 SLE" molecular endotypes (B cell, metabolic and neuromuscular pathway enrichment) and found those same modules in subsets of patients with inactive SLE and idiopathic fibromyalgia, and in active SLE patients with severe fatigue (Robl 2023, PMID 36720488). This suggests a transcriptional state shared between FM and the fatigue-plus-pain presentation within autoimmune disease — relevant to the diagnostic overlap discussed in comorbidities and overlap.

No large, independent, case-control FM blood transcriptome study with replication was identified in this session's searches.

4. Metabolomics and proteomics

Both are covered in diagnostic terms in biomarkers; the mechanistic content is summarised here.

Metabolomics. The most informative design used metabolic and physical stress challenges rather than resting samples: 54 FM patients and 31 controls sampled at baseline and after an oral glucose tolerance test and a cardiopulmonary exercise test. Patients differed from controls in tyrosine and purine pathways at baseline and in the pyrimidine pathway after challenge; several discriminating markers relate to glutamatergic neurotransmission; and 5-hydroxyindole-3-acetic acid and glutamine correlated with fatigue severity (Zetterman 2024, PMID 38411371). The stress-challenge design is the methodological contribution — it suggests FM metabolic abnormality is a response phenotype rather than a resting-state difference, which would explain the inconsistency of resting metabolomic studies.

The bile-acid work (§2.1) is functionally metabolomics with a microbial upstream, and is the only FM metabolite finding with both a mechanistic origin story and a symptom-severity correlation.

Proteomics. Ten observational studies, 3,328 proteins, 145 differentially expressed, converging loosely on complement and coagulation cascades, iron metabolism and oxidative stress (Gkouvi 2024, PMID 38652420). The synthesis is dominated by heterogeneity in matrix (plasma, serum, CSF, saliva), platform and control group; there is no consensus FM proteome.

5. Wearables, digital phenotyping and machine-learning subtyping

Wearables. The promise is dense, longitudinal, objective measurement of the symptoms FM is defined by — activity, sleep, and their variability — replacing a single recall-based questionnaire. The reality is earlier-stage. A systematic review of commercially available activity monitors in musculoskeletal disorders (including FM) found ten eligible studies using ten different device types, with only three comparing trackers against objective reference tools, and concluded that device heterogeneity and differing algorithms currently prevent standardisation (Negrini 2021, PMID 33935067). No validated FM-specific digital phenotype was identified in this session's searches. Trials are beginning to embed continuous measurement — for example an at-home morning bright-light trial in nociplastic pain with sleep stabilisation and n = 390 (NCT06567886) — but the digital-phenotyping literature in FM remains a promissory note.

Machine-learning subtyping. Attempts fall into three families:

  • Imaging-based: resting-state connectivity plus structure separating FM from controls at accuracy 0.95 in n = 56 (Thanh Nhu 2022, PMID 36551758) — classification, not subtyping.
  • Symptom/record-based: random-forest models on electronic medical records achieving AUC 0.810 (Emir 2015, PMID 26089700); an 87-item triage questionnaire reduced to 28 items reaching AUC 0.81 for FM on external validation (Maarseveen 2025, PMID 41248315); a multimodal video-plus-language AI reaching 91% accuracy against rheumatic disease controls, with post-hoc-optimised thresholds explicitly requiring external validation (Venerito 2026, PMID 42442924).
  • Omics-based: metabolomic feature selection to 13 markers at 79% accuracy (Zetterman 2024, PMID 38411371); microbiome classification at AUC 87.8% (Minerbi 2019, PMID 31219947); integrated proteome-plus-metagenome algorithms (Durán-González 2025, PMID 41113645).

The recurring flaw is that nearly all of this is case–control classification presented as subtyping. Genuine subtyping would partition the FM population itself into groups with different biology or different treatment response, then validate the partition prospectively. That study has not been published.

6. Emerging mechanisms

Two mechanistic lines are advancing fastest and are documented in full elsewhere; their omics relevance is noted here.

Autoreactive IgG and satellite glia. Passive transfer of FM patient IgG into mice reproduces core FM phenomenology, with labelling of satellite glial cells and neurons in mouse dorsal root ganglia and binding to human DRG (Goebel 2021, PMID 34196305), extended to sensitisation of Aβ low-threshold mechanoreceptors (Israel 2025, PMID 40898777) and constrained by the finding that anti-satellite-glial-cell reactivity is present in post-COVID and recovered-COVID sera without the pronociceptive functional property (Berwick 2025, PMID 40408228). Full treatment in autoimmunity and inflammation.

The gut–immune–brain axis as a single mechanism. The FMT experiment's mouse phenotype included immune activation and metabolomic alteration alongside pain (Cai 2025, PMID 40280127) — meaning the microbiome line and the neuroimmune line may be one mechanism observed from two ends rather than two competing hypotheses. That reading is currently a hypothesis, not a demonstrated pathway.

7. Precision medicine in FM: an honest assessment

What would precision medicine in FM require, and how far along is each requirement?

Requirement Status
A biologically defined subgroup Not established. Candidates exist (small-fibre-positive, IgG-pronociceptive, microbiome-altered) but none is defined by a validated assay
An assay to assign patients to it Not available. No validated diagnostic or stratification biomarker (biomarkers)
Evidence that the subgroup responds differently to treatment Absent. No published FM trial has prospectively stratified on a biological marker
A druggable target from the biology Newly plausible. GPR52 from the 2026 GWAS is the first genetically-supported, druggable candidate FM target (PMID 42521817)
A trial design that can detect subgroup effects Available but unused — enrichment and randomised-withdrawal designs exist (clinical trials landscape) but are used for regulatory efficiency, not for biological stratification

The realistic sequencing is: GWAS-nominated targets (years to a first-in-human agent), microbiome interventions (already at open-label human stage but requiring blinded, controlled replication), and IgG-directed approaches (mechanistically strongest, with a functional-assay bottleneck). Nothing supports stratified prescribing today, and claims that FM is entering a precision-medicine era should be read against the fact that no biological stratifier has ever been prospectively tested in an FM trial.

Open questions

  • What is the mechanism linking HTT / GPR52 to nociplastic pain, and is GPR52 modulation analgesic? The association is the strongest in the FM genome (PMID 42521817) and has no mechanistic account.
  • Can the brain-exclusive heritability enrichment in FM (PMID 42521817) be reconciled with skeletal-muscle tissue specificity in chronic widespread pain (PMID 33926923)? Either the phenotypes differ biologically or one tissue-enrichment result is an artefact.
  • Why is FM's genetic architecture nearly identical between sexes (PMID 42521817) when its prevalence is not? This makes ascertainment and criteria performance (diagnostic criteria) the leading explanations for the sex ratio, and that has not been tested genetically.
  • Is there a reproducible FM gut microbiome signature, or is the published signature comorbidity- and phenotype-driven? Directly contested between PMID 31219947 / 41390899 and the strictly-controlled null PMID 42550534.
  • Does faecal microbiota transplantation reduce FM pain in a blinded, sham-controlled trial? The human evidence is currently open-label only (PMID 40280127), in the condition with the field's largest documented placebo response.
  • Is SIK1 downregulation a generalisable pharmacodynamic marker of FM treatment response, or specific to manual therapy? Single small study (PMID 39273470, NCT04174300), no replication.
  • Does any published classifier subtype the FM population — rather than separate FM from controls — in a way that predicts differential treatment response? No such study was identified.
  • Are FM metabolic abnormalities resting-state or stress-response phenomena? The challenge-based design found post-challenge pyrimidine-pathway differences absent at baseline (PMID 38411371); no replication exists.

References

  1. Baldi S, Pagliai G, Dinu M, et al. Effect of ancient Khorasan wheat on gut microbiota, inflammation, and short-chain fatty acid production in patients with fibromyalgia. World J Gastroenterol. 2022;28:1965-1980. PMID 35664958
  2. Berwick RJ, Sahbaie P, Kenny G, et al. Postacute COVID-19 syndrome and fibromyalgia syndrome are associated with anti-satellite glial cell IgG serum autoantibodies but only fibromyalgia syndrome serum-IgG is pronociceptive. Pain. 2025;166:e397-e408. PMID 40408228
  3. Bonastre-Férez J, Giménez-Orenga K, Falaguera-Vera FJ, Garcia-Escudero M, Oltra E. Manual therapy improves fibromyalgia symptoms by downregulating SIK1. Int J Mol Sci. 2024;25:9523. PMID 39273470
  4. Cai W, Haddad M, Haddad R, et al. The gut microbiota promotes pain in fibromyalgia. Neuron. 2025;113:2161-2175.e13. PMID 40280127
  5. Celik G, Yanik E, Inan N, Yalinay AM. Absence of a consistent gut or oral microbial signature in fibromyalgia under strictly controlled clinical conditions: a multi-compartment 16S rRNA analysis. Pain Physician. 2026;29:E407-E416. PMID 42550534
  6. Durán-González E, Ramírez-Tejero JA, Pérez-Sánchez M, et al. Fibromyalgia diagnosis from a multi-omics approach: a gut feeling. Front Microbiol. 2025;16:1641185. PMID 41113645
  7. Emir B, Masters ET, Mardekian J, Clair A, Kuhn M, Silverman SL. Identification of a potential fibromyalgia diagnosis using random forest modeling applied to electronic medical records. J Pain Res. 2015;8:277-88. PMID 26089700
  8. Gkouvi A, Tsiogkas SG, Bogdanos DP, Gika H, Goulis DG, Grammatikopoulou MG. Proteomics in patients with fibromyalgia syndrome: a systematic review of observational studies. Curr Pain Headache Rep. 2024;28:565-586. PMID 38652420
  9. Goebel A, Krock E, Gentry C, et al. Passive transfer of fibromyalgia symptoms from patients to mice. J Clin Invest. 2021;131:e144201. PMID 34196305
  10. Gorman BR, et al. Genome-wide association study of dry eye disease reveals shared heritability with systemic comorbidities. medRxiv preprint. 2025. PMID 40166553
  11. Hirsch AG, et al. A comparison of genome-wide association analyses of persistent symptoms after Lyme disease, fibromyalgia, and myalgic encephalomyelitis - chronic fatigue syndrome. BMC Infect Dis. 2025;25:265. PMID 39994562
  12. Israel MR, Berwick R, Vastani N, et al. Aβ low threshold mechanoreceptors contribute to sensory abnormalities in fibromyalgia. Brain. 2025;148:4016-4029. PMID 40898777
  13. Johnston KJA, Adams MJ, Nicholl BI, et al. Genome-wide association study of multisite chronic pain in UK Biobank. PLoS Genet. 2019;15:e1008164. PMID 31194737
  14. Johnston KJA, Ward J, Ray PR, et al. Sex-stratified genome-wide association study of multisite chronic pain in UK Biobank. PLoS Genet. 2021;17:e1009428. PMID 33830993
  15. Kerrebijn I, et al. The genetic architecture of fibromyalgia across 2.5 million individuals. Nat Med. 2026;32:3060-3070. PMID 42521817
  16. Kim Y, Kim GT, Kang J. Microbial composition and stool short chain fatty acid levels in fibromyalgia. Int J Environ Res Public Health. 2023;20:3183. PMID 36833885
  17. Lopez de Coca T, Muñoz-Almaraz FJ, Martinez-Blanch JF, Moreno L. Exploring gut microbiota alterations and their associations with clinical symptoms in fibromyalgia. Sci Rep. 2025;16:2327. PMID 41390899
  18. Maarseveen TD, Reimann F, Al Hasan A, et al. Evaluating the accuracy of the Frysian Questionnaire for Differentiation of Musculoskeletal Complaints for triage of musculoskeletal diseases. JMIR Med Inform. 2025;13:e77345. PMID 41248315
  19. Minerbi A, Gonzalez E, Brereton NJB, et al. Altered microbiome composition in individuals with fibromyalgia. Pain. 2019;160:2589-2602. PMID 31219947
  20. Minerbi A, Fitzcharles MA. Gut microbiome: pertinence in fibromyalgia. Clin Exp Rheumatol. 2020;38 Suppl 123:99-104. PMID 32116215
  21. Minerbi A, Brereton NJB, Anjarkouchian A, et al. Dietary intake is unlikely to explain symptom severity and syndrome-specific microbiome alterations in a cohort of women with fibromyalgia. Int J Environ Res Public Health. 2022;19:3254. PMID 35328942
  22. Minerbi A, Gonzalez E, Brereton N, Fitzcharles MA, Chevalier S, Shir Y. Altered serum bile acid profile in fibromyalgia is associated with specific gut microbiome changes and symptom severity. Pain. 2023;164:e66-e76. PMID 35587528
  23. Minerbi A, Khoutorsky A, Shir Y. Decoding the connection: unraveling the role of gut microbiome in fibromyalgia. Pain Rep. 2024;10:e1224. PMID 39726854
  24. Moscati A, et al. Life is pain: fibromyalgia as a nexus of multiple liability distributions. Am J Med Genet B Neuropsychiatr Genet. 2023;192:171-182. PMID 37334860
  25. Negrini F, de Sire A, Lazzarini SG, et al. Reliability of activity monitors for physical activity assessment in patients with musculoskeletal disorders: a systematic review. J Back Musculoskelet Rehabil. 2021;34:915-923. PMID 33935067
  26. Pan Q, et al. Genome-wide association study identifies novel genetic variants associated with widespread pain in the UK Biobank (n = 172,230). Mol Pain. 2025;21:17448069251346603. PMID 40509746
  27. Petrucci-Nelson T, et al. Complex genetics and regulatory drivers of hypermobile Ehlers-Danlos syndrome: insights from genome-wide association study meta-analysis. medRxiv preprint. 2025. PMID 41001447
  28. Rahman MS, et al. Genome-wide association study identifies RNF123 locus as associated with chronic widespread musculoskeletal pain. Ann Rheum Dis. 2021;80:1227-1235. PMID 33926923
  29. Robl R, Eudy A, Bachali PS, et al. Molecular endotypes of type 1 and type 2 SLE. Lupus Sci Med. 2023;10:e000861. PMID 36720488
  30. Thanh Nhu N, Chen DY, Kang JH. Identification of resting-state network functional connectivity and brain structural signatures in fibromyalgia using a machine learning approach. Biomedicines. 2022;10:3002. PMID 36551758
  31. Venerito V, Lopalco G, Morrone M, et al. fAI-BRO: a multimodal AI decision-support system to address diagnostic delay in fibromyalgia syndrome. RMD Open. 2026;12:e006995. PMID 42442924
  32. Zetterman T, Nieminen AI, Markkula R, Kalso E, Lötsch J. Machine learning identifies fatigue as a key symptom of fibromyalgia reflected in tyrosine, purine, pyrimidine, and glutaminergic metabolism. Clin Transl Sci. 2024;17:e13740. PMID 38411371