An ontology-driven knowledge graph of fibromyalgia research — built link by link from peer-reviewed evidence, with explicit confidence tiers and active bridges to long COVID, ME/CFS, and other adjacent conditions. The goal: identify foundational scientific paths toward a cure.
Pregabalin, duloxetine, milnacipran. Aerobic exercise, sleep hygiene, cognitive behavioral therapy. Today's standard of care reduces symptoms for some patients, but it doesn't fix what's broken — because we don't know exactly what's broken.
What we do know: fibromyalgia is not "in your head," and it's not a single disease. It's a final common pathway that several distinct upstream mechanisms can feed into. Some patients have small-fiber neuropathy. Some have inflammatory glial activation visible on TSPO-PET. A substantial fraction — perhaps 37% — have circulating IgG antibodies that, when transferred to mice, reproduce fibromyalgia-like pain.
Treating all of them with the same antidepressants and gabapentinoids averages across non-responders. The path to a cure runs through subtype-stratified mechanism research — figuring out which patient has which upstream cause, and intervening at that point in the chain.
This project is structured as a living knowledge graph rather than a narrative review. Each entity — every mechanism, molecule, cell type, intervention, biomarker — exists as a typed node. Each relationship between entities carries an explicit confidence tier and citations to supporting papers.
Findings that have been independently replicated across well-powered peer-reviewed studies. The floor of the field's consensus.
establishedPublished findings with strong methodology but not yet independently replicated. Most active research operates at this tier.
emergingClaims derived by triangulating across multiple sources — not directly stated in any single paper. Held with deliberate caution.
inferredHypotheses imported from long COVID, ME/CFS, IBS, lipedema, and other adjacent fields. Speculative but mechanistically motivated.
bridgingNothing gets silently promoted between tiers. Promotion is a deliberate, logged act — recorded in the open-questions log with the triggering evidence. This is the project's epistemic backbone.
For two decades, researchers have pursued the autoimmune hypothesis (FM is driven by circulating IgG antibodies) and the mast-cell hypothesis (FM is driven by histamine-releasing mast-cell activation) as parallel paths.
A May 2025 preprint from Sanchez, Goebel, and Dong showed that FM patient IgG triggers mast cell activation through a specific receptor — MRGPRX2. Genetic deletion of that receptor, or removal of the mast cells themselves, abolishes the IgG-induced pain phenotype in mice.
The two parallel hypotheses are actually one chain. Substance P (a long-established FM biomarker) feeds the same receptor from another direction. Hereditary α-tryptasemia, a genetic trait affecting ~5% of people, amplifies the whole mechanism.
Three further candidate chains staged in May 2026 — H4 (adipose → leptin → microglia, in metabolic-syndrome-comorbid patients), H5 (celiac+T1DM autoimmune-genetic signature within FM-IgG+ subset), and H5b (SOCE-amplifier genetic profile docking into the existing calcium-hypothesis CA-I1 intersection) — supply patient-stratification handles for the H1×H2 unified mechanism rather than competing with it.
All four anchored on clinical-stage compounds. Several have no IND barrier for investigator-initiated FM trials.
Two interactive visualizations. The first traces the mechanism chain from triggers to symptoms. The second shows the structure of the entire ontology — every entity, every relationship, every bridge across domains.
The project's leading chain. Vertical flow from triggers through FM-IgG to the mast cell + MRGPRX2 convergence node — Sanchez/Goebel/Dong 2025 finding collapses the H1 autoimmune and H2 mast-cell hypotheses into a single chain. Six numbered break-points mark where the chain can be interrupted clinically.
Candidate chain 2026-05. Adipose tissue dysfunction → leptin → microglial activation via JAK2/STAT3 + MAPK/ERK + PI3K/Akt. Predicts FM with metabolic-syndrome comorbidity. Six cure-path-arm break-points: GLP-1, LepRb antagonism, JAK inhibitors, minocycline, anti-IL-6, metformin.
Candidate chain 2026-05. The celiac+T1DM autoimmune-genetic signature applied to FM stratification: HLA-DQ2/DQ8 + CTLA4 + PTPN22 (tolerance triad) + IL2/IL21 + IL2RA + TNF + MYO9B + CLDN2 + TGM2. Merges with H1 at the FM-IgG production node. Six break-points across screening, dietary, biologic, and biomarker arms.
Candidate chain 2026-05. Sibling to H5. Six-gene module deliberately straddling three project domains — ORAI1 + STIM1 (immune AND calcium via SOCE — docks directly into CA-I1 intersection of the calcium hypothesis) + MYO9B + OCLN + TJP1 (three layers of the tight junction) + CTLA4. The discriminating cure-path test: CRAC channel inhibitor (CM4620/Auxora) head-to-head vs MRGPRX2 antagonist in stratified patients.
Force-directed network of the full ontology: 187 entities, 201 typed relationships, 43 cross-domain bridges. Filter by class, tier, or relationship type. Click any node to see definition, aliases, and connected entities.
The chain inventory hub. H1, H2, H3, H4, H5, H5b — each with confidence tier, key mechanism, subset, cure-path-arm tractability, and direct link to the interactive mechanism map. The project's recursive-narrowing decision space.
Each is the kind of question that, if answered, would meaningfully change the mechanism map or the subtyping hypotheses. Each carries a status, a tractability estimate, and the experiment that would close it.
The project's highest-leverage cure-path experiment. Crossover RCT of colesevelam in anti-SGC-IgG-positive FM patients, n=50, ~24 months. Full protocol drafted.
Open questionThe most direct cure-path test enabled by Sanchez 2025. Clinical-stage compounds already in Phase 2/3 for chronic urticaria. No IND barrier.
Open questionMignolet 2026 showed LC-IgG passive transfer reproduces sensory pain in mice. If the receptor mechanism extends, the same therapy hierarchy applies to post-COVID FM.
Open questionHighlights from the May 2026 ingestion batches — 20 papers added across the H4 immuno-metabolic, H5 autoimmune-genetic, and H5b immune-barrier-calcium tracks. Each anchor advances a bridge from staging into cross-condition or mechanism-loading territory.
Patient-cohort study, n=45 MS + 45 HC
First patient-cohort SOCE-pathway dysregulation in an autoimmune CNS disease. Stage-progressive serum SARAF (3× control); reduced STIM1/ORAI1 expression; elevated intracellular calcium. Cross-condition anchor for H5b at the CA-I1 intersection.
PaperZhu et al. (Eur J Med Res)
Instantiates the H5b three-domain framing in asthma: STIM1/Orai1 SOCE + Piezo1 → calcium-protease → OCLN degradation → Th2 release. Adds Piezo1 as candidate H5b extension gene.
PaperGeorgopoulos et al. (Sci Rep)
ME/CFS susceptibility HLA alleles bind herpesvirus antigens significantly weaker than protective alleles (P<0.001 across >10,000 antigens). Cross-condition mechanism anchor partially closing B-Gen-1 for the H5 framework.
PaperOmar et al. (Acta Neuropsychiatrica)
sCTLA-4 explains 55-58% of variance in Fibro-Fatigue + HAMD + HAMA scores in preeclampsia (n=90+60). First patient-cohort biomarker linking immune-checkpoint activation to chronic-fatigue phenotype.
PaperSteen et al. (BMC Medicine), n=166,774
Population-scale evidence that FM coaggregates in families with autoimmune disease (λR 1.06-1.24) AND cardiometabolic disease (λR 1.00-1.23). Empirical anchor for both H5 and H4.
PaperProject H4 anchor paper
Proposes leptin → JAK2/STAT3 + MAPK/ERK + PI3K/Akt signaling → reactive microglia as candidate H4 immuno-metabolic chain. Selective leptin resistance framework. New bridge B-IM-1.
Paper