What we know. What we're testing. What we don't know yet.
Endometriosis affects ~190 million people — yet it took decades to be taken seriously, and it still gets a sliver of research funding. OpenEndo is an open project that tracks the evidence, finds new angles, and tests ideas on computers before anyone touches a lab. This page is an honest tour: what is established, what we are testing right now, and what remains unknown. Nothing here is medical advice — everything is sourced and open.
~190M people affected~7 yr average diagnosis wait0.05% of research funding100% open · MIT
Why this exists
A disease that shouldn't still be this invisible
Endometriosis is chronic, painful and common — yet diagnosis takes years and research is starved. The gap isn't a lack of patients. It's a lack of organised, open evidence that researchers, clinicians and patients can all build on.
190M
people live with endometriosis worldwide
WHO estimate — more than diabetes + epilepsy combined
~7yr
average time from symptoms to diagnosis
a disease waved away — the name says it
0.05%
of health research funding goes to it
per burden, it should be tens of times more
What OpenEndo does
Four lanes, one goal: better evidence, faster
live
📡 Watch the evidence
Every endometriosis trial, paper and grant — tracked weekly from public registries (ClinicalTrials.gov, PubMed), in the open.
merged
🎯 Find new targets
Mapped the drug-target landscape — 58 targets, of which 35 are truly novel (no drug has ever been developed against them).
ranked
💊 Repurpose drugs
Screened 9 approved drugs as candidates for endometriosis — ranked by mechanism, evidence and safety (M3).
testing
🧪 Test in silico
New: a virtual-testing pipeline — computer tests on structures and binding before any wet lab (see below).
The big picture
From data to patient — where we are
Every treatment that exists today travelled this road. We are doing steps 1–3 openly, on computers, for ideas nobody else is pursuing. Step 4 needs a lab partner; step 5 needs clinical research.
📡
Open data
trials · papers · funding, weekly, in the open
→
🎯
Targets & candidates
35 novel targets · 9 repurposing leads
→
🖥
Virtual tests
structures · docking · dynamics
→
🔬
Lab tests
cells & models — needs a lab partner
→
🏥
Clinical research
trials — the only real proof
Established
What we know (sourced & verified)
Know Verified facts, with dates and sources
✓The disease is under-researched, not rare. ~190M affected (WHO); diagnosis delay ~7 years; ~0.05% of research funding. Sources: WHO; site stats — verified 2026-08
✓Denmark's access landscape is mapped. Ryeqo has general reimbursement (≈849 kr./28 days); Yselty is approved but not reimbursed for endometriosis (out-of-pocket or individual application); two specialised centres treat severe disease (Rigshospitalet — East, AUH Skejby — West). Lægemiddelstyrelsen / sundhed.dk — verified 2026-09-02
✓58 drug targets mapped — 35 truly novel. Cross-checked against ChEMBL (the drug industry's own database): 35 have no drug ever developed against them. targets.json — 2026-09-02
✓The screening method works. Of 9 repurposing candidates, the computer screen flagged two as "wrong direction" — estradiol (it activates a growth receptor) and dinoprostone (it is the inflammatory signal PGE2). The method caught its own errors. m3-validation.md — 2026-09-02
✓Sirolimus (rapamycin) is the strongest repurposing lead. Its mechanism (mTOR) has independent endometriosis evidence; it has 20+ years of safety data from transplant medicine. M3 validation — 2026-09-02, confidence medium-high
✓MRGPRX2 is a validated pain mechanism. New 2026 studies link this mast-cell receptor to endometriosis pain — a specific, non-hormonal target. M3 validation
✓We can test these ideas right now. Experimental 3D structures exist for every current test candidate (FKBP4, FKBP12, MRGPRX2, xCT); 27 of 35 novel targets already have predicted structures in AlphaFold DB. Phase 0 audit — 2026-09-02
27 of 35 targets — structure already exists
8 to fold
already in AlphaFold DB (free, predicted)need one folding run (~$1–3 total)
Experimental (cryo-EM / X-ray) structures also exist for the drug candidates being tested: FKBP4 · FKBP12 (control) · MRGPRX2 · xCT · ACVR1B.
In progress
What we're testing right now (hypotheses, on computers)
Testing Each test produces a hypothesis — never a conclusion
1Does sirolimus actually fit its target? Molecular docking: sirolimus → FKBP4, with FKBP12 as a control — because we need to know if it can engage the endometriosis-relevant protein at all. Phase 1 · Vina docking · ~$5
2Can cetrorelix reach the pain receptor? Docking the peptide drug against the MRGPRX2 cryo-EM structure — testing the pain-pathway idea computationally. Phase 1 · ~$3
3Does sulfasalazine's target move in the right direction? Its target (xCT) is real — but a 2026 paper suggests ferroptosis may drive lesion growth, not shrink it. Direction must be resolved before this ranks higher. literature + docking · Phase 1
4Are the targets actually switched on in disease tissue? Next free step: endometriosis-lesion gene data (GEO/single-cell) — is MRGPRX2 elevated in lesion mast cells specifically? Phase 0.5 · ~$0
5Fold the 8 targets nobody has a structure for. AlphaFold on a rented GPU — a few dollars, a few hours. Phase 1 · ~$1–3
The evidence ladder — where each candidate stands
For a repurposed drug, the honest path has five rungs. We are climbing rungs 2–3 on computers; rungs 4–5 need laboratories and clinical research.
1
Approved & safe for another disease
All 9 candidates passed this — that's the point of repurposing. Sirolimus has 20+ years of safety data.
2
Mechanism fits endometriosis biology
Done for 9/9 (M3 validation): sirolimus strongest · MRGPRX2 axis validated · two candidates failed this rung honestly.
3
Physically binds the target
In progress — docking on experimental structures (Phase 1). Controls included so we don't fool ourselves.
4
Works in endometriosis models
Needs a wet lab (cells / animal models). We are looking for a lab partner.
5
Helps patients in trials
Years away — and only clinical research can ever prove it. Our work exists to make that more likely, cheaper and better-targeted.
The 9 candidates, after ranking
Sirolimus (rapamycin)
mTOR pathway — independent endometriosis evidence, huge safety history
Top tier
Cetrorelix
already used in endo — validates MRGPRX2 pain target, not a new drug
real mechanisms but open questions: toxicity, immunosuppression, or drug-target mismatch
Watchlist
Estradiol · Dinoprostone (PGE2)
wrong direction — would feed the disease, not fight it. Caught by the screen.
Wrong direction
Honest gaps
What we don't know yet
Unknown We'd rather say "we don't know" than pretend
?Does sirolimus actually help endometriosis? No endometriosis trial exists. Docking can say "it fits"; only a lab and trials can say "it helps". the honest ceiling of computer work
?Which direction does ferroptosis go in endometriosis? A Sept 2026 paper links it to lesion progression — if true, drugs that trigger it (sulfasalazine) could backfire. Unresolved. must-read before ranking
?Is MRGPRX2 really elevated in lesion mast cells? It's absent from healthy bulk tissue (expected for a mast-cell receptor) — the disease-specific question needs lesion data. Phase 0.5
?Does a good computer binding predict a good real binding? Usually yes-ish, sometimes no. That's why docking is followed by dynamics, and why both still need a lab. methodological humility
?Would a selective ACVR1B inhibitor work? The biology says yes — but no such drug exists. Building one is a decade-scale project. parked honestly
?Funding. All of this runs on evenings and small GPU rentals. A lab partnership or grant would change the pace completely.
Plain words
What the terms mean
What is a "drug target"?
A specific protein that a disease depends on. If a drug can switch that protein off (or on), the disease loses an advantage. Our 35 novel targets are proteins involved in endometriosis that no drug has ever been aimed at — overlooked on purpose, because the disease was under-funded.
What is "repurposing"?
Taking a drug already approved for one disease and testing it against another. Huge shortcut: safety is already known, so years of early testing are skipped. It's how we found sirolimus.
What is AlphaFold / a "predicted structure"?
A 3D shape of a protein, computed from its genetic sequence. AlphaFold (the 2024 Nobel Prize in Chemistry) made this possible for almost any protein. Predicted shapes are models — useful for planning, never proof.
What is "docking"?
A computer test that tries to fit a drug molecule into its target protein, like testing if a key turns in a lock. It scores how well they fit. Fast and cheap — a first filter, not a verdict.
What is "molecular dynamics"?
A simulation of the drug-protein pair wiggling in water over time (nanoseconds). If the drug falls off instantly, the docking was probably an illusion. More expensive than docking, so we only run it on the best hits.
What is ChEMBL?
The drug industry's open database of drug-target relationships. When we say a target is "novel", we mean ChEMBL has no drug against it — verified, not assumed.
What does "confidence" mean on our pages?
Our honesty label. High = multiple sources, verified. Medium = single source or fast-moving. Low = hypothesis, verify before relying. Every claim we publish carries one.
Join in
How you can help
Share
Send this page to one person who should know the disease is under-researched.
Labs
Are you in endometriosis or drug research? We need a wet-lab partner for rung 4.
Data
Spot a missing trial, paper or funding call? Add it — the dataset is open (GitHub).
Fund
GPU hours cost dollars, not millions. Support open endometriosis research.
Read this first
OpenEndo is open research infrastructure — not a medical service, and nothing here is medical advice. Every claim carries a source and a confidence label; computational results are hypotheses until confirmed in the lab. If you have endometriosis symptoms, talk to a clinician — and in Denmark, Endometriose Fællesskabet (endo.dk) offers free counselling.