Coverage and gaps
How the map's landmarks fall across lanes, kinds and phases, and how complete the project thinks it is.
Landmarks by lane and outcome
| Foundations | 4 | 0 | 0 | 0 | 0 | 4 |
| Synuclein misfolding & spread | 11 | 13 | 4 | 0 | 3 | 31 |
| Synuclein lowering | 5 | 6 | 2 | 0 | 2 | 15 |
| Gut–brain origin | 2 | 9 | 0 | 1 | 2 | 14 |
| Mitochondria & energy | 7 | 14 | 1 | 1 | 4 | 27 |
| Oxidative stress | 0 | 3 | 2 | 2 | 6 | 13 |
| Lysosome & GBA | 3 | 11 | 1 | 0 | 3 | 18 |
| LRRK2 | 10 | 4 | 2 | 1 | 2 | 19 |
| Selective vulnerability | 2 | 9 | 0 | 0 | 1 | 12 |
| Cell survival & anti-apoptosis | 1 | 6 | 2 | 1 | 4 | 14 |
| Inflammation & immunity | 0 | 16 | 4 | 0 | 3 | 23 |
| Metabolic drug repurposing | 2 | 4 | 1 | 1 | 5 | 13 |
| Environment & lifestyle | 2 | 8 | 0 | 0 | 1 | 11 |
| Genetic architecture | 7 | 14 | 0 | 0 | 4 | 25 |
| Cell replacement | 9 | 6 | 6 | 1 | 4 | 26 |
| Growth factors | 2 | 4 | 1 | 1 | 3 | 11 |
| Gene therapy & delivery | 7 | 4 | 0 | 0 | 3 | 14 |
| Biomarkers | 9 | 7 | 0 | 1 | 0 | 17 |
| Trial methods & regulation | 7 | 6 | 2 | 3 | 2 | 20 |
| Infrastructure & funding | 8 | 3 | 0 | 0 | 0 | 11 |
| Models | 6 | 2 | 0 | 0 | 0 | 8 |
By kind
| Kind | Landmarks |
|---|---|
| Discovery | 146 |
| Trial in people | 129 |
| Research tool | 29 |
| Research infrastructure | 19 |
| Research model | 13 |
| Lab study (cells or animals) | 10 |
Trials by phase
| Phase | Trials |
|---|---|
| Phase 1 | 18 |
| Phase 1/2 | 2 |
| Phase 2 | 73 |
| Phase 3 | 18 |
| Phase 4 | 1 |
| not recorded | 17 |
Least verified landmarks
Almost every file lists some "Not yet verified" points, so the count is a guide to where a second check would help most, not a mark of quality. The ten with the most open points:
| Landmark | Kind | Outcome | Open points |
|---|---|---|---|
| BIIB122 / DNL151 (LUMA) | Trial in people | Did not hold up | 7 |
| GT-02287 Phase 1b (GCase modulator) | Trial in people | Still being studied | 7 |
| VY-AADC Phase 1b and RESTORE-1 | Trial in people | Did not hold up | 7 |
| ASAP and GP2 | Research infrastructure | Still being studied | 6 |
| VTX3232 (NLRP3 inhibitor) | Trial in people | Still being studied | 6 |
| Parkinson Study Group founded | Research infrastructure | Confirmed and built on | 6 |
| Idebenone in REM sleep behaviour disorder | Trial in people | Result due | 6 |
| RNDP-001 (Kenai) REPLACE | Trial in people | Result due | 6 |
| Freed: sham-controlled fetal grafts | Trial in people | Did not hold up | 6 |
| DNAJC13 and TMEM230: disputed genes | Discovery | Did not hold up | 6 |
Coverage of disease-modifying trials
Checked 19 Oct 2026. Measures how complete the trial layer is against outside lists (goal item 1 in CLAUDE.md). The source tables, built by research agents, are in [coverage/](coverage/). Back to the index.
Sources
- 2020–2024: [
coverage/pipeline-2020-2026.md](coverage/pipeline-2020-2026.md). This covers the five annual pipeline reviews by McFarthing, Stott and colleagues (J Parkinson's Dis 2020–24), plus their 2015–2024 retrospective (May 2026), with status checked on ClinicalTrials.gov on 18 Oct 2026. No 2025 or 2026 edition exists, so trials registered from 2025 onward must be found separately. - Before 2020: [
coverage/pre2020-trials.md](coverage/pre2020-trials.md). Built mainly from Athauda & Foltynie (Nat Rev Neurol 2015), Tables 1–2. Lang & Espay (2018) could not be read, so this list is less complete.
Tally
| Source | Agents or trials | Covered | Mentioned inside a landmark | Missing |
|---|---|---|---|---|
| Pipeline reviews, 2020–24 (Phase 2+) | 92 agents | 46 | 8 | 38 |
| Before 2020 (Phase 2+) | 35 trials | 27 | 6 | 2 |
Updated 18 Oct 2026 after batches 17–20. All group 1 trials now have landmarks (161–190), and all group 2 mentions are added. Everything still missing is in group 3. Batch 25 (227–286) moved BIA 28-6156 from mentioned to covered (235) and added AAV2-GAD (252); it also added trials outside these lists (DNL151 Phase 1, GT-02287, NR-SAFE, the terazosin pilot, PDGF-BB, ENT-01, EFFACE-PD, Soroka, idebenone in RBD). Batch 26 (287–346) promoted ten group 3 agents with a result or a clear end (AKST4290, GRF6021, EPI-589, fasudil, lithium, N-acetylcysteine, hydrogen, and the APM001 arms alogliptin, nilvadipine and albuterol) and gave PT320 and the Stockholm exenatide trial their own files; trials started in 2025–26 are tracked in [coverage/sweep-2025-2026.md](coverage/sweep-2025-2026.md). The first check (before batch 17) found 23/5/64 and 14/4/17.
Group 1: new landmarks (done: landmarks 161–190)
Trials that add a dead end, change a branch's status, or shaped how slowing trials are designed.
| Branch | Year | Trial | Why |
|---|---|---|---|
| trial-methods | 2004 | ELLDOPA (levodopa, 361) | Washout design; scans and clinical results disagreed |
| trial-methods | 2002–09 | TEMPO (rasagiline, 404) | First delayed-start design, before ADAGIO (62) |
| trial-methods | 2002–03 | CALM-PD and REAL-PET (dopamine-agonist imaging trials) | Started the scan-versus-clinic debate |
| trial-methods | 2006–07 | NET-PD futility screens FS-1/FS-2 (minocycline, creatine, GPI-1485, CoQ10) | The screening design behind 70 and 105 |
| trial-methods | 2013 | PROUD (pramipexole, 535) | Delayed-start, negative |
| trial-methods | 2019 | LEAP (levodopa, 445) | Delayed-start, negative; settled the question of whether levodopa is toxic |
| mitochondria | 2007 | PRECEPT (CEP-1347, 806) | Stopped for futility; ended the mixed-lineage-kinase line |
| mitochondria | 2006 | TCH346 (301) | Anti-apoptotic drug, negative |
| mitochondria | 2010 | MitoQ PROTECT (128) | Mitochondria-targeted antioxidant, negative |
| mitochondria | 2026 | NOPARK (nicotinamide riboside, Phase 3, 410) | Large NAD-boosting trial, reported negative |
| mitochondria | 2020s | CNM-Au8 (REPAIR-PD) | Brain-energy nanocrystal; results posted, not yet read |
| cell-replacement | 2008 | Spheramine STEPS (71) | Retinal-cell grafts failed against sham surgery |
| mitochondria | 2025 | Risvodetinib (IkT-148009, c-Abl, 137) | Second-generation c-Abl drug after nilotinib (68); check efficacy |
| synuclein-spread | 2025 | Exidavnemab (EXIST Phase 2) | Antibody against synuclein aggregates; check result |
| synuclein-spread | 2024–25 | UB-312 vaccine (Phase 1/2) | Active vaccine after PD01A and ACI-7104 (103) |
| inflammation | 2024–25 | VTX3232 (NLRP3 inhibitor, Phase 2) | First NLRP3 test in patients after 55 |
| lrrk2 | pending | NEU-411 (NEULARK, Phase 2, 150) | Second LRRK2 inhibitor; keeps the branch alive |
| growth-factors | 2022 | Fosgonimeton (SHAPE) | HGF-pathway drug; trial terminated, endpoint missed |
Group 2: mentions added to existing landmarks (done 18 Oct 2026)
| Landmark | Add |
|---|---|
| 145 LUMA | LIGHTHOUSE Phase 3 terminated in 2023 with 7 people enrolled |
| 67 deferiprone | SKY Phase 2 (140), also negative |
| 40 ambroxol | AMBITIOUS, GREAT and high-dose trials |
| 20 GLP-1 story | Exenatide Stockholm Phase 2; PT320 (sustained-release exenatide); semaglutide trial (confirm which one) |
| 139 faecal transplant | Warsaw EFFACE-PD and Soroka trials |
| 146 GDNF infusions | Nutt 2003 intraventricular GDNF (50), negative |
| 62 ADAGIO | Florida rasagiline imaging Phase 2 (96) |
| 144 isradipine | STEADY-PD Phase 2 dose-finding (99) |
| 55 NLRP3 / 52 inflammation | AZD3241 (verdiperstat) microglial PET study |
| 68 c-Abl | Radotinib Phase 2 (ongoing) |
Group 3: left out (recheck yearly)
- Symptom or cognition trials labelled disease-modifying by the reviews: blarcamesine, istradefylline, donepezil, memantine, citalopram, ceftriaxone in Parkinson's dementia.
- Trials focused on multiple system atrophy: Lu AF82422 (amlenetug), MEDI1341/TAK-341.
- Small or unpublished Phase 2s with no result found: about 35 agents, including lithium, NAC, sulforaphane, hydrogen, montelukast, gemfibrozil, mannitol, cilostazol, lovastatin, carvedilol, sodium oligomannate, tocotrienol, KM-819, PT320, AKST4290, EPI-589, GRF6021, TJ0113, mesenchymal stem-cell products, and other repurposed drugs. Promote any of these to group 1 if it publishes a clear result or reaches Phase 3.
- Before 2020, still missing: riluzole (outcome unverifiable), Cogane (no peer-reviewed paper). AAV2-GAD was added in batch 25 (252) for its role in the gene-delivery branch.
From the project's graph/coverage.md, shown as written, with its own date.
Gap check
Checked 23 Oct 2026, with the map at 346 landmarks and 526 links (43 links inferred). Scores the map against the coverage goal in CLAUDE.md. It replaces the 19 Oct 2026 version (280 landmarks), whose to-do list batches 26–31 have since cleared. Back to the index.
Summary
| Goal | Target | Now | Status |
|---|---|---|---|
| 1. Disease-modifying trials (Phase 2+, notable Phase 1) | ~150–250 | 129 trial files (101 finished, 28 pending) | Below the range. The coverage check ([coverage.md](coverage.md)) shows what is still left out, and the stop rule in CLAUDE.md applies: add a trial only if it adds a parent link, a dead end or a change in a branch. |
| 2. Genetics: causal genes and GWAS milestones | ~40 | 25 landmarks in the genetics lane, plus gene landmarks in other lanes | The genes and milestones the 19 Oct check listed as missing now have files: HTRA2 (327), DNAJC13 and TMEM230 (328), ATP10B (329), PSAP (330), DCTN1 (331), the progression GWAS (332), polygenic scores (333), SNCA Rep1 (334) and SNCA duplications (335). |
| 3. Branches: 10–25 landmarks each | 10–25 | 19 of 21 lanes are at 11 or more | Below 10: models (8) and foundations (4). Foundations is closed on purpose. |
| 4. Enablers | ~50 | 61 tools, infrastructure and models | Met. |
| Total | 340–360 | 346 | In range. |
What changed since 19 Oct 2026
- The thin branches in the old check are filled: environment 11, synuclein-dosage 15, vulnerability 12, cell-survival 14, gene-delivery 14, inflammation 23.
- The bundled landmarks flagged for splitting were split: separate files now exist for the fetal-graft trials (317, 318, 319), the ambroxol trials (320, 321), the c-Abl drugs (323), the vaccines (324, 325), PT320 (309) and the AAV-AADC trials (305).
- The 2025–26 registry sweep is logged in [sweeps/2026-10.md](sweeps/2026-10.md); 23 trial files start in 2025 or later.
What is still open
- Trials with no single size or phase. About 24 trial landmarks bundle several registered studies, or have no registry record, so they carry no
norphase.validate.pylists them as warnings. Each needs a decision on which study is the headline one. - Models lane (8). Candidates are in next-landmarks.md only if they pass the coverage test.
- Yearly work. Pending-result checks, errata and retraction checks, and the October source sweep are listed in next-landmarks.md under Maintenance.
From the project's graph/gaps.md, shown as written, with its own date.