R26

log/R26.md — Round 26 entries

Sub-file of log — see parent for index.

[2026-05-06] verify — molecules/compounds/coq10.md (pass 2)

Pages verified: 1 (second-pass; six new sources added by intermediate edit)

Sources checked:

• Dohlmann 2022 (10.3390/antiox11091698) — PMC full text (PMC9495827) — VERIFIED
• Kovacic 2025 (10.1017/jns.2025.10043) — PMC full text (PMC12554813) — VERIFIED
• Ahmad 2024 (10.7759/cureus.68316) — PMC full text (PMC11441719) — VERIFIED
• Alehagen 2025 (10.3390/antiox14060685) — PMC full text (PMC12190002) — VERIFIED
• Wei 2022 (10.1007/s11845-021-02651-x) — PubMed abstract only (closed-access) — ABSTRACT-VERIFIED
• Raizner 2021 (10.1016/j.jacc.2020.12.009) — PubMed abstract only (closed-access) — ABSTRACT-VERIFIED

Corrections made (pass 2):

1. Ahmad 2024 screened count “5,000 records screened” → 5,461 records identified / 2,665 screened (per PRISMA flowchart; abstract rounds to “5,000 identified”)
2. Alehagen 2025 mean age ~78 → ~77 (PDF Table 1: 77.0 / 77.3)
3. Kovacic 2025 article number e34 → e72 (per PDF title page)
4. Kovacic 2025 p-value p<0.05 → p=0.043 (exact value from PDF Table 3); heterogeneity I²=93.3% added to footnote; subgroup finding added (benefit only in low-baseline-pain subgroup)
5. Dohlmann 2022 correlation p-value p=0.948 → p=0.946 (per PDF Figure 2b caption); footnote expanded with CS activity values and analysed n per arm
6. Alehagen 2025 IHD claim “equal benefit in IHD subjects” → nuanced: female p=0.02, male p=0.06 (nominally non-significant); Cox HRs 0.29 vs 0.61 added to body and footnote
7. Raizner 2021 footnote expanded with publication date, exact quoted conclusions, diastolic-HF framing confirmed

Wei 2022 and Raizner 2021: All claims in the wiki match the abstract/PubMed data exactly. No corrections needed.

Framing note: Kovacic 2025’s I²=93.3% is a critical qualifier for the “marginally positive” characterization — the CI barely excludes zero AND heterogeneity is extreme. The page’s existing framing (“CI that hugs zero”) remains appropriate; the high heterogeneity strengthens the case for skepticism. No synthesis revision triggered, but the I²=93.3% is now in the footnote for the next reader to judge.

Downstream pages: No new downstream propagation required from pass 2 corrections (all corrections are numeric refinements within the coq10.md page itself).

[2026-05-06] R26 — Round 26 housekeeping (started)

Round 26 scope: housekeeping-only (no new pages). User direction (2026-05-06): “clean up all the housekeeping items before seeding any new pages.” 5 sub-rounds:

• R26a — CLAUDE.md schema cleanup (DONE, this entry)
• R26b — Verify R24 atomic pages (in progress)
• R26c — Verify R23 atomic pages (queued)
• R26d — Open Targets druggability-tier sweep + UniProt-REST canonical-DB drift check (DONE, this entry)
• R26e — Lint-pass carryover items (queued)

R26a — CLAUDE.md schema cleanup (DONE)

Applied 5 schema escalations from R25 close + 2 R24 carryovers:

1. type: pathway block — added verified / verified-date / verified-by / verified-scope fields (mtor.md + heat-shock-response.md precedent formalized). Pathway pages originate quantitative claims (key-node activities, EC50s, KO phenotypes) and carry the same verification discipline as type:protein and type:process.
2. type: pathway druggability-tier ambiguity (R25 close) — added prose convention: pathway pages use aging-context tier (not max-druggability of any node). Examples: mTOR=tier 1 (rapamycin = aging-tractable); HSR=tier 2 (HSP90 inhibitors = oncology-only).
3. mr-causal-evidence enum extension (R24 carryover) — added not-applicable-somatic value for proteins whose aging relevance is mediated by somatic mutation (CHIP drivers — DNMT3A, ASXL1, TET2, JAK2 V617F, SF3B1).
4. mr-causal-evidence partial vs not-tested threshold (R25 close) — defined: partial = GWAS hits exist but no MR study published (instrument-available, study-pending); not-tested = no instruments described or MR attempted-but-underpowered.
5. type: cell-type Cell Ontology gap pattern (R25 close, FAPs precedent) — documented null + #gap/needs-canonical-id pattern when no CL term resolves at appropriate granularity. Lint pass should aggregate into standing watchlist for periodic CL update sweeps.
6. type: hypothesis Mode-B schema cleanup (R24 carryover) — key-evidence-for / key-evidence-against are REQUIRED for Mode A (evidence-aggregating) but OPTIONAL for Mode B (conceptual-frame, e.g., antagonistic-pleiotropy). Lint pass should suppress missing-field warnings for Mode B pages.
7. cardiolipin-stabilizer intervention class — promoted from frameworks/intervention-classes.md roadmap-classes table to main inventory (with elamipretide as canonical first compound). Dataview block added; roadmap row removed.

R26d — Open Targets sweep + UniProt-REST canonical-DB drift check (DONE)

Re-queried Open Targets Platform for the 6 mito-dynamics cluster pages flagged with null druggability-tier due to R25 5xx errors. Result: 5 of 6 already filled by R25 verifier sweep (fis1=3, mid49=3, mid51=3, oma1=3, yme1l=4); only mff was still null.

MFF druggability tier resolved (queried 2026-05-06, ENSG00000168958): tier 3 (predicted druggable). No approved drug, no advanced clinical, no structure-with-ligand, no high-quality ligand or pocket, no druggable-family classification across SM/AB/OC modalities. Positive signals confined to PROTAC modality — Database Ubiquitination = true and Half-life Data = true. Body annotation updated to document the PROTAC-tractable interpretation.

Canonical-DB drift cross-check (UniProt REST against 11 R24 verified:false protein pages): 3 drifts caught:

• MFF — Ensembl ENSG00000168772 (CXXC4) → ENSG00000168958 (MFF). The wrong ID was actually present on the verified mff.md page (stale R25 seeder); R25 verifier-sweep did not catch this. Fixed in mff.md frontmatter, verified-scope, body Identity table, and body druggability paragraph.
• caspase-1 — Ensembl ENSG00000141510 (TP53) → ENSG00000137752 (CASP1). R24d seeder drift; never validated against UniProt REST. Fixed in frontmatter and body Identity bullet.
• TET3 — HGNC 11591 (STK24) → 28313 (TET3). R24b seeder drift. Fixed in frontmatter and body Identity bullet.

Cluster-wide pattern (R25 + R26 combined): 8 canonical-DB drifts caught across mito-dynamics + R24 protein pages over two consecutive rounds. The R25 close recommendation — “lint-pass UniProt REST cross-check against all verified: false protein pages” — is now upgraded to a permanent lint-pass step against ALL protein pages regardless of verified status. Drift pattern is not isolated to verified:false pages (mff was verified:true with wrong ID, only caught when re-querying Open Targets).

Schema escalation surfaced:

• The R25 close noted that cell-ontology-id and mr-causal-evidence fields are commonly null. These under-population audit sweeps remain queued.
• Open Targets tier null on the rest of the R24 set will be filled by R26b verifier sweep (in progress).

R26e — Lint-pass carryover items (DONE)

Four standing carryover items addressed:

1. HSC verified-scope stale on Jaiswal 2014 CIs (R12 carryover) — FIXED. Re-verified against NEJM PDF (Jaiswal 2014, doi:10.1056/NEJMoa1408617, abstract + body Results §). Two CI corrections caught: incident-CHD upper-CI 3.5 → 3.4; ischemic-stroke lower-CI 1.3 → 1.4. Also added prevalence CIs (5.0–6.3 / 8.4–10.8 / 8.6–15.7 / 12.1–27.0) and case-counts (219/2,300; 37/317; 19/103) that were absent from the page. Top-line “90+” age bracket clarified to “90–108” per paper. verified-scope updated to record the re-verification date and what was corrected.

2. mitofusins.md drafting gap: Karbowski 2002 MFN2-at-apoptotic-fission-sites (R25 carryover) — FIXED. Added new section ## MFN2 at apoptotic fission sites between CMT2A and Aging-relevance. Quantitatives (92% ± 10.5% MFN2/BAX co-localization, n=847; 68% ± 10% reciprocal, n=1,238) transcribed verbatim from the verified mitochondrial-dynamics page (Karbowski 2002 PDF was verified there 2026-05-05). New [^karbowski2002] footnote added. verified-scope updated to disclose the verbatim-transcription provenance (rather than independent re-verification on this page).

3. mode: lifestyle vs dietary audit (R23c outputs) — FIXED. Audited 8 lifestyle-folder intervention pages. Only one drift: caloric-restriction was mode: lifestyle (R10 seeding pre-dated schema clarification); flipped to mode: dietary. R23c outputs (intermittent-fasting, time-restricted-eating, methionine-restriction, ketogenic-diet) were already correct. Updated frameworks/interventions-by-modality.md Watchdog 1 commentary + Limitations §: post-audit modality counts are lifestyle (3 — exercise, sleep, heat-exposure), dietary (5 — caloric-restriction + R23c outputs). Documented that the lifestyle/ folder is the catchall directory for non-pharmacological non-cellular interventions (folder name does not constrain mode: value).

4. Lint script alias-blindness (R10e standing item) — DOCUMENTED. Added a ⚠️ Known limitation callout to sops/lint-pass.md § “Broken wikilinks”, explicitly listing the five known-aliased false-positives (inflammaging, irs1, bnip3l, bcl2-family-pathway, mfn1/mfn2) and proposing the alias→file map fix as a versioned-script feature request. Tracked here in log + ROADMAP standing items rather than as a separate FEATURE_REQUESTS file (a local paper archive’s FEATURE_REQUESTS.md is for the archive tool, not the wiki lint scripts).

R26e status: all four carryover items closed. No new gaps surfaced.

[2026-05-06] R26 — Round 26 close

Round 26 status: COMPLETE. All 5 sub-rounds finished:

• R26a — CLAUDE.md schema cleanup (7 escalations applied: pathway verified+druggability-tier-aging-context conventions, mr-causal-evidence not-applicable-somatic + partial/not-tested threshold, Cell Ontology gap pattern, hypothesis Mode-B optional-fields, cardiolipin-stabilizer class promoted).
• R26b — R24 atomic-page verification sweep (17/17 pages verified, ~80+ corrections, ~15 scientifically critical).
• R26c — R23 atomic-page verification sweep (22/22 pages verified, ~110 corrections, ~25 scientifically critical, 4 fabricated experimental claims removed, 4 reversed-direction errors caught).
• R26d — Open Targets druggability-tier sweep + UniProt-REST canonical-DB drift check (3 drifts caught: MFF Ensembl, caspase-1 Ensembl, TET3 HGNC).
• R26e — Lint-pass carryover items (HSC verified-scope CIs re-verified body-vs-abstract; mitofusins Karbowski 2002 added; mode:lifestyle/dietary audit; lint script alias-blindness documented).

Wiki-wide state after Round 26 close: 0 atomic pages remain verified: false for the first time in the wiki’s history. All ~330+ atomic content pages now ship verified: true with documented verified-scope lines distinguishing PDF-verified end-to-end claims from partial/abstract-only/closed-access scoped verification.

User direction (2026-05-06): “clean up all the housekeeping items before seeding any new pages.” This was the explicit Round 26 mission and is now satisfied. R27+ can resume new-page seeding.

---