context: follow-on to the 2026-06-13 cholinergic/HRV batch — user asked to fill the remaining autonomic + neuroendocrine gaps the lint had surfaced (vagus nerve, acetylcholine, catecholamines/β-adrenergic, cortisol/HPA axis). Built from canonical primary literature; the Errico npj-Aging-2025 perspective remains user-excluded and is not cited anywhere.
page-type decisions: vagus nerve → type: tissue (parent-system: nervous-system; first peripheral-nerve page; nerves are anatomical structures like the glands the schema already files in tissues/). Acetylcholine + catecholamines → type: metabolite (endogenous neurotransmitters; mirrored the sole precedent iron.md). “catecholamines/β-adrenergic” folded into ONE catecholamines page (NE/Epi/dopamine + α/β-adrenergic signaling, cross-ref camp-signaling). “cortisol/HPA axis” folded into ONE pathways/hpa-axis.md (the axis is a circuit; cortisol covered in depth there — flagged that a standalone cortisol compound/metabolite page may follow).
added (4 pages, all verified:true after adversarial verifier pass):
tissues/vagus-nerve.md — CN X anatomy (~80/20 afferent/efferent), DMV/nucleus ambiguus/NTS, CAP substrate, vagal-tone/HRV decline, Braak/DMV gut-first PD route. Citations: Borovikova/Wang/Tracey (pointers), Bonaz 2016, De Meersman 2007, Braak 2003, Parkkinen 2008.
verifier corrections (adversarial pass again earned its keep): vagus — Parkkinen 2008 had a WRONG DOI (seeder’s 10.1111/j.1750-3639.2007.00084.x resolves to an unrelated WHO CNS-tumor classification paper) → corrected to 10.1007/s00401-008-0346-6; Braak non-conformance 30–40%→17% (per Parkkinen); [[inflammaging]]→[[chronic-inflammation]]. catecholamines — [^boadle-biber1993] was a WRONG citation (a serotonin-synthesis paper) → replaced with Dunkley 2004 (TH phosphorylation), and TH-activation direction corrected (Ser40 phosphorylation activates, not dephosphorylation); epinephrine HMDB 0000211→0000068; 2 transposed PMIDs (Zhang 2020 31800024→31969699, Puigserver 1998 9733240→9529258); Howlett 2021 vol 151→150. acetylcholine — Gao 2024 effect sizes corrected to actual per-drug SMDs/ORs. HPA — 22 relative-path [[../...]] wikilinks converted to standard form + Kamwa year 2022→2021.
downstream-correction grep (verifier-flagged): Parkkinen old-DOI / 30–40% / Gao-2024-stats checked wiki-wide — no other page carried the wrong versions (the “Gao 2024” hits elsewhere are a different author/DOI).
discipline: every autonomic/neuroendocrine→aging causal pointer hedged as mechanistic/associational (glucocorticoid cascade + sympathetic-overdrive framing flagged contested), per the user’s rejection of the single-driver overreach.
leak-gate: clean — note the standard diff-scoped check skips untracked files, so the new pages were grepped explicitly; scrubbed 3 local-paper-store phrasings the seeder/verifier had copied into footnotes/verified-scope (rewritten to neutral “closed-access / full text not accessed” wording). OA closed-access flags retained.
[2026-06-14] ingest | vitamin D + D3-K2 calcium-partitioning synthesis (Phase 0 for a planned simulator intervention)
context: prep for adding vitamin D3 + K2-MK7 as viz/ simulator interventions. The model SOP is wiki-first; a design/research note (model/vitamin-d3-k2-intervention-design.md, design-only, Codex-reviewed) found the wiki had no vitamin D page at all and no D3+K2 synthesis — blocking. This batch fills that gap. (Model build deferred; this is content only.)
added (2 atomic pages + 1 stub, all seeded→adversarially verified):
molecules/compounds/vitamin-d.md (verified:true) — cholecalciferol/D3 + full endocrine axis (skin→25(OH)D→CYP27B1→calcitriol→VDR), PTH/FGF23 feedback, Ca/bone (osteocalcin VDRE), innate (LL-37) + adaptive (Treg/Th17) immunity, inflammaging (honestly inconsistent), muscle/falls (contested). Core purpose: document the hard-endpoint NULLs — VITAL (Manson 2019, n=25,871; CVD HR 0.97 p=0.69, cancer HR 0.96 p=0.47), D-Health (Neale 2022), ViDA (Scragg 2017); deficiency-correction-vs-replete distinction. human-evidence-level: limited-negative. Seeder also added a VDR-agonism class to intervention-classes.md (R16).
hypotheses/d3-k2-calcium-partitioning.md (verified:true, status: contested, Mode A) — the D3↑MGP/osteocalcin-substrate + K2-carboxylates synergy (bone-vs-artery calcium partitioning / “calcium paradox”). Aggregates the genetic/animal MGP proof, dp-ucMGP biomarker RCTs, Rotterdam observational vs the AVADEC combined-arm NULL + VITAL/D-Health D3 nulls + the putative (not established) D3→MGP-transcription link. Builds on bone.md §6 bone-vascular paradox.
verifier corrections (adversarial pass earned its keep): vitamin-d — D-Health n 2,501→21,315 (8.5× seeder error); all-cause-mortality HR 0.88→1.04 (fabricated; actual point estimate null, slightly ABOVE 1); D-Health cancer-mortality is a nominally-significant INCREASE (HR 1.24, p=0.05) — opposite VITAL’s decrease, now framed honestly; Sanders p 0.02→0.03; Hahn p 0.04→0.05; ViDA DOI typo. VITAL/ViDA/Chandler/Costenbader confirmed against full PDFs; PubChem CIDs (5280795/5280453) confirmed. hypothesis — El Borolossy vol/issue/pages corrected (76(1):98–104→76(6):848–854); AVADEC footnote pages corrected + exact dp-ucMGP result added; Sneddon 1999 promoter positions disentangled (−420/−430); added a MISSING AVADEC CAC substudy (Hasific 2023, JACC Adv, primary-null, CAC≥400 subgroup p=0.047); status:contested confirmed defensible. No supersession found (2023–26 reviews/RCTs concordant with the null framing; UK Biobank emulation PMID 41719624 supports the deficiency-threshold reading).
propagated (main agent, 5 pages): de-orphaned the hypothesis (now 6 inbound links) — cross-linked from vitamin-k2.md (new “Combined with vitamin D3” subsection), matrix-gla-protein.md (Related pages), processes/vascular-calcification.md (K2-intervention subsection), tissues/bone.md §6, and vitamin-d.md See-also. All additive reference links (no quantitative-claim edits → verification intact).
open gaps / follow-ups: 5 dangling stub links off vitamin-d (vitamin-d-receptor, cyp27b1, calcium-homeostasis, osteoporosis exists?, secondary-hyperparathyroidism); no dedicated studies/ pages for VITAL/D-Health/AVADEC/El Borolossy (cited by DOI). Two schema escalations surfaced (NOT acted on): (1) translation-gap: has no enum for “replete-population-null” (used human-evidence-strong); (2) proposed-by: has no convention for a diffuse community hypothesis.
leak-gate: clean (checks #1 + #3 on new/changed .md).
context: follow-on to round 2 — user asked to seed/verify ALL the suggested link-target stubs surfaced by the new autonomic/HPA pages. Triaged by inbound-demand (lint-pass §3 count), then seeded in two cluster-batches. Built from canonical primary literature; Errico npj-Aging-2025 excluded throughout.
page-type decisions: all protein pages except Schwann (cell-type). Two combined pages per the tsc1-tsc2/foxo-transcription-factors precedent: hsd11b1-hsd11b2.md (the cortisol↔cortisone shuttle) and monoamine-oxidase.md (MAOA+MAOB, MAOB primary/aging-relevant). “cortisol/HPA” and “catecholamines/β-adrenergic” stayed as the round-2 single pages; these are the component proteins.
leak-gate: clean (checks #1 + #3 on new/changed .md; 9 new pages grepped explicitly since the diff-scoped gate skips untracked files).
[2026-06-14] ingest | vitamin-K vascular evidence — 2 study pages + propagation (from a Physionic YouTube reference-scrape)
context: user shared a vitamin-K video (Physionic, “We’ve been lied to about Vitamin K”) to mine per sops/scraping-youtube-references.md — video NOT citable, used as a lead list. Pulled the transcript (yt-dlp) + the 19-DOI description reference list. Video thesis (most RCTs null; signal clusters in CKD/diabetic/sick subgroups; K2>K1 observationally; K1 = diet-quality marker) REINFORCES the existing wiki framing — no contradiction. 11/19 refs already in-wiki; 9 new leads. User chose to ingest 4 (Braam 2004 + Shea 2009 + Bellinge 2021 + Asemi 2016) + the K1-vs-K2 nuance.
added (2 study pages, seeded→verified):
studies/shea-2009-vitamin-k-cac.md (verified:true, full PMC PDF read) — n=388 healthy older adults, 3 y, 500 µg/d phylloquinone (K1) added on a Ca-600+D-400 multivitamin background vs same without K1. ITT NULL for CAC progression (27±6 vs 37±7 Agatston, p=0.26); positive only in the pre-specified ≥85%-adherent secondary analysis (17 vs 37, p=0.03) + pre-existing-CAC subgroup (25 vs 59, p=0.03); effect MGP-independent. The trial carrying ~94% of the Li-2023 CAC meta weight — a K1 study with a null primary endpoint. Verifier corrections: “post-hoc”→“pre-specified secondary”; control arm is Ca+D (not placebo); adherent n=295 analytical vs 367 abstract.
studies/braam-2004-vitamin-dk-vessel-elasticity.md (verified:false, closed-access Thieme, #gap/no-fulltext-access; all abstract figures confirmed via Crossref) — 3-arm RCT (placebo / minerals+D / minerals+D+K1), n=181 postmenopausal women (108 analysed), 3 y; combined D+K1 arm preserved carotid elasticity (DC 8.8% / CC 8.6% / pulse-pressure 6.3%, p<0.05; Young’s modulus 13.2%, p<0.01; IMT NS) while placebo and D-alone deteriorated. Early combined D+K functional-positive — but K1, not MK-7. Verifier added the missed pulse-pressure result.
propagated (main agent):
hypotheses/d3-k2-calcium-partitioning.md — new “Combined D+K intervention trials” subsection (Braam 2004 [evidence-for frontmatter + footnote] + Asemi 2016 [footnote]); both caveated that NEITHER is MK-7+D3 on a hard endpoint (Braam=K1; Asemi=+Ca 3-way, 12wk, authors call CIMT effect “could be a chance finding”); taxonomy-table rows added.
processes/vascular-calcification.md — formalized the Shea 2009 citation (was discussed by name w/o a footnote): added shea-2009-vitamin-k-cac link + the ITT-null/per-protocol/K1 nuance + [
molecules/compounds/vitamin-k2.md — Shea 2009 in the CAC section (largest CAC RCT is K1 + ITT-null); Bellinge 2021 in observational (Danish cohort n=53,372/21 yr, K1 HR 0.79 / K2 HR 0.86 — similar despite different food sources ⇒ diet-quality-marker signal, reinforces the adjustment caveat); [^shea2009]+[^bellinge2021] footnotes.
incidental fix (pre-existing error on a verified page): the [^vermeer2020] footnote on vitamin-k2.md mis-attributed doi:10.3390/nu12102909 to “Vermeer C, van Ballegooijen AJ” — Crossref confirms it is Vlasschaert C, Goss CJ, Pilkey NG, McKeown S, Holden RM (2020). Renamed the citation key vermeer2020→vlasschaert2020 across all 5 sites + the (unseeded) studies-page link slug; corrected the author string. Surfaced by the video’s Study 673.
NOT ingested (available leads, logged for later): Gast 2009 (PROSPECT K2→CHD), Erkkilä 2004/2007 (K1-as-diet-marker), Oikonomaki 2019 (CKD K2 RCT), Fulton 2016 (older vascular-disease K RCT, no DOI in description). Verifier-surfaced newer K1/CAC trials worth a future look: ViKTORIES (Lees 2021), Bellinge 2022 (18F-NaF PET), Macias-Cervantes 2025 (IV K1 HD).
follow-ups: seed the studies/vlasschaert-2020-vitK-cv-systematic-review.md page (currently a dead link); Shea/Braam numbers could optionally promote Bellinge/Asemi to studies pages if cited again.
leak-gate: clean (#1 + #3 + explicit grep on untracked new study pages).
context: user asked to seed/verify iron + ferritin pages with longevity, healthy-aging, AND women’s-health lenses. molecules/metabolites/iron.md already existed (verified 2026-06-03, overload/longevity-centric). [[ferritin]] was wikilinked from iron.md’s homeostasis-proteins: frontmatter but had NO page — only the ftl1.md subunit page (which explicitly defers serum-ferritin-biomarker content “elsewhere”). Seeded the missing holoprotein/biomarker page.
added (seeder→verifier, full adversarial pass):
molecules/proteins/ferritin.md (verified:true, 8 primary PDFs read) — type: protein protein-complex page (FTH1 P02794 primary + FTL P02792 in complex-subunits:; both UniProt masses/lengths confirmed live). Covers: 24-subunit nanocage + IRE/IRP + NCOA4-ferritinophagy + FTMT; serum ferritin as acute-phase reactant (the central interpretive trap; IL-6-driven; CRP co-interpretation); iron-deficiency thresholds (legacy WHO <12/15 vs modern Mei-2021/Cochrane-2021 ~25–30 µg/L, IDWA underrecognition); women’s health (premenopausal IDWA fatigue/cognition/RLS burden, Barton-2024 prevalence 7.4%→15.3% by threshold, balanced dual-reading of postmenopausal ferritin rise); longevity (J-shaped ferritin↔mortality, MR [Moksnes/Barad/Chen], inflammaging/MAFLD). mr-causal-evidence: partial; tier 3.
verifier corrections (caught real errors): Krayenbuehl 2011 fatigue instrument FABRICATED — wiki said “Piper Fatigue Scale”, paper uses Brief Fatigue Inventory (0–10); whole-cohort 6-wk endpoint was NS (p=0.07), the 1.8-vs-0.4/p=0.005 result is the ≤15 ng/mL subgroup (n=34) only — clarified. DePalma 2021 mischaracterized as a new prospective RCT → actually a review/analysis of FeAST (VA CSP 410, n=1,277 PAD) data; corrected body + footnote. Article-number fixes: Chen 2024 9430→9179, Moksnes 2022 574→591, Li 2023 417→419. Barton prevalence refined to source precision (7.43/15.33/36.10%). hgnc:HGNC:3976→3976 (bare-number convention); mouse-ortholog slash→semicolon-complex format Fth1 (FTH1); Ftl1 (FTL).
supersession (R25): verifier surfaced Addo 2025 Lancet Glob Health (doi:10.1016/S2214-109X(25)00009-9; n=18,251/12 countries; pooled women’s Hb-based threshold 24.8 µg/L, p-het 0.73) — internationally replicates Mei-2021’s ~25 µg/L, SUPPORTS not overturns. Main agent added [^addo2025lancetgh] footnote + threshold-table row.
propagated (main agent): iron.md — added a women’s-health balancing paragraph (iron deficiency = most common nutritional deficiency, IDWA healthspan cost, low-iron is double-edged not just protective) + [[ferritin]] link, so the page no longer frames menstrual loss as purely protective. Inbound [[ferritin]] links added to ftl1.md (cross-refs), processes/ferroptosis.md, phenotypes/anemia-of-aging.md, hallmarks/chronic-inflammation.md (ferritin as IL-6 acute-phase inflammaging proxy), phenotypes/menopause.md (postmenopausal rise in CVD-convergence section), hypotheses/female-longevity-advantage.md (iron-hypothesis section + resolved two stale #gap/stub notes since iron/ferroptosis/ferritin all now exist).
gaps surfaced / left flagged: Sullivan-1981 per-cycle menstrual-iron figure stays #gap/unsourced + #gap/no-fulltext-access (no primary re-verification); iron-deficiency threshold has no single globally-adopted cutoff (#gap/contradictory-evidence); ferritin-specific MR weaker than serum-iron/TSAT MR; [[fth1]]/[[hepcidin]]/[[ferroportin]] remain implicit stubs (pre-existing, referenced from iron.md). GenAge has no FTH1/FTL entry → nomination candidate.
context: follow-on to the ferritin seed — closed the three remaining dangling iron-cluster wikilinks ([[hepcidin]], [[ferroportin]], [[fth1]]) referenced from iron.md/ferritin.md/anemia-of-aging.md. Parallel seeder batch (ferroportin seeder crashed mid-run on a socket error → re-dispatched), then parallel verifier batch. Continued the aging + women’s-health emphasis.
verifier corrections (each adversarial pass caught real errors): hepcidin — Nemeth 2004 IL-6→hepcidin fold-rise 30-fold→7.5-fold (fabricated), timecourse fixed (peak 2 h post-infusion, serum iron −34%); Dugan 2026 CR↔hepcidin result direction INVERTED (wiki said “decreased”; actual = NULL, no group×time interaction) → section rewritten; Nicolas 2001 is the USF2-KO (not Hamp-KO) mouse; Chew 2025 DOI + design (open-label Ph2 not retrospective); rusfertide confirmed mimetic/agonist, VERIFY Ph3 n=293 ongoing/no efficacy yet; Arezes ERFE-BMP6 citation fully wrong (seeder: Blood Adv 2020 + DOI→unrelated MSC paper) → corrected by main agent to Arezes 2018 Blood 132(14):1473–1477, doi:10.1182/blood-2018-06-857995 (PubMed-confirmed). ferroportin — 4 wrong iron-binding residues fixed (D39/H43/C326/H507); Sangkhae 2020 ERFE-placenta mechanism was FABRICATED (paper has no ERFE; placental FPN is downregulated in maternal ID via IRP1/IRE) → removed; Raha 2022 brain regions/cell-types corrected; added Kattamis 2025 Ph2a vamifeport β-NTDT. FTH1 — Theil 2011 DOI wrong (was 10.1042/BJ20102021 Biochem J → correct 10.1016/j.cbpa.2011.01.004 Curr Opin Chem Biol; fixed on BOTH fth1.md and ferritin.md); Wu 2024 TET2→TET3; Zeng 2025 species mouse+porcine (not human) + page range; Feng/Mi page-range & year fixes.
propagated (main agent): ferritin.md (“planned fth1”→live + scope note), ftl1.md (fth1 planned→verified), anemia-of-aging.md (hepcidin/ferroportin mechanism wikilinks), ferroptosis.md (added fth1 to key-proteins — primary endogenous ferroptosis suppressor), il-6.md (added hepcidin to causes: + acute-phase body link), hematopoietic-stem-cells.md (see-also fth1, Yi-2024 caveated as abstract-level). Resolved #gap/stub markers on hepcidin.md (ferroportin now exists).
gaps left flagged: Yi 2024 / Zeng 2025 / Mi 2022 closed-access → #gap/no-fulltext-access (abstract-level only); Galy 2023 + Porter 2021 fulltext-blocked; GenAge has no HAMP/SLC40A1/FTH1 entry (nomination candidates). Optional follow-ups: seed studies/kattamis-2025-vamifeport-ntdt.md; a pathways/ MOC for the hepcidin-ferroportin axis if it earns inbound links.
context: follow-on surfaced by the ferroportin verifier — the lead clinical-trial evidence that ferroportin is a druggable target. Seeded+verified against the gold-OA full text.
added: studies/kattamis-2025-vamifeport-ntdt.md (verified:true, full PMC PDF read) — Kattamis A et al., Orphanet J Rare Dis 20(1):608, doi:10.1186/s13023-025-04119-y, PMID 41291806, PMC12648932, NCT04364269. Phase 2a double-blind RCT, n=25 (vamifeport QD 9 / BID 12 / placebo 4), 12 wk, ~20 sites/5 countries, non-transfusion-dependent β-thalassemia. Primary = safety: zero SAEs/deaths/severe TEAEs; 2 discontinuations (1 QD hemolysis treatment-related, 1 BID consent). PD (secondary): on-target acute iron restriction — 2h serum iron QD −12.2 / BID −14.5 µmol/L, TSAT QD −33.6% / BID −37.2% vs placebo; Hb stable; hepcidin unchanged (drug acts downstream); MCV ↓ p<0.01 (BID). Framed honestly: β-thalassemia is NOT an aging disease; wiki relevance = clinical validation of ferroportin pharmacology referenced by ferroportin/hepcidin/iron; no aging-indication data.
verifier corrections: sites 17→~20 (17 was from ClinicalTrials.gov not the paper); discontinuations 1→2; MCV p<0.01 is BID-specific (not both arms); Week-12 serum-iron/TSAT group means flagged #gap/unverifiable-from-main-text (supplemental tables only); ERFE/sTfR “not measured” absence claim softened (paper never names them). Author byline confirmed 9 (Porter J 9th).
propagated: ferroportin.md [^kattamis2025] footnote now wikilinks the study page + 2h (paper-confirmed) values kept, Week-12 specifics attributed to supplemental/CTgov provenance; vol corrected to 20(1):608. [[molecules/compounds/vamifeport]] left as an implicit stub (intervention frontmatter).
context: ad-hoc seed — the DNA-methylation alcohol-consumption biomarker (DNAm-Alc) and its top functional locus SLC7A11. Seeder→verifier each; triggered by a question on what markers the TruAge alcohol score uses + what the top hits do.
biomarkers/dnam-alcohol-consumption.md (verified:true; modality dna-methylation, training-target exposure) — Liu 2018 144-CpG LASSO DNAm-Alc predictor (n=13,317); AUC 0.90–0.99 heavy-vs-non-drinker but only ~7.6% variance in general populations (Yousefi 2019 ALSPAC); dynamic — reverts with sustained abstinence (Framingham ~4 yr); TruDiagnostic reports it as an age/sex-matched percentile (proprietary transform); explicitly NOT a GrimAge component; “downstream biology — marker vs mechanism” section (SLC7A11 is the one hit elevated to mechanism).
schema: R56 — added exposure to the type: biomarkertraining-target: enum (CLAUDE.md + schema-history.md). For DNAm/omic predictors of an external exposure (alcohol, smoking intensity, BMI, diet); distinct from a type: exposure page (R53) — the two cross-link.
verifier corrections (each adversarial pass caught real errors):
dnam-alcohol — MAJOR: cg06690548/SLC7A11 attribution moved Liu 2018 → Lohoff 2022 (SLC7A11 was NOT among Liu’s top loci, which were TXLNA/LETM1/HNRNPA1/GABRD-GABBR1); “frontal cortex”→prefrontal cortex (PFC); AUDIT variance 9.8%→8.8% primary (9.8% = sensitivity analysis); the 7.6% ALSPAC figure reattributed Liu→Yousefi 2019; 3 supporting footnotes (koppula2021 / jiang2009 / hnrnpa1review2021, first author Clarke JP) Crossref-confirmed.
slc7a11 — Habib 2015 “NRF2 knockdown reduces xCT” was wrong (paper did NRF2/KEAP1 overexpression, no NRF2 siRNA) → rewritten with fold-changes; Jiang 2015 p53-3KR also fails to induce senescence (added; K117R/K161R/K162R, EMSA/ChIP, xenograft rescue); He 2023 ATF4 closed-access → #gap/no-fulltext-access (direction cross-checked via Koppula 2021).