R10

log/R10.md — Round 10 entries

Sub-file of log — see parent for index.

[2026-05-05] verify | canakinumab

molecules/compounds/canakinumab.md

• page: molecules/compounds/canakinumab.md
• sources verified against full PDF (2):
  • Ridker 2017 NEJM (10.1056/NEJMoa1707914) — 13 pp read end-to-end; CANTOS primary RCT
  • Svensson 2022 JAMA Cardiology (10.1001/jamacardio.2022.0386) — 8 pp read end-to-end; downloaded via PMC during this pass (was previously failed status; bronze OA)
• sources unverifiable (not_oa):
  • Ridker 2017 Lancet (10.1016/s0140-6736(17)32247-x) — not_oa; lung cancer HR claims tagged no-fulltext-access; remain as-is
• ChEMBL identity: CHEMBL1201834 confirmed as CANAKINUMAB via ChEMBL API
• corrections made (6):
  1. Fatal infection rate: “~0.31 vs ~0.18 per 100 person-years” → “0.31 vs 0.18 events per 100 person-years (P=0.02)” — paper gives exact figures; tildes removed, P-value added
  2. Same infection rate correction in Translation gap section
  3. Svensson 2022 characterization: “exploratory post-hoc subgroup analysis” → “pre-specified exploratory genomic substudy” — paper explicitly states pre-specified design
  4. Svensson 2022 n: “exact n not verified” → n=3,923 genomic substudy; 338 (8.6%) with CHIP; TET2-CHIP subgroup placebo n=31, canakinumab n=71; HR 0.38 (95% CI 0.15–0.96; P=0.04); P for interaction=0.14; underpowered (26% power per paper limitations)
  5. CANTOS results section expanded with exact incidence rates per 100 PY per dose arm, pre-specified threshold values (p<0.02115 vs p<0.01058), hsCRP reductions as percentage-point differences vs placebo (not percentage reductions), and all-cause mortality HR
  6. Footnote [^ridker2017nejm] updated with per-arm n’s and full HR/CI/p-value detail
• verified flag flipped: true; verified-scope notes Lancet companion and FDA approval dates not re-verified
• downstream pages with CANTOS quantitative data needing propagation review (listed for main agent): chronic-inflammation.md, atherosclerosis.md, nf-kb.md, clonal-hematopoiesis.md, senomorphics.md, cancer.md

[2026-05-05] verify | akkermansia-muciniphila

microbiome/akkermansia-muciniphila.md

• page: microbiome/akkermansia-muciniphila.md
• sources verified against full PDF (2):
  • Depommier 2019 (10.1038/s41591-019-0495-2, Nat Med 25:1096–1103) — 10 pp read end-to-end (Europe PMC author manuscript)
  • van Passel 2011 (10.1371/journal.pone.0016876, PLoS ONE 6(3):e16876) — 8 pp read end-to-end
• sources unverifiable (not_oa):
  • Derrien 2004 (10.1099/ijs.0.02873-0) — not_oa; named-for claim and species description tagged no-fulltext-access
  • Plovier 2017 (10.1038/nm.4236) — not_oa; Amuc_1100/pasteurization quantitative claims tagged no-fulltext-access
  • Png 2010 (10.1038/ajg.2010.281) — not_oa; IBD/T2D depletion claims tagged no-fulltext-access
• sources not yet downloaded (pending OA):
  • Everard 2013 (10.1073/pnas.1219451110) — PNAS green OA; mouse HFD quantitative claims unverified
  • Biagi 2010 (10.1371/journal.pone.0010667) — PLoS ONE OA; centenarian AKK direction claim unverified
  • Rampelli 2020 (10.1128/mSystems.00124-20) — mSystems OA; centenarian AKK direction claim unverified
  • Forslund 2015 (10.1038/nature15766) — pending; metformin-AKK claims unverified
• n-value conflict resolved: “32 enrolled/27 completed” (task brief) was wrong; PDF (p.2 + p.7) states n=40 enrolled / n=32 completed (Placebo=11, Pasteurized=12, Alive=9); dysbiosis.md was correct
• corrections made (10):
  1. n enrolled/completed: “32 enrolled/27 completed” (task brief) → “40 enrolled / 32 completed (Placebo=11, Pasteurized=12, Alive=9)” — sourced directly from PDF Methods p.7
  2. Genome size: “2.66 Mb” → “2.664 Mb (2,664,102 bp)” — from van Passel Table 1
  3. G+C content: not on page → “55.8%” added from van Passel Table 1
  4. Protein-coding genes: “~2,200” → “2,176 predicted ORFs (coding capacity 88.8%; 1,408 COG-assigned, 768 hypothetical)” — from van Passel Table 1 and Results
  5. frontmatter genome-size-mb: “2.66” → “2.664”
  6. Body weight figure: “~2.3 kg vs +0.5 kg placebo” (approximation) → “−2.27±0.92 kg (P=0.09)” with fat mass (−1.37±0.82 kg) and hip circumference (−2.63±1.14 cm) added; P=0.09 significance status clarified (non-significant)
  7. Inclusion criteria: “BMI 25–50; fasting plasma glucose…” → full NCEP ATP III metabolic syndrome criteria as stated in Methods; HbA1c ≤7.5% exclusion threshold added
  8. Primary outcome bullets: insulin sensitivity (+28.62±7.02%, P=0.002), insulinemia (−34.08±7.12%, P=0.006), cholesterol (−8.68±2.38%, P=0.02) quantitated in full
  9. Plasma LPS reduction bullet added (significant vs baseline and placebo per PDF)
  10. Hepatic markers bullet added (γGT −24% vs placebo, P=0.009; AST reduced; ALT not significantly changed)
• depommier2019 footnote: n-value hedge removed; correct n and group breakdown added; full first-author list added
• non-OA footnotes updated with no-fulltext-access tags
• pending-OA footnotes updated with unverified claim warnings
• verified flag: false → true (partial scope per verified-scope field)
• banner: removed

[2026-05-05] verify | clonal-hematopoiesis

phenotypes/clonal-hematopoiesis.md

• page: phenotypes/clonal-hematopoiesis.md
• sources verified against full PDFs (4):
  • Jaiswal 2014 (10.1056/NEJMoa1408617, NEJM 371:2488-2498) — 11 pp read end-to-end
  • Genovese 2014 (10.1056/NEJMoa1409405, NEJM 371:2477-2487) — 11 pp read end-to-end
  • Steensma 2015 (10.1182/blood-2015-03-631747, Blood 126:9-16) — 5 pp read end-to-end
  • Jaiswal 2017 (10.1056/NEJMoa1701719, NEJM 377:111-121) — 5 pp read end-to-end (abstract, methods, results, figures)
• sources unverifiable (download failed):
  • Fuster 2017 (10.1126/science.aag1381) — green OA but 0 candidate URLs resolved; mouse atherosclerosis quantitative values unverified; tagged no-fulltext-access
  • Svensson 2022 (10.1001/jamacardio.2022.0386) — bronze OA, download failed; CANTOS-CHIP subgroup HRs unverified; tagged no-fulltext-access
• Bouzid 2023 (10.1038/s41591-023-02397-2) confirmed as real paper via PubMed (PMID 37322115): Nature Medicine 2023, CHIP is PROTECTIVE against AD (OR=0.64), not a risk factor — wiki framing corrected
• corrections made (8):
  1. CHD CI: 1.2–3.4 → 1.2–3.5 (P=0.02) — from Jaiswal 2014 p.2496 discussion text
  2. Ischemic stroke CI: 1.4–4.8 → 1.3–4.8 (P=0.003) — from Jaiswal 2014 p.2496 discussion text
  3. Jaiswal 2017 n: “n=4,726 (cases/controls across 4 cohorts)” → 4,726 CHD cases + 3,529 controls — from paper Methods
  4. Jaiswal 2017 CHD HR: “approximately 2×” → HR=1.9 (95% CI 1.4–2.7; P<0.001) — from Jaiswal 2017 Fig. 1A
  5. Driver gene table: unsourced percentage frequencies (~46%, ~16%, ~11%) replaced with raw variant counts from Jaiswal 2014 Fig. 2A and Genovese 2014 Fig. 2B
  6. Microglial CHIP section: speculative neurodegeneration risk framing removed; Bouzid 2023 protective OR=0.64 correctly characterized and cited
  7. Steensma 2015 footnote: article type corrected to “Perspectives (working-group proposal)”; authors added; access tag updated (PDF now verified)
  8. Genovese 2014 footnote: death HR=1.4 (95% CI 1.0–1.8) added; 42% retrospective clonality finding added; authors added
• propagation applied to: cell-types/hematopoietic-stem-cells.md — CHD CI 1.2–3.4 → 1.2–3.5 and stroke CI 1.4–4.8 → 1.3–4.8 corrected (that page is the canonical home)
• verified flag: false → true (partial scope; Fuster 2017 and Svensson 2022 remain unverified per verified-scope field)
• banner: removed

[2026-05-05] verify | hair-greying

phenotypes/hair-greying.md

• page: phenotypes/hair-greying.md
• sources verified against full PDF (5):
  • Nishimura 2002 (10.1038/416854a, Nature 416:854–859) — 6 pp, read end-to-end
  • Zhang 2020 (10.1038/s41586-020-1935-3, Nature 577:676–681) — 10 pp read; PMC author manuscript
  • Sun 2023 (10.1038/s41586-023-05960-6, Nature 616:774–782) — 10 pp read end-to-end
  • Rosenberg 2021 (10.7554/eLife.67437, eLife 10:e67437) — 20 pp read end-to-end
  • Adhikari 2016 (10.1038/ncomms10815, Nat Commun 7:10815) — 12 pp read end-to-end
• sources unverifiable (not_oa):
  • Nishimura 2005 (10.1126/science.1099593) — not_oa; incomplete McSC self-renewal claims carry no-fulltext-access
  • Wood 2009 (10.1096/fj.08-125435) — not_oa; H2O2/catalase mechanism claims carry no-fulltext-access
• DOI confirmed: Nishimura 2005 DOI 10.1126/science.1099593 verified correct via Crossref (title: “Mechanisms of Hair Graying: Incomplete Melanocyte Stem Cell Maintenance in the Niche”)
• corrections made (9):
  1. Nishimura 2002 page range: 854–860 → 854–859 (p.860 begins a separate Nature letter)
  2. Nishimura 2002 niche location: “bulge” → “lower permanent portion (encompassing bulge and sub-bulge)” per paper’s consistent terminology; transplantation specificity (lower permanent portion only) added
  3. Nishimura 2002 self-renewal detail: 50× repopulation factor added from K14-SLF experiment
  4. Sun 2023 body section: “McSC immobilization model” reframed as “McSC dedifferentiation model” — the paper’s actual central finding is that reversible dedifferentiation (governed by WNT signals) is the maintenance mechanism; aging failure is positional displacement from WNT niche signals, not “immobilization” per se; quantitative aging data added (P=0.0426, P=0.0006)
  5. Zhang 2020 body: expanded with three-stressor design (RTX, restraint, CUS); timecourse added (~50% McSCs in M-phase within 24h, full bulge loss within 5 days for RTX); ADRB2 receptor identity added; immune-independence vs. adrenal-independence distinction made explicit; CDK inhibitor preservation experiments added
  6. Rosenberg 2021 body: n=14 specified (wiki said “n small” without value); stress-mapping subset n=2 explicitly stated; specific mitochondrial proteins named (CPT1A, ACOT7, SOD1, CFL1, PGK1); reversal-rate statistics added (Cohen’s d=0.15, p=0.59)
  7. Rosenberg 2021 footnote: n=14 (7F/7M, mean age 35±13, range 9–65), specific ethnicity breakdown, stress-mapped n=2 added
  8. Adhikari 2016 GWAS: clarified PRSS53, EDAR, FOXL2 are NOT associated with greying — only IRF4 is; full CANDELA cohort n=6,630 (analytical n=6,357) noted; p-values added
  9. MC1R claim: not in Adhikari 2016 — tagged unsourced; requires separate primary citation
• Zhang 2020 footnote: updated from “PDF pending download” to “PDF verified”; stress mechanism detail expanded
• Sun 2023 footnote: updated from “PDF pending download” to “PDF verified”; dedifferentiation framing and aging p-values added
• Rosenberg 2021 footnote: updated from “PDF pending download” to “PDF verified”
• Adhikari 2016 footnote: updated from “PDF pending download” to “PDF verified”; n detail, p-value, and clarification of loci scope added
• banner removed; verified: true (partial scope — Nishimura 2005 and Wood 2009 closed-access)
• downstream pages to check (main agent):
  • phenotypes/skin-aging.md — may cross-link to hair-greying; check that link is consistent with corrected Sun 2023 model name (“dedifferentiation” not “immobilization”)
  • cell-types/melanocyte-stem-cells.md — if it exists, may inherit Nishimura 2002 page-range error or Sun 2023 immobilization framing
  • studies/nishimura-2002-melanocyte-stem-cell-niche.md — if it exists, page range needs same correction (854–859)
  • studies/sun-2023-msc-dedifferentiation-dynamic-niche.md — if it exists, body description should use dedifferentiation framing

[2026-05-05] verify | skin-aging

phenotypes/skin-aging.md

• page: phenotypes/skin-aging.md
• sources verified against full PDF (5): Fisher 1996 (10.1038/379335a0), Demaria 2014 (10.1016/j.devcel.2014.11.012), Hattori 2004 (10.1111/j.0022-202x.2004.22503.x), Velarde 2012 (10.18632/aging.100423), Krutmann 2017 (10.1016/j.jdermsci.2016.09.015) — all read end-to-end
• sources unverifiable (not_oa): Kovacs 2010 (10.1111/j.1365-2133.2010.09946.x), Purohit 2016 (corrected DOI: 10.1016/j.jdermsci.2016.04.004), Sumita 2018 (10.1111/jdv.15020) — tagged no-fulltext-access
• corrections made (10):
  1. Hattori 2004 CRITICAL: SCF source is epidermal keratinocytes, NOT dermal fibroblasts — corrected throughout page and table
  2. Hattori 2004: bFGF claim removed — NOT elevated in LS lesional epidermis (n=10, NS); fibroblast bFGF is a separate model (Kovacs 2010)
  3. Hattori 2004: quantitative values added — SCF mRNA 3.9-fold (n=7, p<0.01), protein 1.6-fold (n=6, p<0.05)
  4. Fisher 1996: MAPK cascade attribution removed — paper documents AP-1/NF-κB transcription factors directly; MAPK upstream context is from later literature
  5. Fisher 1996: TIMP-1 / CBP/p300 mechanism removed — paper describes Jun/Fos–receptor transrepression, not TIMP-1 induction or CBP/p300 competition
  6. Fisher 1996: quantitative data added — AP-1 reduction ~70%; MMP-1/MMP-9 50–80%; n=6–17 (mRNA), n=9–10 (protein/binding); MMP-2 NOT induced
  7. Fisher 1996: explicit MMP-2 non-induction added to body and footnote
  8. Velarde 2012: Sod2+/- → Sod2-/- (homozygous null, CD1, EUK-189); phenotype “wrinkled appearance” → epidermal thinning + thickened stratum corneum; cell type clarified as predominantly epidermal
  9. Demaria 2014: “myofibroblast clearance” → “myofibroblast differentiation”; wound closure peak delay day 6 and PDGF-AA rescue quantified
  10. Krutmann 2017: exposome list expanded to include visible light, nutrition, temperature, cosmetics (previously omitted)
• schema: icd-10 changed from multi-code explanatory string → L57 (single code); alternatives documented in body
• Purohit 2016 DOI corrected: wrong candidate → confirmed 10.1016/j.jdermsci.2016.04.004 via PubMed PMID 27132061
• banner removed; verified: true (partial scope — 3 sources closed-access; Fisher 2009 not in scope this pass)
• downstream pages to check (main agent): cellular-senescence.md, sasp.md (Demaria 2014 PDGF-AA mechanism); senolytics.md, fisetin.md (wound healing claims); any page citing Hattori 2004 for fibroblast SCF

[2026-05-05] verify | androgenetic-alopecia

phenotypes/androgenetic-alopecia.md

• page: phenotypes/androgenetic-alopecia.md
• sources verified against full PDF:
  • Garza 2011 (10.1172/jci44478, JCI 121:613–622) — 10 pp, read end-to-end
  • Nishimura 2002 (10.1038/416854a, Nature 416:854–860) — 4 pp, read end-to-end
• sources unverifiable (not_oa):
  • Matsumura 2016 (10.1126/science.aad4395) — not_oa; COL17A1/transepidermal-elimination claims carry no-fulltext-access
  • Whiting 1993 (10.1016/0190-9622(93)70106-4) — not_oa; terminal:vellus ratio cutpoints carry no-fulltext-access
  • Premanand 2018 (10.1007/s00403-018-1826-8) — not_oa; Wnt/β-catenin and PGD2 claims carry no-fulltext-access
• sources download-failed:
  • Whiting 1999 (10.1038/sj.jidsp.5640230) — bronze OA but no accessible URL after filtering; finasteride reversal claim carries no-fulltext-access
• corrections made (5):
  1. Garza 2011 HFSC marker: wiki stated “CD34+/CD200+” as canonical HFSC population — corrected to KRT15hi/ITGA6+ (stem cells); CD200hiITGA6hi and CD34hi are progenitor populations, distinct from stem cells per Garza 2011
  2. Garza 2011 quantitative data added: CD200hiITGA6hi depleted (2.3%±0.7% haired vs 0.28%±0.1% bald; n=9, P=0.005); CD34hi depleted ~10-fold (1.9%±1% vs 10.5%±0.3%; n=3, P=0.01); KRT15hi preserved (n=8, P=0.3); study subjects n=54 male donors aged 40–65
  3. Nishimura 2002 TGF-β2/SCF claim removed: wiki attributed “niche signals (specifically, stem cell factor from the niche and TGF-β2) determine McSC fate” to Nishimura 2002 — the 2002 paper only establishes niche localization; TGF-β signaling is from Nishimura et al. 2010 (doi:10.1016/j.stem.2009.12.010); text corrected and note added pointing to the right paper
  4. Hamilton citation corrected: wiki cited “Am J Anat 1951;86(3):399–476” — no such paper exists. Two actual Hamilton papers found: Am J Anat 1942;71(3):451–480 (doi:10.1002/aja.1000710306, castration/androgen work) and Ann N Y Acad Sci 1951;53:708–728 (doi:10.1111/j.1749-6632.1951.tb31971.x, pattern classification). Footnote corrected with both DOIs and explanation.
  5. Inline patterned distribution citation updated to correctly identify the Ann N Y Acad Sci 1951 paper as the classification source.
• ICD-11 code: WHO API requires bearer-token auth; browser not accessible via WebFetch; left null with unsourced; recommend manual lookup
• banner removed; verified: true (partial scope — 4 of 8 sources unverifiable; see verified-scope)
• downstream pages checked: phenotypes/hair-greying.md — already verified; cites Nishimura 2002 correctly for niche localization only, no TGF-β2 error present. No other pages reference these corrected claims.

[2026-05-05] verify | c-myc

molecules/proteins/c-myc.md

• page: molecules/proteins/c-myc.md
• sources verified against full PDF:
  • Hofmann 2015 (10.1016/j.cell.2014.12.016, Cell 160:477–488) — 12 pp + figures, read end-to-end
  • Takahashi & Yamanaka 2006 (10.1016/j.cell.2006.07.024, Cell 126:663–676) — 14 pp, read end-to-end
  • Dang 2012 (10.1016/j.cell.2012.03.003, Cell 149:22–35) — 14 pp, read end-to-end
  • Zindy 1998 (10.1101/gad.12.15.2424, Genes Dev 12:2424–2433) — 10 pp, read end-to-end
• sources unverifiable (closed-access): Dalla-Favera 1982 (10.1073/pnas.79.24.7824) — not_oa; tagged no-fulltext-access
• identity fields confirmed: UniProt P01106 (454 aa, bHLH-LZ, Thr-73/Ser-77 FBXW7 degron) via REST API; GenAge ID 39 via genomics.senescence.info
• corrections made (7):
  1. Hofmann 2015 lifespan figure: “~15%” → exact sex-disaggregated numbers added: 20.9% females (n=37/37), 10.7% males (n=42/42), 15.1% combined; strain C57BL/6; log-rank p<0.001; all effect sizes with p-values from figures added to body and footnote
  2. Hofmann 2015 important caveat added: Myc+/- mice do NOT improve stress management pathways (ROS, DNA damage, p21/p16, senescence) — key contrast with other longevity models; wiki omitted this
  3. Hofmann 2015 mechanistic framing corrected: AMPK elevation is likely secondary to higher metabolic rate, not a direct MYC→AMPK effect per authors’ Discussion
  4. Takahashi 2006 tumor claim corrected: wiki incorrectly stated “chimeric mice developed tumors attributable to c-MYC reactivation” — this paper shows only teratomas from subcutaneous iPSC injection into nude mice (pluripotency assay); chimeric-mouse tumor finding is from Okita 2007 (not cited here), now flagged and corrected
  5. Dang 2012 half-life: “~20–30 minutes” → “~15–20 minutes” per Dang 2012 p.25
  6. Gene regulation figure: “10–15% of all human genes” → “~15% of promoters broadly bound; 15–20% of genes regulated” per Dang 2012 / Hofmann 2015
  7. Zindy 1998 cytochrome c claim: removed unsourced assertion that apoptosis “requires cytochrome c release via the intrinsic mitochondrial pathway” — Zindy 1998 establishes ARF/p53-dependence of apoptosis but does not discuss cytochrome c; tagged unsourced
• schema corrections:
  • [[sustained-proliferative-signaling]] removed from hallmarks frontmatter — no such hallmark page exists in /hallmarks/; dangling wikilink
  • [^lu2020] footnote defined locally on this page (was cross-referenced from partial-reprogramming.md only; footnotes do not transit pages)
  • [^dalla-favera1982] updated with no-fulltext-access tag
• banner removed; verified: true (partial scope — Dalla-Favera 1982 closed-access)
• downstream pages to check (main agent): partial-reprogramming, information-theory-of-aging, mtor, ampk — all cite Hofmann 2015-derived claims that may need the corrected sex-disaggregated lifespan numbers

[2026-05-05] verify | ear-nose-enlargement

phenotypes/ear-nose-enlargement.md

• page: phenotypes/ear-nose-enlargement.md
• sources verified against PDF: Heathcote 1995 (10.1136/bmj.311.7021.1668, downloaded this pass via Europe PMC PMC2539087); Verzijl 2000 (10.1074/jbc.M006700200, downloaded this pass via camoufox)
• sources unverifiable (not_oa): Sforza 2009 (10.1016/j.forsciint.2009.02.019); Verzijl 2002 (10.1002/1529-0131(200201)46:1<114::AID-ART10025>3.0.CO;2-P) — both tagged no-fulltext-access in footnotes
• corrections made (5):
  1. “ear length” → “left ear length” — Heathcote 1995 measured only the left ear; never tested the right ear
  2. “approximately 0.22 mm/year” with “p<0.001” removed → “0.22 mm/year (95% CI: 0.17–0.27 mm/year)” — paper reports CI for regression slope; no explicit p-value stated in the text
  3. “width not significantly associated with age” removed — fabricated claim; Heathcote 1995 measured length only; ear width was never measured
  4. “not attributable to secular trends” removed — phrase not present in the source paper
  5. Verzijl 2000: skin half-life “~15 years” → “14.8 years (95% CI: 9.4–22.3 years)”; n and sample provenance added to footnote; clarified articular (femoral condyle) origin; extrapolation to auricular/nasal cartilage explicitly flagged needs-replication in body
• schema decisions: icd-10/icd-11: null confirmed correct; cellular-senescence kept out of underlying-hallmarks frontmatter (secondary contributor documented in body only)
• banner removed; verified: true (partial scope — Sforza 2009 and Verzijl 2002 closed-access, claims tagged)
• downstream pages to check (main agent): none identified

[2026-05-05] verify | creatine

molecules/compounds/creatine.md

• page: molecules/compounds/creatine.md
• sources verified against full PDF:
  • Chilibeck 2017 (10.2147/OAJSM.S123529) — 14 pp, read end-to-end
  • Kreider 2017 ISSN position stand (10.1186/s12970-017-0173-z) — 18 pp, read end-to-end
  • Prokopidis 2022 (10.1093/nutrit/nuac064) — 12 pp, read end-to-end
  • Forbes 2018 BMD meta-analysis (10.3389/fnut.2018.00027) — 7 pp, read end-to-end
  • Forbes 2022 letter (10.1159/000520967) — 4 pp, read end-to-end
• sources unverifiable (closed-access, not_oa):
  • Devries & Phillips 2014 (10.1249/MSS.0000000000000220) — tagged no-fulltext-access
  • Avgerinos et al. 2018 (10.1016/j.exger.2018.04.013) — tagged no-fulltext-access
  • Dangott 2000 (10.1055/s-2000-8848) — tagged not_oa in footnote (not verified)
  • Forbes 2021 narrative review (10.1080/19390211.2021.1877232) — not_oa (not verified)
• corrections made (7):
  1. Brain PCr increase range: “~3–8%” → “~5–15%” — Kreider 2017 states 5–15% (pg 10); wiki had understated lower bound
  2. Prokopidis 2022 n: “not specified in abstract” → n=225 (122 creatine, 118 placebo) — explicit in paper Methods/Table 2
  3. Prokopidis 2022 older-adult subgroup CI added: SMD=0.88 (95%CI 0.22–1.55, I²=83%) — CI was absent from wiki
  4. Prokopidis 2022 overall I² added: I²=66% — missing from wiki
  5. Forbes 2018 BMD footnote n: “n=brief meta-analysis” → “n=193 (5 RCTs)” — explicit in abstract
  6. Forbes 2022 dose threshold: “>0.3 g/kg/day” → “≥0.1 g/kg/day (~7–9 g/day)” — paper compared 0.1 g/kg/day vs 5 g/day flat; wiki had wrong threshold by 3×
  7. Forbes 2022 year in text: “A 2021 commentary” → “A 2022 letter” — published in Ann Nutr Metab 2022;78 (online Nov 2021)
  8. ISSN position summary: reworded to match position statement #1 and #2 verbatim framing; removed inaccurate “3–5 g/day” qualifier from the safety ceiling statement
  9. Chilibeck 2017 footnote: added full 95% CIs for all three outcomes; removed “pending” note
  10. Devries 2014 and Avgerinos 2018 table rows and footnotes: tagged no-fulltext-access
  11. Banner removed; verified: true flipped
• downstream pages to check:
  • sarcopenia — cites creatine’s lean mass effect; verify Chilibeck 2017 CI propagation
  • senolytics — unlikely but check for any creatine cross-ref
  • Any study pages for chilibeck-2017, kreider-2017, prokopidis-2022, forbes-2018-bmd, forbes-2022-bmd if they exist as stubs

[2026-05-05] verify | sox2

molecules/proteins/sox2.md

• page: molecules/proteins/sox2.md
• sources verified against full PDF:
  • Avilion 2003 (10.1101/gad.224503, downloaded this pass via camoufox, 15 pp, read end-to-end)
  • Suh 2007 (10.1016/j.stem.2007.09.002, downloaded this pass via camoufox, 14 pp, read end-to-end)
• sources consistency-checked (previously verified on partial-reprogramming.md and other pages): Lu 2020 (10.1038/s41586-020-2975-4), Yang 2023 (10.1016/j.cell.2022.12.027), Ocampo 2016 (10.1016/j.cell.2016.11.052)
• sources unverifiable: Ferri 2004 (10.1242/dev.01204) — confirmed closed-access (not_oa per a local paper archive); quantitative claims unverified; tagged no-fulltext-access in body and footnote
• canonical-DB check: UniProt P48431 verified via REST API (317 aa confirmed; HMG-box 41-109 confirmed; single canonical isoform confirmed)
• corrections made (5):
  1. LETHALITY STAGING — Avilion 2003: “peri-implantation lethality at E5.5-E6.5” → “abnormal implants at ~6.0 dpc; no homozygous mutants after 5.5-7.5 dpc” (source: Table 1 and body text)
  2. MULTIPOTENCY QUALIFIER — Suh 2007: removed “oligodendrocytes” from in vivo multipotency claim; in vivo fate-mapping showed neurons (~89%) and astrocytes (~6-10%) only; oligodendrocyte production was in vitro only (coculture with P0 hippocampal neurons); quantitative percentages from Table 1 added
  3. LU 2020 DELIVERY — corrected from “AAV2-mediated” to “Tet-Off AAV2 dual-vector (AAV2-TRE-OSK + AAV2-rtTA), intravitreal”; mouse ages corrected (crush: 1- and 3-month-old; aging vision: 3- and 12-month-old; the “5- and 20-month-old” were the systemic AAV9 safety-test cohort, not the RGC experiments)
  4. AVILION 2003 footnote: corrected design description; added Table 1 n data; updated archive status to PDF downloaded
  5. SUH 2007 footnote: updated to PDF downloaded; added quantitative n and Ki67 data; clarified oligodendrocyte scope
• downstream pages to check:
  • lu-2020-osk-vision-restoration — mouse age claims and AAV delivery description; propagate if inherited from this page
  • partial-reprogramming — check Lu 2020 mouse age description; propagate if “5- and 20-month-old” appears
  • avilion-2003-sox2-null (if stub exists) — check lethality staging claim

[2026-05-05] verify | oct4

molecules/proteins/oct4.md

• page: molecules/proteins/oct4.md
• sources verified against PDF: Nichols 1998 (10.1016/s0092-8674(00)81769-9, local PDF, 13 pp, read end-to-end); Niwa 2000 (10.1038/74199, local PDF, 5 pp, read end-to-end); Takahashi 2006 (10.1016/j.cell.2006.07.024, downloaded during this pass via camoufox, 14 pp, read end-to-end)
• sources spot-checked for oct4.md-specific claims: Ocampo 2016, Lu 2020, Yang 2023 (full verification recorded in partial-reprogramming.md)
• canonical-DB check: UniProt Q01860 PTM annotations verified via REST API
• corrections made (6):
  1. CRITICAL — Niwa 2000 dosage bullet: “~150% increase” → “>50% increase above normal (>1.5× normal)” — threshold is ±50% of diploid level, not 150% above it
  2. Niwa 2000 footnote: added ZHTc6/ZHBTc4 system description and ±50% threshold language
  3. Takahashi 2006: added “similar, but not identical to ES cells” qualifier; noted Oct3/4 promoter partial methylation; removed unsourced POU-family replacement claim (not in this paper) with unsourced tag
  4. Takahashi 2006 footnote: updated to PDF-available, correct cell types/strain, Nanog-dispensable finding
  5. Nichols 1998 footnote: added hybrid strain background (129×CBA and 129×MF1), more precise n (89 pups), trophoblast giant cell clarification
  6. Frontmatter: removed loss-of-proteostasis from hallmarks (unsupported by body); corrected key-ptms — Lys123-ubiquitination → WWP2-mediated-ubiquitination (UniProt does not annotate K123 as ubiquitination site; K123 is confirmed sumoylation site only)
• downstream pages to check: nichols-1998-oct4-pluripotency, niwa-2000-oct4-dosage, takahashi-2006-ips-oskm, partial-reprogramming, information-theory-of-aging

[2026-05-05] verify | egcg

molecules/compounds/egcg.md

• page: molecules/compounds/egcg.md
• sources verified against full PDF: EFSA 2018 (10.2903/j.efsa.2018.5239, gold OA PMC, 89 pp. read)
• sources verified against abstract only (closed-access): Van Aller 2011 (10.1016/j.bbrc.2011.02.010, Europe PMC abstract); Ullmann 2003 (10.1177/147323000303100205, Europe PMC abstract); Nakagawa 1997 (PMID 9177723, PubMed abstract); Kuriyama 2006 (10.1001/jama.296.10.1255, Europe PMC abstract)
• sources not verified: Gu 2014, Ye 2015, Yu 2017, Naumovski 2015, James 2018, Acosta 2023, Brown 2009, Mielgo-Ayuso 2014, Basu 2010, Pallauf 2017 (all closed-access or not priority; recommend next lint pass)
• corrections made (6):
  1. Nakagawa 1997 DOI: 10.1006/abio.1997.2122 (Ras protein paper — completely wrong) → 10.1006/abio.1997.2098 (confirmed correct via PMID 9177723); bioavailability value 0.32% confirmed correct
  2. EFSA hepatotoxicity threshold in body text: “≥800 mg/day total catechins” → “≥800 mg/day EGCG” (EFSA conclusion explicitly specifies EGCG, not total catechins)
  3. EFSA language: “possible risk of liver injury” → accurate quotation of EFSA conclusion: “statistically significant increase of serum transaminases” (the paper does not use “possible risk” language — its conclusion is stronger: evidence-based finding from interventional trials)
  4. Risk threshold bullet: “≥800 mg/day total catechins (approximately 600+ mg/day EGCG equivalent)” → “≥800 mg/day EGCG (as a food supplement); no evidence of hepatotoxicity below 800 mg EGCG/day in clinical studies up to 12 months”
  5. Ullmann 2003 n: n=18 (wrong) → n=60 (10 per group across 6 dose levels; 8 active + 2 placebo per group); sex clarified as male volunteers
  6. Ullmann 2003 Tmax: “~1.3–2.5 h” → “1.3–2.2 h” (per abstract: “1.3–2.2 h”)
• confirmed correct: Van Aller Ki values (PI3K=380 nM, mTOR=320 nM); Kuriyama n=40,530, age 40–79; all-cause and CVD mortality inverse association; cancer non-significant; EFSA panel identity; EFSA abstract and conclusion language; Nakagawa 0.32% bioavailability; Ullmann Cmax/AUC/t½z ranges
• additional correction: Kuriyama HR attribution — wiki had “HR ~0.77 for CVD mortality”; corrected to specify HR 0.77 (95% CI 0.67–0.89) = all-cause mortality in women; CVD mortality in women = HR 0.69 (95% CI 0.53–0.90); follow-up clarified as 11 yr all-cause / 7 yr cause-specific
• downstream pages to check: studies/nakagawa-1997-egcg-plasma-hplc (wikilink exists on egcg.md; study page filename should be confirmed correct); studies/efsa-2018-green-tea-catechin-safety; studies/ullmann-2003-egcg-ascending-dose-pk; studies/kuriyama-2006-green-tea-mortality-ohsaki

[2026-05-05] verify | taurine

molecules/compounds/taurine.md

• page: molecules/compounds/taurine.md
• sources verified against PDF: Singh 2023 (10.1126/science.abn9257, gold OA, local PDF read end-to-end); Chen 2021 (10.3389/fphys.2021.700352, gold OA, local PDF read end-to-end)
• sources verified against abstract only: Fernandez 2025 (10.1126/science.adl2116, closed-access — Crossref abstract); da Silva 2014 (10.1139/apnm-2012-0229, closed-access — PubMed abstract)
• canonical-DB identity fields: PubChem CID 1123, InChIKey XOAAWQZATWQOTB-UHFFFAOYSA-N, MW 125.15 independently re-verified via PubChem REST API; ChEMBL/DrugBank/CAS not independently re-checked
• corrections made (8):
  1. Mouse taurine decline %: “~80%” → “~70%” (132.3 ± 14.2 ng/ml at 4 wk → 40.2 ± 7.1 ng/ml at 56 wk; slope −25.7, p < 2×10⁻¹⁶; Fig. 1A)
  2. NHP taurine decline %: “~60%” → “85% lower in 15-year-old vs 5-year-old monkeys” (per Fig. 1B legend)
  3. Mouse strain: “C57BL/6JN” → “C57BL/6J” (paper consistently uses C57BL/6J throughout; JN substrain was erroneous)
  4. Mouse cohort n’s added: n=62 vehicle / n=60 taurine females; n=64 vehicle / n=60 taurine males (from Fig. 1D, E legends); caveat “exact n’s not yet verified” removed
  5. Mechanism: “telomere shortening” → “protection against telomerase deficiency” (paper tests zebrafish tert−/− model, not WT telomere shortening reversal)
  6. Competing interest: “one investigator disclosed a Columbia University patent” → “Columbia University has filed provisional patent applications (plural) on which V.K.Y. is listed as an inventor” (per p.11 of PDF)
  7. Chen 2021 dose range: “approximately 1–6 g/day” → “0.05–6 g/day across studies” (0.05 g is the strength-exercise dose per da Silva 2014; 1–6 g is the aerobic range only). Also clarified: 10 studies included; systematic review only, no pooled meta-analysis; majority of studies high-risk-of-bias.
  8. Banner updated: removed “PDF not yet read / download pending” language; replaced with verified-source summary and retained contested-claim warning
• confirmed correct: human taurine decline >80% (Fig. 1C, p < 2.2×10⁻¹⁶); 14-mo age start; 1000 mg/kg/day dose; 10–12% lifespan extension (both sexes, life expectancy at 28 mo +18–25%); health-span endpoints (body weight, energy expenditure, neuromuscular, grip, glucose, bone, immune — all confirmed from Fig. 2); NHP dose 250 mg/kg/day for 6 mo; da Silva 2014 n=21 (10+11); Fernandez 2025 authors list (confirmed via Crossref); Fernandez 2025 key finding (taurine increased or unchanged with age — confirmed from abstract); EPIC-Norfolk n=11,966; contradictory-evidence framing preserved and accurate
• downstream pages to check: singh-2023-taurine-deficiency-aging, chen-2021-taurine-exercise-dose-response, hallmarks-of-aging, cellular-senescence, mitochondrial-dysfunction

[2026-05-04] verify | round-10d

cgas-sting

• page: pathways/cgas-sting.md
• sources verified against PDF: Sun 2013 (10.1126/science.1232458, downloaded); Ishikawa 2008 (10.1038/nature07317, downloaded); Glück 2017 (10.1038/ncb3586, downloaded); Yang 2017 (10.1073/pnas.1705499114, downloaded via camoufox during this pass — was previously “pending”)
• sources unverifiable: Wu 2013 (10.1126/science.1229963, not_oa; claims about 2’3’-cGAMP non-canonical linkage retained with no-fulltext-access)
• corrections made (8):
  1. KEGG ID: null → hsa04623 (Cytosolic DNA-sensing pathway; confirmed via KEGG)
  2. STING transmembrane topology: “four-pass” → “five-pass” — Ishikawa 2008 explicitly states “five putative transmembrane regions” (human STING, amino acids 1–230 containing the five TM motifs)
  3. Ishikawa 2008 primary virus model: footnote said “STING-/- mice fail to clear HSV-1” — corrected to primary demonstration being VSV-GFP susceptibility; HSV-1 also tested but VSV is the principal in vivo model in the paper
  4. Sun 2013 cell models: footnote “HEK293T + THP-1 cells” expanded to reflect full model set: L929, HEK293T-STING, Raw264.7, THP1, BMDM, plus purified-protein biochemical reconstitution
  5. cGAS subcellular localization: “normally nuclear-excluded in resting cells” → corrected per Yang 2017: cGAS is cytoplasmic in nondividing cells and enters the nucleus to associate with chromatin during mitosis; the directionality is cytoplasm-to-nucleus during division, not constitutive nuclear exclusion
  6. Glück 2017 cell model: “IMR90 + WI-38 human fibroblasts (OIS and replicative senescence)” → corrected to WI-38 human lung fibroblasts + cGAS KO MEFs; IMR90 is from Dou et al. 2017, a separate companion paper; senescence types include oxidative stress, irradiation (12 Gy), HRasV12, palbociclib — not solely OIS
  7. Yang 2017 aged tissue claim: removed the unsupported claim “cGAS-STING activation is observed in naturally aged tissues” — Yang 2017 demonstrates only senescence induced by etoposide, irradiation (3 Gy X-ray), and spontaneous immortalization in MEFs and BJ human fibroblasts; no aged tissue experiments
  8. Yang 2017 footnote model: “MEFs + aged mouse tissues” → “cGas−/− MEFs (C57BL/6) + cGas−/− BJ human foreskin fibroblasts (TALEN-generated) + B16F10 melanoma cells; no aged tissue data”
• added: STING trafficking citation gaps flagged (#gap/unsourced for ERGIC/Golgi/Ser366 details — those are from Shang 2019 / Ergun 2019, not Ishikawa 2008)
• confirmed correct: Reactome R-HSA-1834949; cGAS as nucleotidyltransferase; cGAMP from ATP+GTP; STING ER residency; TBK1-IRF3 and TBK1-NF-κB dual arms; CCF/lamin B1 mechanism in Glück 2017; lung adenocarcinoma survival data in Yang 2017 (log-rank p=5e-05, HR=1.82 [1.4–2.5], n=451 low vs n=222 high)
• banner removed; verified: true (partial scope — Wu 2013 closed-access; WikiPathways ID not confirmed)
• downstream pages to check (main agent): cellular-senescence (may cite Glück 2017 using IMR90 or the wrong senescence-type list); tbk1 (verify STING trafficking phosphorylation site Ser366 attribution — that page may cite Ishikawa 2008 for details that actually come from later structural papers); chronic-inflammation / inflammaging (may repeat the unsupported “cGAS-STING active in aged tissues” claim attributed to Yang 2017)

nampt

• page: molecules/proteins/nampt.md
• sources verified against PDF: Yoshida 2019 (10.1016/j.cmet.2019.05.015, downloaded); Yoshino/Baur/Imai 2018 review (10.1016/j.cmet.2017.11.002, downloaded)
• sources unverified (still pending download): Revollo 2004 (10.1074/jbc.M408388200), Rongvaux 2002 (10.1002/1521-4141(200211)32:11<3225::AID-IMMU3225>3.0.CO;2-L), Rongvaux 2008 (10.4049/jimmunol.181.7.4685), Khaidizar 2017 (10.1111/gtc.12542), Imai & Guarente 2014 (10.1016/j.tcb.2014.04.002); Fukuhara 2005 not_oa (retraction paper, claims preserved correctly)
• corrections made (7):
  1. eNAMPT localization: “circulates in plasma; also transported in extracellular vesicles” → corrected to “exclusively within EVs in both mice and humans — not detectable free in plasma” (Yoshida 2019 key finding; plasma proteinase K digestion + sucrose gradient experiments confirmed EV-exclusive localization)
  2. eNAMPT lifespan claim conflation: wiki conflated ANKI genetic model with EV pharmacological injection into a single statement. Split into two clearly labelled experiments with precise numbers.
  3. ANKI lifespan n: “n=~20/group (approximate; verify)” → ANKI female control n=39, ANKI n=40; ANKI male control n=39, ANKI n=39
  4. ANKI lifespan extension: imprecise → median 786 vs 693 days, 13.4% extension, p=0.014 (Gehan-Breslow-Wilcoxon χ²=6.043, df=1); males: no effect p=0.56
  5. EV injection lifespan: not previously stated → median 926 vs 840 days, 10.2% extension, p=0.0009 (χ²=11.10, df=1); n=11–12; treatment started at 26 months; EVs from 4–12-month-old C57BL/6J mice
  6. NAD+ decline “~50%” claim: body text corrected to acknowledge tissue-specificity and wide range (~30–90%); lede updated; Yoshino 2018 Table 3 specific values cited
  7. Pharmacological activators section: “recombinant eNAMPT protein” → “EV-contained eNAMPT” with explicit note that free recombinant protein is NOT internalized and does not enhance NAD+ (Yoshida 2019 Figure 7B directly shows this)
• confirmed correct: UniProt P43490; 491 aa canonical; NCBI Gene 10135; HGNC 30092; obligate homodimer; iNAMPT/eNAMPT distinction; SIRT1-dependent K53 deacetylation predisposes secretion (cited from Yoshida 2019 Discussion, citing Yoon et al. 2015); eNAMPT plasma decline ~33% female / ~74% male (6→18 months); human linear decline ages 37–80 R²=0.6212, p=0.0014, n=13; sex difference (females > males for lifespan extension) documented and noted
• banner removed; verified: true (partial scope — Revollo/Rongvaux/Khaidizar/Imai & Guarente claims not verified against PDF)
• downstream pages to check (main agent): none identified as actively citing nampt.md-specific quantitative claims yet; nmn and nr pages cite NAMPT biology generally but don’t appear to replicate the corrected numbers; sirt1 may mention NAMPT upstream supply — check if it uses the “~50%” decline figure

trem2

• page: molecules/proteins/trem2.md
• sources verified: Guerreiro 2013 (10.1056/NEJMoa1211851, local PDF); Hammond 2019 (10.1016/j.immuni.2018.11.004, local PDF)
• sources unverifiable: Jonsson 2013 (10.1056/NEJMoa1211103, download failed — green OA handle redirects to HTML only; re-attempt recommended); Paloneva 2002 (10.1086/342259, pending); Wang 2015 (10.1016/j.cell.2015.01.049, pending); Keren-Shaul 2017 (10.1016/j.cell.2017.05.018, download failed)
• corrections made (4):
  1. Guerreiro 2013 R47H OR description: removed fabricated “initial discovery OR=4.5, 95% CI 1.7–11.9” figure (not stated in the paper for R47H alone in the discovery set); corrected to combined series OR=5.05 (95% CI 2.77–9.16; p=9.0×10⁻⁹) in 1,994 cases + 4,062 controls
  2. Guerreiro 2013 footnote n: “~25,000 (combined discovery + replication)” → “8,147 total (1,092 cases + 1,107 controls discovery; plus 1,887 cases + 4,061 controls additional direct genotyping)” — the ~25,000 figure has no basis in the paper
  3. Hammond 2019 aged-brain TREM2 claim: “TREM2-high microglial cluster (corresponding to DAM) is expanded in aged mouse brains” → corrected to the paper’s actual finding: Ccl4-enriched (OA2) and interferon-responsive (OA3) clusters expand in aged C57BL/6J brains; Trem2 is a broadly expressed canonical marker uniformly expressed across clusters, not selectively elevated in the aged clusters; aged clusters partially overlap with DAM (shared: Spp1, Lpl, Apoe) but are distinct
  4. Hammond 2019 footnote: age range “E14.5 through 12 months” → “E14.5 through P540 (18 months)”; n corrected to “41 mice; 76,149 microglia sequenced”; strain corrected to “C57BL/6J”
• confirmed correct: UniProt Q9NZC2; protein length 230 aa (including 19-aa signal peptide); NCBI Gene 54209; HGNC 17761; Guerreiro 2013 p-value 9.0×10⁻⁹; Jonsson 2013 OR=2.92 (95% CI 2.09–4.09; p=3.42×10⁻¹⁰) — not independently verified vs PDF but retained from seeder (Jonsson PDF unavailable); NHD mutation T66M mention is present in Guerreiro paper but as one of several FTD-like variants; Q33X is the Guerreiro-confirmed NHD variant; wiki correctly attributes NHD to “biallelic TREM2 mutations” generally; DAM two-stage model (Keren-Shaul 2017) claims retained unverified
• banner removed; verified: true (partial scope — see verified-scope field)
• downstream pages to check (main agent): alzheimers-disease (may cite OR=4.5 discovery figure from Guerreiro or n=~25,000); microglia (may repeat TREM2-high/DAM-expansion claim attributed to Hammond 2019); guerreiro-2013-trem2-ad (study page n and OR fields need correction if they match old values); hammond-2019-microglia-aging (study page should reflect aged clusters are Ccl4+/IFN-responsive, not TREM2-high)

lipa

• page: molecules/proteins/lipa.md
• sources verified against live databases: UniProt P38571 (REST API — confirmed signal peptide aa 1–27; propeptide aa 28–76; mature chain aa 77–399; calculated MW ~45.4 kDa; 6 N-glycosylation sites N36/N72/N101/N161/N273/N321); Burton 2015 trial framing checked against PubMed abstract PMID 26352813
• sources unverifiable (archive pending; task instructions prohibit download triggering): Anderson 1991 (10.1016/s0021-9258(18)54597-x, hybrid OA pending); Burton 2015 full PDF (10.1056/NEJMoa1501365, bronze OA pending); Morris 2017 (10.1161/atvbaha.116.308734, pending); Evans 2019 (10.1161/atvbaha.119.313443, pending)
• corrections made (4):
  1. Signal peptide positions: “aa 1–21” → “aa 1–27” (UniProt P38571, evidence PMID 8112342)
  2. Propeptide: previously absent → added “propeptide aa 28–76 removed during lysosomal maturation; mature chain begins aa 77” (UniProt P38571, PMID 8112342 and 15269241)
  3. Precursor MW: “~49 kDa before post-translational processing” → “~45.4 kDa calculated MW” (UniProt P38571)
  4. Burton 2015 footnote enriched: added n=66 (36 sebelipase, 30 placebo), primary result (ALT normalization 31% vs 7%, p=0.03), hepatic fat reduction (p<0.001), “double-blind” qualifier; verified framing against PubMed abstract PMID 26352813
• confirmed correct: UniProt P38571 accession; 399 aa precursor length; NCBI Gene 3988; HGNC 6617; LAL-D spectrum (Wolman = complete loss; CESD = partial); FDA approval December 2015; Singh 2009 lipophagy axis (trusted from lipophagy.md verification); rs1051338 CAD variant (Morris 2017 citation DOI confirmed in archive)
• banner removed; verified: true (partial scope — Anderson 1991, Morris 2017, Evans 2019 full-text not independently verified)
• downstream pages to check (main agent): lipophagy (check whether mature chain start or signal peptide position appears in any claim there — unlikely but confirm); chaperone-mediated-autophagy (propeptide note does not affect any claim there)

tau

• page: molecules/proteins/tau.md
• sources verified against PDF: Hutton 1998 (10.1038/31508, local PDF — confirmed available)
• sources verified against live databases: UniProt P10636 (REST API + FASTA), NCBI Gene 4137 (web)
• sources unverifiable (not_oa): Lee 2001 (10.1146/annurev.neuro.24.1.1121) — tauopathy taxonomy, kinase sites, Mapt KO phenotype, biomarker claims retained unverified; Goedert 1989 (10.1093/emboj/8.2.393, not in archive) — 6-isoform claim retained unverified; Mukrasch 2009 (10.1371/journal.pbio.0070034, not in archive) — NMR/IDP claims retained unverified
• critical bad citation found: Cuervo 2004 footnote DOI 10.1038/nn1387 resolves to “Coactivation and timing-dependent integration of synaptic potentiation and depression” (Nat Neurosci 8:187-193, 2005) — completely unrelated paper; confirmed via Crossref API; tau/CMA claim flagged no-fulltext-access; footnote annotated with full warning
• corrections made (4):
  1. Splice-site mutations: “+14, +16 in intron 10” → “+13, +14, and +16 in intron 10” — Hutton 1998 explicitly identifies three intronic mutations (Table 1 and text confirm +13 in DDPAC/Man19 family clusters, +14 in Aus1/FTD002/Man6/Man23/FTD004, +16 in Aus1/FTD002/Man6/Man23/FTD004)
  2. Gene size: “~143 kb” → “~134 kb” (NCBI Gene 4137 GRCh38.p14: 45,894,554–46,028,334 = 133,780 bp; ~134 kb)
  3. Isoform section: expanded to clarify UniProt lists 9 total isoforms (including PNS-tau at 758 aa as canonical); wiki table correctly covers the 6 CNS-expressed isoforms; added explicit note that PNS-tau (758 aa) is the UniProt canonical/longest; table column header “aa (canonical)” → “aa”; table notes updated to “largest/smallest CNS isoform”
  4. Cuervo 2004 footnote: replaced misleading citation with full warning about wrong DOI and no-fulltext-access marker
• confirmed correct: UniProt P10636 accession; 16 exons (NCBI Gene); 17q21.31 chromosomal location; missense mutations G272V (exon 9), P301L (exon 10), R406W (exon 13) per Hutton 1998; FTDP-17 autosomal dominant inheritance; P301L only affects 4R isoforms (exon 10 present); splice-site mutations increase exon 10 inclusion (4R excess); HFTD1 family (Netherlands) is the largest reported with 49 affecteds; Nature 393:702-705 journal/volume/pages correct; cited_by ~3,484 confirmed
• banner removed; verified: true (partial scope — see verified-scope field)
• downstream pages to check (main agent): chaperone-mediated-autophagy (may cite the wrong Cuervo 2004 DOI); hutton-1998-mapt-ftdp17 (study page — check if splice mutations are listed correctly as +13/+14/+16, not +14/+16); any page citing tau/CMA without a verified DOI

partial-reprogramming

• page: processes/partial-reprogramming.md
• sources verified against PDF: Ocampo 2016 (10.1016/j.cell.2016.11.052, local PDF); Lu 2020 (10.1038/s41586-020-2975-4, local PDF); Yang 2023 (10.1016/j.cell.2022.12.027, local PDF)
• sources not re-verified: Schooling 2025 (10.1186/s40246-025-00852-4, PDF available; MR n and clock details trusted from prior information-theory-of-aging pass); Takahashi 2006 (10.1016/j.cell.2006.07.024, pending download — background-only citation, no load-bearing quantitative claims on this page)
• corrections made (1):
  1. Ocampo 2016 WT aged context: “improved recovery from metabolic disease (high-fat diet)” → “improved recovery from metabolic disease (streptozotocin-induced beta-cell ablation; glucose tolerance improved at 2 weeks post-injury)” — the paper uses low-dose STZ to ablate beta cells in 12-month-old WT 4F mice, not a high-fat diet model; muscle injury correctly described as cardiotoxin (CTX) to tibialis anterior
• confirmed correct: Ocampo 2016 LAKI model (G609G knockin Lmna); cyclic protocol exactly 2d doxycycline on / 5d off; no teratomas in cyclic protocol; Lu 2020 AAV2 intravitreal delivery to RGCs; TET1/TET2 required — TET3 not required (paper confirms “Tet1 and Tet2, but not Tet3, were upregulated by OSK”); Yang 2023 ICE system uses I-PpoI under ER^T2/tamoxifen control; C57BL6/J background; DSBs induced 4–6 months; epigenetic aging ~50% faster (p<0.0001); OSK reversal up to 57% of DNAme age; 16-month post-treatment assessment confirmed
• banner removed; verified: true (partial scope — see verified-scope field)
• downstream pages to check (main agent): information-theory-of-aging (cross-check WT aging model description if it reproduces the high-fat-diet claim); fisetin (no partial-reprogramming claims expected — no action needed); footnote in [^ocampo2016] on this page is correct (no high-fat-diet wording there)

cardiac-fibroblasts

• page: cell-types/cardiac-fibroblasts.md
• sources verified against PDF: Acharya 2012 (10.1242/dev.079970 — downloaded this pass via Europe PMC)
• sources unverifiable against PDF: Pinto 2016 (10.1161/CIRCRESAHA.115.307778 — green OA listed but DOI lookup failed, 0 candidate URLs); Tallquist 2017 (10.1038/nrcardio.2017.57 — same failure)
• sources trusted from prior verification: Anderson 2019 and Lewis-McDougall 2019 (previously verified on cardiomyocytes, per task instructions)
• canonical identity: CL:0002548 confirmed as “Fibroblast of cardiac tissue” via OLS4 API
• corrections made (3):
  1. Footnote [^acharya2012] genetic tool: “Tcf21-MerCreMer + Rosa26R” → “Tcf21^iCre (tamoxifen-inducible Cre knocked into the Tcf21 locus) + R26R^YFP/R26R^tdT” — MerCreMer is a different construct; paper uses a Tcf21-locus knock-in of iCre; R26R^tdT and R26R^YFP are the actual reporters used
  2. Footnote [^acharya2012] page numbers: “Development 139(13):2359–2369” → “Development 139(12):2139–2149” — confirmed from PDF masthead
  3. Body text, Developmental origin section: “epicardial-mesothelial transition (EMT)” → “epicardial epithelial-to-mesenchymal transition (EMT)”; also removed unsupported “majority of adult resident CFs” claim — paper shows Tcf21-lineage cells are CF-restricted but does not quantify their fraction of all CFs; ~6% of total cardiac single-cell suspension lineage-traced, enriched for CF marker genes
• footnotes updated: Pinto 2016 and Tallquist 2017 archive status changed from “pending download” to “download failed — no-fulltext-access”
• banner: replaced ⚠️ auto-extraction banner with partial-verification notice; verified: true (partial scope)
• downstream pages to check (main agent): none identified — cardiomyocytes holds Anderson/Lewis-McDougall claims and was already verified

[2026-05-04] ingest | cardiac-fibroblasts (round-10d)

• added: cell-types/cardiac-fibroblasts.md (type: cell-type, verified: false)
• entity: cardiac fibroblasts; Cell Ontology CL:0002548; no UniProt/PubChem applicable
• inbound links satisfied: cardiomyocytes (verified-partial), cardiac-fibrosis (verified-partial), heart-failure (verified-partial), myocardium (R10c tissue stub), cellular-senescence (verified hallmark), plus cardiovascular-aging (drafted)
• canonical DOIs confirmed in archive:
  • doi:10.1242/dev.079970 — Acharya 2012 (Tcf21 lineage tracing) — green OA (PMC) — pending download — 484 cites
  • doi:10.1161/CIRCRESAHA.115.307778 — Pinto 2016 (cell composition stereology) — green OA — pending download — 1,360 cites
  • doi:10.1038/nrcardio.2017.57 — Tallquist & Molkentin 2017 (CF identity review) — green OA (PMC) — pending download — 540 cites
  • doi:10.15252/embj.2018100492 — Anderson 2019 — local PDF available (previously verified on cardiomyocytes.md)
  • doi:10.1111/acel.12931 — Lewis-McDougall 2019 — local PDF available (previously verified on cardiomyocytes.md)
• DOI corrections (two brief-supplied DOIs were BUG-2 mismatches):
  • brief doi:10.1242/dev.072082 → not found in archive; correct Acharya 2012 DOI is 10.1242/dev.079970 (confirmed via Crossref + DOI lookup, title confirmed “The bHLH transcription factor Tcf21…”)
  • brief doi:10.1038/nrc.2017.50 → resolves to “Expanding the use of PARP inhibitors” (oncology; 4 cites) — wrong paper entirely; correct Tallquist 2017 DOI is 10.1038/nrcardio.2017.57 (confirmed via PubMed PMID 28436487 + DOI lookup, title “Redefining the identity of cardiac fibroblasts”)
• gaps surfaced: needs-human-replication (CM SASP → CF activation in human; senolytic cardiac trials; CF-to-CM ratio in human); needs-replication (CF-intrinsic senescence characterization; CPC SASP → CF data is conditioned-medium in vitro); contradictory-evidence (cardiomyocyte–fibroblast gap junction functional coupling); no-mechanism (SGLT2i anti-fibrotic); unsourced (CF-derived IGF-1/FGF-2 paracrine quantification; CF-intrinsic senescence)
• implicit stubs created: tgf-beta (new wikilink not previously registered)
• ROADMAP updated: cardiac-fibroblasts row added under Cell types section; myocardium stub reference to “planned R10d” removed

[2026-05-04] ingest | ep300 (round-10d)

• added: molecules/proteins/ep300.md (type: protein, verified: false)
• entity: EP300 / E1A Binding Protein p300 / KAT3B; UniProt Q09472; NCBI Gene 2033; HGNC 3373
• inbound links satisfied: spermidine (verified-partial — EP300-specific HAT inhibition mechanism per Pietrocola 2015 verified; corrected from HAT-broad to EP300-selective), autophagy (verified-partial — EP300 acetylates ATG proteins to inhibit autophagy), p53 (verified — EP300/CBP acetylates p53 K382), cbp-p300 (planned/aliased paralog page)
• canonical DOIs confirmed in archive (5 footnotes):
  • doi:10.1101/gad.8.8.869 — Eckner 1994, EP300 cloning (Genes Dev; 1,047 cites; diamond OA; download pending)
  • doi:10.1016/s0092-8674(00)81165-4 — Yao 1998, Ep300 KO mouse embryonic lethal (Cell; 1,001 cites; 100th percentile; local PDF available) — DOI CORRECTED from brief-supplied 10.1016/s0092-8674(00)80432-x (not found in archive)
  • doi:10.1074/jbc.M807135200 — Lee & Finkel 2009, EP300 acetylates ATG proteins / autophagy regulation (JBC; 261 cites; 100th percentile; download pending)
  • doi:10.1038/cdd.2014.215 — Pietrocola 2015, spermidine inhibits EP300 acetyltransferase (Cell Death Differ; 300 cites; FWCI 9.5; 100th percentile; local PDF available — previously verified on spermidine.md)
  • doi:10.1086/429130 — Roelfsema 2005, EP300 mutations in Rubinstein-Taybi syndrome 2 (AJHG; 467 cites; 100th percentile; download pending) — DOI CORRECTED from brief-supplied 10.1086/430408 (archive matched astrophysics paper — title mismatch caught by BUG-2)
• DOI corrections this round:
  • Brief doi:10.1016/s0092-8674(00)80432-x → correct doi:10.1016/s0092-8674(00)81165-4 (Yao 1998; PMID 9590171; confirmed via PubMed efetch)
  • Brief doi:10.1086/430408 → correct doi:10.1086/429130 (Roelfsema 2005; PMID 15706485; archive title for 430408 was astrophysics paper)
• implicit stubs created: histone-acetylation (process page), caloric-restriction-mimetic (intervention/concept page), cbp-p300 (paralog complex page — planned/aliased)
• gaps surfaced:
  • needs-human-replication — EP300 inhibition → autophagy in aged human tissue; longevity endpoint undemonstrated
  • contradictory-evidence — EP300 vs CBP individual contributions at specific loci/organisms
  • long-term-unknown — systemic EP300 inhibitor safety in healthy older adults
  • no-mechanism — EP300 occupancy shifts in human aging tissue (causal vs correlative)
  • needs-replication — CR-mimetic claim for EP300 inhibitors beyond autophagy endpoint
  • unsourced — H3K27ac EP300 vs CBP partition at any given enhancer; aging-specific EP300 ChIP-seq datasets

[2026-05-04] ingest | hsc70 (round-10d)

• added: molecules/proteins/hsc70.md (type: protein, verified: false)
• entity: HSC70 / HSPA8 / Heat shock cognate 71 kDa protein; UniProt P11142; NCBI Gene 3312; HGNC 5241
• inbound links satisfied: chaperone-mediated-autophagy (verified-partial), loss-of-proteostasis (verified hallmark), alpha-synuclein (R10d parallel)
• canonical DOIs confirmed in archive (none locally downloaded):
  • doi:10.1083/jcb.99.2.723 — Schlossman 1984 — 442 cites; bronze OA pending
  • doi:10.1126/science.2799391 — Chiang 1989 — 959 cites; not_oa
  • doi:10.1126/science.273.5274.501 — Cuervo 1996 — 937 cites; not_oa
  • doi:10.1126/science.1101738 — Cuervo 2004 — 1,966 cites; not_oa
  • doi:10.1007/s00018-004-4464-6 — Mayer & Bukau 2005 — 2,811 cites; bronze OA pending
• DOI corrections: task brief 10.1007/s00018-004-4464-1 not in archive → correct 10.1007/s00018-004-4464-6 (PMID 15770419); task brief 10.1126/science.1197754 is BUG-2 mismatch → correct Cuervo 2004 is 10.1126/science.1101738 (PMID 15333840)
• gaps surfaced: needs-human-replication; contradictory-evidence (HSC70 preservation vs decline); needs-replication (non-rate-limiting claim; tau CMA); no-mechanism (co-chaperone aging shifts); dose-response-unclear; GenAge: no HSPA8 entry confirmed
• implicit stubs: none (all wikilinks to existing pages); lamp-2a standalone protein page flagged as future seeding candidate

[2026-05-04] ingest | lipa (round-10d)

• added: molecules/proteins/lipa.md (type: protein, verified: false)
• entity: LIPA / Lysosomal Acid Lipase; UniProt P38571; NCBI Gene 3988; HGNC 6617
• inbound links satisfied: lipophagy (verified-partial — Singh 2009 axis confirmed), chaperone-mediated-autophagy (drafted — PLIN degradation prerequisite cross-link)
• canonical DOIs confirmed in archive (all pending download — no local PDFs):
  • doi:10.1016/s0021-9258(18)54597-x — Anderson 1991, LIPA cDNA cloning (JBC; 145 cites) — DOI lookup confirmed; NOT the DOI supplied in task brief (10.1042/bj2740731 — resolves to fibronectin paper, BUG-2 mismatch)
  • doi:10.1038/nature07976 — Singh 2009, lipophagy (Nature; 3,816 cites; local PDF available — previously verified on lipophagy.md)
  • doi:10.1056/NEJMoa1501365 — Burton 2015, sebelipase alfa Phase 3 RCT (NEJM; 238 cites) — confirmed; corrected from brief-supplied 10.1056/NEJMoa1407972 (not in archive)
  • doi:10.1161/atvbaha.116.308734 — Morris 2017, rs1051338 functional characterization (ATVB; 38 cites)
  • doi:10.1161/atvbaha.119.313443 — Evans 2019, CAD-associated LIPA variant characterization (ATVB; 24 cites)
• DOI corrections (two brief-supplied DOIs had archive title-mismatches):
  • brief doi:10.1042/bj2740731 → resolves to “Human plasma fibronectin” (fibronectin paper, BUG-2); correct Anderson 1991 LIPA cDNA DOI is 10.1016/s0021-9258(18)54597-x (confirmed via Crossref + archive)
  • brief doi:10.1136/jmg.6.4.448 → resolves to insect-virus microbiology journal issue (BUG-2); correct Patrick & Lake Wolman disease DOI could not be confirmed — flagged unsourced with manual verification note on page
  • brief doi:10.1056/NEJMoa1407972 → not in archive; correct Burton Phase 3 DOI confirmed as 10.1056/NEJMoa1501365 via PubMed PMID 26352813
• GenAge: LIPA not listed in GenAge human genes database (no genage-id)
• gaps surfaced: needs-human-replication (age-associated LIPA decline in human hepatocytes; lysosomal pH effect on LIPA activity in aged cells); no-mechanism (relative contribution of autophagy flux decline vs pH rise to LIPA substrate deprivation); long-term-unknown (sebelipase alfa in adult CESD with established fibrosis); unsourced (Patrick & Lake DOI unconfirmed); needs-canonical-id (GenAge absent)
• implicit stubs created: none (all cross-linked pages already exist)

[2026-05-04] ingest | partial-reprogramming (Round 10d)

• added: processes/partial-reprogramming.md (type: process, verified: false)
• entity: partial reprogramming — transient OSK/OSKM Yamanaka-factor-driven epigenetic rejuvenation
• central process page anchoring information-theory-of-aging (verified); 4 inbound cross-references satisfied (information-theory-of-aging, epigenetic-alterations, homo-sapiens, caloric-restriction)
• Mode A discipline applied: safety and single-lab caveats are load-bearing; banner note instructs editors not to soften them
• canonical DOIs confirmed via DOI lookup:
  • doi:10.1016/j.cell.2006.07.024 — Takahashi 2006 — 26,108 citations (100th percentile) — local PDF: pending (bronze OA)
  • doi:10.1016/j.cell.2016.11.052 — Ocampo 2016 — 903 citations (100th percentile) — local PDF: available
  • doi:10.1038/s41586-020-2975-4 — Lu 2020 — 771 citations (100th percentile) — local PDF: available
  • doi:10.1016/j.cell.2022.12.027 — Yang 2023 — 560 citations (100th percentile) — local PDF: available
  • doi:10.1186/s40246-025-00852-4 — Schooling 2025 — 0 citations (preliminary) — local PDF: available
• implicit stubs created: oct4, sox2, klf4, c-myc
• gaps surfaced: needs-human-replication (all evidence mouse-only); needs-replication (single-lab OSK claims); long-term-unknown (no lifespan data in normally aged mice; no human safety data); dose-response-unclear (teratoma safety window); contradictory-evidence (Schooling 2025 MR null — preliminary)
• schema note: pathways: null used for process page with no canonical signaling pathway (OSK acts via TF-driven epigenome remodeling, not a defined pathway node); null is valid YAML; no escalation required
• verification priority: (1) Ocampo 2016 cyclic dosing protocol + teratoma-absence claim; (2) Lu 2020 TET1/TET2 vs TET3 requirement; (3) Yang 2023 ICE ~50% clock acceleration and ~57% reversal figures — all three have local PDFs

[2026-05-04] ingest | deptor (round-10d)

• added: molecules/proteins/deptor.md (type: protein, verified: false)
• entity: DEPTOR / DEP domain-containing mTOR-interacting protein; UniProt Q8TB45; NCBI Gene 64798; HGNC 24784; 409 aa
• inbound links satisfied: mtor (verified-partial), raptor (verified-full), rictor (verified-partial), fkbp12 (verified-partial), hyperfunction-theory (verified)
• canonical DOIs confirmed in archive (none locally downloaded — all pending):
  • doi:10.1016/j.cell.2009.03.046 — Peterson 2009 — DEPTOR discovery — bronze OA pending — 1,182 citations (100th percentile, impact 0.91)
  • doi:10.1016/j.molcel.2011.08.029 — Gao 2011 — SCF^β-TrCP + RSK1/S6K1 degron — bronze OA pending — 267 citations (100th percentile)
  • doi:10.1016/j.molcel.2011.09.005 — Zhao 2011 — CK1α + mTOR auto-amplification loop — bronze OA pending — 192 citations (100th percentile)
  • doi:10.18632/aging.203959 — Miller 2022 — GHR-KO mTOR complex proteomics — hybrid OA pending — 3 citations
• DOI corrections (two brief-supplied DOIs had archive title-mismatches):
  • brief doi:10.1016/j.molcel.2010.12.030 → archive title “The Structural Basis for Tight Control of PP2A Methylation” (wrong paper); correct Gao 2011 DOI = 10.1016/j.molcel.2011.08.029 (confirmed archive title matches)
  • brief doi:10.1016/j.molcel.2010.12.024 → not verified separately; corrected by analogy and confirmed as Zhao 2011 (doi:10.1016/j.molcel.2011.09.005) via PubMed efetch PMID 22017877
• gaps surfaced:
  • unsourced — DEPTOR expression trajectory with normal mammalian aging not identified in literature search
  • needs-human-replication — all mechanistic data from cell lines / mouse models; no human aging phenotype data
  • contradictory-evidence — GHR-KO long-lived mice show REDUCED DEPTOR (Miller 2022); simple model predicts elevated DEPTOR in long-lived animals
  • long-term-unknown — Deptor-/- aging phenotype unpublished; overexpression lifespan study not published
  • needs-replication — multiple myeloma IRS-1 disinhibition mechanism (Peterson 2009); USP7/OTUB1 deubiquitination functional significance
• implicit stubs created: none (all linked pages already seeded: raptor, rictor, fkbp12, s6k1, 4ebp1, akt, sgk1, autophagy, mtor, insulin-igf1, deregulated-nutrient-sensing, disabled-macroautophagy, hyperfunction-theory)
• schema gap: none encountered

[2026-05-04] ingest | nampt (round-10d)

• added: molecules/proteins/nampt.md (type: protein, verified: false)
• entity: NAMPT / Nicotinamide Phosphoribosyltransferase; UniProt P43490; NCBI Gene 10135; HGNC 30092
• inbound links satisfied: sirt1 (verified), sirtuin (verified-partial), nmn (verified-partial), nr (verified-partial), plus ampk pathway cross-link
• canonical DOIs confirmed in archive (none locally downloaded — all pending):
  • doi:10.1074/jbc.M408388200 — Revollo 2004 — hybrid OA pending — 969 citations (100th percentile)
  • doi:10.1016/j.cmet.2017.11.002 — Yoshino 2018 — bronze OA pending — 952 citations (100th percentile)
  • doi:10.1016/j.cmet.2019.05.015 — Yoshida 2019 — OA pending — eNAMPT lifespan extension in mice
  • doi:10.1002/1521-4141(200211)32:11<3225::AID-IMMU3225>3.0.CO;2-L — Rongvaux 2002 — OA status unconfirmed; PMID 12555668
  • doi:10.4049/jimmunol.181.7.4685 — Rongvaux 2008 — OA pending
  • doi:10.1016/j.tcb.2014.04.002 — Imai-Guarente 2014 — OA pending
  • doi:10.1126/science.1097243 — Fukuhara 2005 — not_oa — RETRACTED; documented on page
  • doi:10.1111/gtc.12542 — Khaidizar 2017 — OA pending
• DOI corrections (two brief-supplied DOIs had archive title-mismatches):
  • brief doi:10.1002/eji.200324134 → correct doi:10.1002/1521-4141(200211)32:11<3225::AID-IMMU3225>3.0.CO;2-L (Rongvaux 2002; PMID 12555668); archive confirmed correct title
  • brief doi:10.1126/science.1109389 → correct doi:10.1126/science.1097243 (Fukuhara 2005 visfatin; PMID 15604363); archive confirmed correct title and retraction status
• gaps surfaced: eNAMPT mechanism (#gap/no-mechanism); tissue-specific NAMPT decline (#gap/contradictory-evidence); human clinical NAMPT data (#gap/needs-human-replication); NAMPT activator status (#gap/unsourced); long-term overexpression safety (#gap/long-term-unknown)
• implicit stubs created: nad-salvage (pathway page does not exist), parp1 (protein page not yet seeded), cd38 (protein page not yet seeded)
• visfatin retraction fully documented; retraction DOI confirmed in PubMed (retraction notice Science 318:565)

[2026-05-04] ingest | opa1 (round-10d)

• added: molecules/proteins/opa1.md
• entity type: protein
• inbound references resolved: 4 (from mitofusins R10c verified, cytochrome-c verified-partial, mitochondrial-dysfunction drafted, mitophagy verified-partial)
• canonical IDs confirmed: UniProt O60313 (Swiss-Prot reviewed), NCBI Gene 4976, HGNC 8140
• primary-source DOIs cited (4 footnotes):
  • doi:10.1038/79944 — Alexander 2000, OPA1 cloning / ADOA (Nat Genet; 1,302 cites; 100th percentile; local PDF available)
  • doi:10.1038/79936 — Delettre 2000, OPA1 ADOA / splice isoforms (Nat Genet; 1,412 cites; 100th percentile; local PDF available)
  • doi:10.1016/j.cell.2006.06.025 — Frezza 2006, OPA1 cristae junction / cytochrome c gating (Cell; 1,616 cites; 100th percentile; local PDF available; previously verified on cytochrome-c.md)
  • doi:10.1083/jcb.201308006 — Anand 2014, YME1L + OMA1 cleavage balance (JCB; 755 cites; 100th percentile; not_oa, no local PDF — no-fulltext-access)
• implicit stubs created: oma1, yme1l
• gaps surfaced: no-fulltext-access (Anand 2014); needs-human-replication (OPA1 decline in human aging); no-mechanism (OMA1 sensitivity age-related changes); long-term-unknown (pharmacological OMA1 inhibition)
• GenAge: OPA1 not listed

[2026-05-04] ingest | cgas-sting (round-10d)

• added: pathways/cgas-sting.md (type: pathway, verified: false)
• entity type: pathway
• inbound references resolved: 5 (from chronic-inflammation, tbk1, mitochondrial-dysfunction, cellular-senescence, ROADMAP)
• canonical IDs confirmed:
  • Reactome R-HSA-1834949 (Cytosolic sensors of pathogen-associated DNA) — confirmed via API
  • KEGG: null — no dedicated cGAS-STING KEGG entry found as of 2026-05-04; needs-canonical-id
  • WikiPathways: null — no WP ID confirmed; needs-canonical-id
• primary-source DOIs cited (5 footnotes):
  • doi:10.1126/science.1232458 — Sun 2013, cGAS discovery (Science; 4,556 cites; 100th percentile; local PDF available)
  • doi:10.1038/nature07317 — Ishikawa 2008, STING discovery (Nature; 3,350 cites; 100th percentile; local PDF available)
  • doi:10.1126/science.1229963 — Wu 2013, cGAMP second messenger (Science; 2,374 cites; 100th percentile; archive: not_oa / closed-access)
  • doi:10.1073/pnas.1705499114 — Yang 2017, cGAS in senescence (PNAS; 940 cites; 100th percentile; archive: bronze OA, download pending)
  • doi:10.1038/ncb3586 — Glück 2017, cGAS-CCFs-SASP (Nature Cell Biology; 1,053 cites; 100th percentile; local PDF available)
• implicit stubs created: cgas (planned protein page), sting (planned protein page), irf3 (planned protein page)
• existing verified cross-links used: tbk1 (verified-partial), nf-kb (existing page), chronic-inflammation (drafted), cellular-senescence (verified), mitochondrial-dysfunction (drafted)
• gaps surfaced:
  • needs-canonical-id — KEGG and WikiPathways IDs absent
  • no-fulltext-access — Wu 2013 (closed-access); Yang 2017 (download pending)
  • needs-human-replication — human in vivo aging activation data; interventional data
  • needs-replication — LINE-1/retrotransposon arm in human aging
  • long-term-unknown — safety of STING/cGAS inhibitors in aged hosts

[2026-05-04] ingest | round-10d-targeted-adds

jak-stat-pathway

• added: pathways/jak-stat-pathway.md (type: pathway, verified: false)
• entity: JAK-STAT signaling pathway; KEGG hsa04630; Reactome R-HSA-6785807; WikiPathways null (#gap/needs-canonical-id)
• inbound references resolved: 5 (chronic-inflammation verified, immunosenescence verified-partial, senomorphics verified-partial, growth-hormone verified-partial, ROADMAP)
• primary-source DOIs confirmed in archive (5 footnotes):
  • doi:10.1126/science.8197455 — Darnell 1994, JAK-STAT discovery review (Science; 6,055 cites; 100th percentile; archive: closed, no local PDF)
  • doi:10.1038/nrm909 — Levy 2002, STATs review (NRMolCellBio; 3,151 cites; 100th percentile; local PDF available)
  • doi:10.1073/pnas.1515386112 — Xu 2015, JAKi senomorphic / frailty (PNAS; 839 cites; 100th percentile; local PDF available; previously verified on senomorphics.md)
  • doi:10.1038/nature03546 — James 2005, JAK2 V617F in MPNs (Nature; 3,499 cites; 100th percentile; local PDF available)
  • doi:10.1056/NEJMoa1707914 — Ridker 2017, CANTOS IL-1β/IL-6 axis in CV disease (NEJM; 8,619 cites; 100th percentile; local PDF available; used narrowly for IL-6/inflammation axis validation)
• implicit stubs created: jak1, jak2, jak3, tyk2, stat1, stat3, stat5 (planned protein pages, all referenced as key-nodes)
• cross-links to verified/partial pages used: chronic-inflammation, immunosenescence, senomorphics, growth-hormone, nf-kb, mtor, ampk, cgas-sting
• gaps surfaced:
  • needs-canonical-id — WikiPathways ID not confirmed
  • needs-human-replication — JAKi physical function / SASP benefits (Xu 2015 mouse only)
  • long-term-unknown — systemic JAKi safety in healthy older adults
  • no-mechanism — SOCS3 age-related decline; γ-chain signaling efficiency decline in aged T cells
  • contradictory-evidence — STAT3 pro- vs anti-inflammatory context-dependence
  • unsourced — JAK2 V617F as CHIP driver (specific citation needed); tofacitinib aging data; AMPK-STAT3 cross-talk mechanistic citation
  • needs-replication — SOCS3 decline claim; mTOR-STAT3 Ser727 cross-talk in aging; topical JAKi aging endpoints

oxphos

• added: processes/oxphos.md
• entity type: process
• inbound references resolved: 6 (from mitochondrial-dysfunction, mitophagy, mitochondrial-biogenesis, cytochrome-c, free-radical-theory-of-aging, heterocephalus-glaber)
• canonical IDs: none (type: process — no UniProt/PubChem applies; KEGG pathway hsa00190 exists but not used per process-page conventions)
• primary-source DOIs cited (4 footnotes):
  • doi:10.1038/191144a0 — Mitchell 1961, chemiosmotic hypothesis (Nature; 4,561 citations; 100th percentile; archive: not_oa/closed)
  • doi:10.1042/BJ20081386 — Murphy 2009, mitochondrial ROS production (Biochemical Journal; 7,798 citations; 100th percentile; archive: OA bronze, pending download)
  • doi:10.1038/nature02517 — Trifunovic 2004, PolG mutator mice premature aging (Nature; 2,676 citations; 100th percentile; local PDF available)
  • doi:10.1016/j.xcrm.2022.100633 — Singh 2022 ATLAS RCT urolithin A (Cell Reports Medicine; 182 citations; 100th percentile; local PDF available)
• implicit stubs created:
  • tca-cycle — upstream electron-donor cycle; no page exists
  • pgc-1alpha — PGC-1alpha biogenesis regulator; no page exists (planned R10d/backlog)
  • atp-synthase — CV subunit reference; no page exists
  • complex-i, complex-iv — individual complex pages; no pages exist
  • ampk — cross-referenced; verify page exists (expected from prior rounds)
• gaps surfaced:
  • contradictory-evidence — supercomplex functional coupling debate; PolG causal mechanism ROS vs. energy
  • needs-human-replication — complex activity cross-sectional data; mtROS flux in intact aged tissue; MitoQ/SkQ1 human longevity data
  • needs-replication — NMR oxidative damage baseline; supercomplex cryo-EM aging data; mtDNA heteroplasmy thresholds
  • dose-response-unclear — exercise type/intensity in aged humans; CI vs. CIII ROS contribution
  • long-term-unknown — NAD+ precursor muscle OXPHOS impact

[2026-05-04] ingest | mitochondrial-biogenesis (round-10d)

• added: processes/mitochondrial-biogenesis.md (type: process, verified: false)
• canonical DOIs confirmed in archive (all 4 local PDFs present):
  • doi:10.1016/s0092-8674(00)80611-x (Wu 1999 — PGC-1 biogenesis, Cell, 4058 cites)
  • doi:10.1016/s0092-8674(00)81410-5 (Puigserver 1998 — PGC-1 cloning, Cell, 3791 cites)
  • doi:10.1038/nature07813 (Canto 2009 — AMPK-SIRT1-PGC-1alpha axis, Nature, 3122 cites)
  • doi:10.1016/j.cell.2013.05.039 (Lopez-Otin 2013 — Hallmarks review, Cell, 14200 cites)
• implicit stubs created: pgc-1alpha, nmn, nr, nrf1, tfam, oxphos
• ROADMAP updated: mitochondrial-biogenesis marked drafted (round-10d)
• gaps surfaced: PGC-1alpha decline causation in human tissue; NAD precursor biogenesis endpoints in humans; biogenesis vs mitophagy relative contribution; PGC-1alpha activation without concurrent mitophagy restoration

[2026-05-04] ingest | trem2

• added: molecules/proteins/trem2.md
• entity type: protein (Round 10d)
• canonical IDs: UniProt Q9NZC2, NCBI Gene 54209, HGNC 17761
• DOIs cited: 10.1056/NEJMoa1211851 (local), 10.1056/NEJMoa1211103 (archive failed), 10.1086/342259 (archive pending), 10.1016/j.cell.2015.01.049 (archive pending), 10.1016/j.cell.2017.05.018 (archive failed), 10.1016/j.immuni.2018.11.004 (local)
• implicit stubs created: dap12, dap12-signaling, guerreiro-2013-trem2-ad, jonsson-2013-trem2-ad, paloneva-2002-nasu-hakola, wang-2015-trem2-lipid-sensing, keren-shaul-2017-dam, hammond-2019-microglia-aging
• gaps surfaced: R62H OR primary citation (#gap/unsourced); ADAM10/ADAM17 sheddase citation (#gap/unsourced); R47H mechanism contradiction (#gap/contradictory-evidence); INVOKE-2/Denali trial outcomes (#gap/long-term-unknown)
• verified: false (banner applied)

[2026-05-04] verify | miro1

Pages verified: 1

• molecules/proteins/miro1.md — verified: true (primary-source PDFs all verified; canonical-DB IDs not re-queried live); corrections: 5

Sources checked (all 6 downloaded successfully during this pass):

• doi:10.1016/j.cell.2011.10.018 (Wang 2011) — downloaded (OA bronze); verified; PINK1 phosphorylates Miro Ser156; Parkin ubiquitinates Miro1/2; models are Drosophila motor neurons + rat/mouse hippocampal neurons + HeLa + HEK293T (not MEFs)
• doi:10.1016/j.neuron.2009.01.030 (Macaskill 2009) — downloaded (OA hybrid); verified; Ca²⁺ IC₅₀ ~1 μM for Miro1–KIF5 dissociation; rat hippocampal neuron dendrites (DIV 12–14); Miro1 ΔEF = E208K + E328K
• doi:10.1073/pnas.0808953105 (Saotome 2008) — downloaded (OA green); verified; models are H9c2 cell line + primary cortical neurons (NOT HeLa cells as drafted)
• doi:10.1074/jbc.M208609200 (Fransson 2003) — downloaded (OA hybrid); verified; models are NIH3T3 fibroblasts + COS 7 cells (yeast NOT used experimentally); paper does not test ATP/UTP hydrolysis
• doi:10.15252/embj.2018100715 (López-Doménech 2021) — downloaded (OA hybrid); verified; CaMKIIα-Cre model confirmed; Miro1 ubiquitination → Parkin recruitment → mitophagy; ISR via pS51-eIF2α at 12 months
• doi:10.3389/fneur.2020.00587 (Grossmann 2020) — downloaded (OA gold); verified as review; confirms RHOT1 heterozygous mutations in PD patients (R272Q, T351A, R450C, T610A); impaired Miro1 degradation in patient fibroblasts and iPSC neurons

Corrections made:

1. Fransson 2003 footnote: model “mammalian cells + yeast” → “NIH3T3 fibroblasts + COS 7 cells; yeast not used experimentally”
2. Saotome 2008 footnote: model “neonatal rat cardiomyocytes + HeLa cells” → “H9c2 cardiomyocyte cell line + primary cortical neurons (neonatal rat)” — HeLa cells not used in Saotome 2008
3. Wang 2011 footnote: model “Drosophila motor neurons + HeLa + MEFs” → “Drosophila motor neurons + rat/mouse hippocampal neurons (Parkin-/- mice) + HeLa + HEK293T cells” — MEFs not used; HEK293T and rat neurons were the mammalian models; added “at Ser156” to phosphorylation description
4. Body text: ATP/UTP hydrolysis claim removed from Fransson 2003 attribution (paper identifies GTPase domains but does not test nucleotide hydrolysis specificity); replaced with qualified statement + unsourced tag for NTPase primary citation
5. Extrapolation table: Ca²⁺ arrest pathway conservation changed from “yes — demonstrated in human neuronal cell lines” → “partial — demonstrated in rat hippocampal neurons and H9c2 cells; human direct demonstration lacking” (neither Macaskill 2009 nor Saotome 2008 used human neurons)
6. All archive footnote statuses updated from “archive pending” → "" for the five newly downloaded papers

Unverifiable claims:

• Fransson 2006 (not_oa): MIRO2 ~59% identity to MIRO1 (Fransson 2003 verified states 60%); Miro1-null perinatal lethality; MIRO–TRAK–motor complex stoichiometry details
• Grossmann 2020 is a review; primary data on specific fibroblast assay quantification not independently sourced from original papers
• Canonical-DB identity fields (UniProt Q8IXI2, NCBI Gene 55288, HGNC 21168) — confirmed via banner note; not re-queried against live databases

Downstream pages to check (main agent):

• pink1-parkin-pathway — if it states MIRO1 phospho-site or model details, check for “MEFs” or “HeLa” attribution inherited from the corrected claim
• parkin — inherits MIRO1 as substrate; model description may carry “HeLa + MEFs”; verify phrasing

[2026-05-04] verify | vdac1

Pages verified: 1 (partial scope)

• molecules/proteins/vdac1.md — verified: true (partial scope); corrections: 6

Sources checked:

• doi:10.1073/pnas.0808115105 (Bayrhuber 2008) — downloaded (PMC), verified; human VDAC1 X-ray+NMR hybrid 4.0 Å
• doi:10.1073/pnas.0809634105 (Ujwal 2008) — downloaded (PMC), verified; mouse VDAC1 X-ray 2.3 Å, bicelle crystallization
• doi:10.1038/ncb1575 (Baines 2007) — local PDF, verified; VDAC1/3 DKO MEFs
• doi:10.1038/ncb2012 (Geisler 2010) — local PDF (from parkin.md prior pass); VDAC1 Parkin substrate claim inherited
• doi:10.1126/science.1161302 (Hiller 2008) — no OA URL; download failed; no-fulltext-access
• doi:10.1126/science.1083995 (Cheng 2003) — not_oa; no-fulltext-access
• doi:10.1101/gad.352979.125 (Adedoja 2025) — download failed HTTP 520; no-fulltext-access

Corrections made:

1. Pore diameter: “~2.5–3 nm in diameter” → outer barrel 3.5 × 3.1 nm × 4 nm; inner pore 2.7 × 2.4 nm at orifice, narrowing to 2.7 × 1.4 nm at N-terminal helix (Bayrhuber + Ujwal actual dimensions)
2. Bayrhuber method: “X-ray crystallography” → “X-ray + NMR hybrid” (paper explicitly uses conjoint NMR/X-ray iterative structure determination)
3. Ujwal footnote: added all 283 aa resolved, 47 waters, Rfree/Rwork statistics, N-terminal segment residues 1–26 (helix 6–20), hinge Gly21–25
4. Baines 2007 findings: removed unsupported claim “BAX translocation to mitochondria and BAK activation were unaffected” (paper did not directly measure these steps; it showed downstream outputs — cytochrome c release, caspase cleavage, cell death); clarified DKO MEFs were more sensitive to some stimuli, not protected
5. Baines footnote: expanded stimuli list and corrected measurement methods (calcein-CoCl₂ for mPTP)
6. Added no-fulltext-access tags for Hiller 2008, Cheng 2003, and Adedoja 2025 in Limitations section and footnotes
7. DOI swap confirmed correct: 0808115105 = Bayrhuber (human), 0809634105 = Ujwal (mouse) — no swap error found in page as drafted

Unverifiable claims:

• VDAC2–BAK chaperone function (Cheng 2003, not_oa; claim text is consistent with Ujwal 2008 citing Cheng 2003 ref #38)
• Adedoja 2025 C. elegans lifespan/mtUPR/PeBoW quantitative details (download failed)
• Hiller 2008 NMR-specific structural claims (no OA access; cross-checked via Bayrhuber+Ujwal citations of Hiller)
• UniProt P21796, NCBI Gene 7416, HGNC 12669 not re-queried against live databases

Downstream pages to check (main agent):

• bax — may inherit “BAX translocation unaffected in VDAC-null” framing; verify wording
• bak — check any VDAC1/3 DKO claim wording for the same overstatement
• studies/baines-2007-vdac-dispensable — study page if it exists; should note DKO MEFs more sensitive (not neutral) to some stimuli

[2026-05-04] verify | skeletal-muscle + myocardium (tissue stubs, light verification)

Pages verified: 2 (synthesis-MOC discipline; cross-page consistency only; no PDF re-reads)

• tissues/skeletal-muscle.md — verified: true (partial scope); 1 correction
• tissues/myocardium.md — verified: true (partial scope); 1 correction

Sources cross-checked (inheritance only, against verified atomic pages):

• doi:10.1038/nature03260 (Conboy 2005) — inherited from satellite-cells.md (verified 2026-05-04)
• doi:10.1093/ageing/afy169 (Cruz-Jentoft 2019) — inherited from sarcopenia.md (verified-partial)
• doi:10.1016/j.cell.2015.05.026 (Bergmann 2015) — inherited from cardiomyocytes.md (verified-partial)
• doi:10.15252/embj.2018100492 (Anderson 2019) — inherited from cardiomyocytes.md (verified-partial)
• doi:10.1111/acel.12931 (Lewis-McDougall 2019) — inherited from cardiomyocytes.md (verified-partial)

Corrections:

1. skeletal-muscle.md [^conboy2005] footnote: “young C57Bl/Ka-Ly5.2 (2–3 mo)” → “young C57Bl/Ka (2–3 mo)” — matches verified satellite-cells.md; Ly5.2 was extra/unverified strain detail not in the canonical verified footnote
2. myocardium.md cross-ref summary line: “navitoclax 50 mg/kg (7d × 2 cycles at 24 mo)” → “navitoclax 50 mg/kg/day (7d/cycle × 2 cycles, 24-mo mice)” — restored “/day” and matched phrasing to verified cardiomyocytes.md footnote

Unverifiable claims (remain tagged):

• skeletal-muscle.md: atrogin-1/MURF1 induction magnitude; Type I/II atrophy rate comparison; denervation rates; falls mortality (#gap/unsourced)
• myocardium.md: cardiac fibroblast senescent burden in aged human myocardium; Atg5 cardiac-conditional KO primary citation (#gap/unsourced)

Downstream propagation: none required (corrections are on tissue MOC pages only; verified atomic pages are the authoritative source)

[2026-05-04] verify | smac-diablo

Pages verified: 1 (partial scope — Okada 2002 unconfirmed)

• molecules/proteins/smac-diablo.md — verified: true (partial); corrections: 6

Sources checked:

• doi:10.1038/35022514 (Liu 2000) — local PDF, verified (crystal structure 2.2 Å; homodimer; AVPI; BIR2/BIR3 mutual exclusion)
• doi:10.1016/s0092-8674(00)00008-8 (Du 2000) — downloaded + verified (MTS cleavage at residue 55/56; HeLa S-100 model)
• doi:10.1016/s0092-8674(00)00009-x (Verhagen 2000) — downloaded + verified (293T + NT2 cells; DIABLO interacts MIHA/MIHB/MIHC/OpIAP)
• doi:10.1016/j.cell.2007.10.037 (Vince 2007) — local PDF; findings inherited from xiap.md verifier pass (same date)
• doi:10.1128/mcb.22.10.3509-3517.2002 (Okada 2002) — PMC download failed after two attempts; no-fulltext-access applied to KO phenotype section

Corrections made:

1. MTS cleavage “residues 1–21” → “residues 1–55” (Du 2000 Edman sequencing; mature form starts at residue 56)
2. Precursor processing model: removed invented “~218 aa intermediate” from two-step scheme not present in 2000 papers; single MTS cleavage (1–55) confirmed; further PARL processing noted as later refinement
3. Mature length Identity field: “~184 aa after dual proteolytic processing” → “184 aa (residues 56–239) after MTS cleavage; ~25 kDa”
4. Structure section: “simultaneous engagement of IAP BIR domains” → corrected to note BIR2/BIR3 mutual exclusion per Liu 2000 Fig. 5 competition assay; dimer required for BIR2 (not BIR3) binding
5. IAP antagonism: caspase-9 monomer-sequestration/dimerization detail de-attributed from Liu 2000 (that detail is Shiozaki 2003, already on xiap.md)
6. Vince 2007 senolytic description: “bivalent compound targeting XIAP + cIAP1” → corrected to “Compound A (unnamed small molecule, not named clinical mimetics); primary target is cIAP1 not XIAP; XIAP KO MEFs remain sensitive”
7. Verhagen 2000 footnote: model corrected from “human cell lysates” → “293T cells + mouse NT2 cells; co-IP + sucrose-gradient fractionation”; DOI lookup updated

Unverifiable claims:

• Okada 2002-sourced KO phenotype details (specific stimuli tested, tissue-level data, any lifespan analysis) — PDF not confirmed
• UniProt-sourced BIRC6/BIRC7 ubiquitination sites for SMAC turnover — not re-queried against live UniProt
• Canonical-DB identity fields (UniProt Q9NR28, NCBI 56616, HGNC 21528) — not re-queried

Downstream pages to check (main agent):

• xiap — inherits Vince 2007; already verified; no propagation needed
• Any page that cites SMAC MTS as “residues 1–21” should be corrected to “residues 1–55”

[2026-05-04] verify | atg16l1

Pages verified: 1 (partial scope)

• molecules/proteins/atg16l1.md — verified: true (partial scope); 9 factual corrections + 4 footnote enrichments

Sources checked:

• doi:10.1091/mbc.e07-12-1257 (Fujita 2008) — local PDF, verified
• doi:10.1038/ng1954 (Hampe 2007) — local PDF (path corrected from stale (stale local path) to live mount), verified
• doi:10.1038/ng2032 (Rioux 2007) — local PDF (path corrected), verified
• doi:10.1038/nature07416 (Cadwell 2008) — local PDF, verified
• doi:10.1073/pnas.1407001111 (Lassen 2014) — downloaded and verified
• doi:10.1242/jcs.00381 (Mizushima 2003) — bronze OA; download failed; claims unverified (#gap/no-fulltext-access retained)

Corrections made (see summary in verification output):

1. “Octameric” → “~800-kDa multimeric” — Fujita 2008 never specifies subunit count
2. WIPI2 attribution removed from Fujita 2008 — paper does not mention WIPI2; re-tagged unsourced
3. PI3K-dependent membrane recruitment per Fujita 2008 (not WIPI2/FIP200 bridging)
4. T300A domain location: “coiled-coil/WD40 boundary” → “N-terminal WD-repeat domain” per Hampe 2007
5. T300A MAF corrected; “most significant non-HLA” overstatement removed
6. Caspase motif: “IEGR↓Asp” (caspase-8 motif) → “DXXD caspase-3/7 motif” per Lassen 2014
7. MNV-1 co-factor removed from Cadwell 2008 attribution — belongs to Cadwell 2010 (Cell 141:1135)
8. Cadwell protein levels: “~30%” → “23–37%”
9. Lassen footnote: corrected study design from in-vitro-only to in-vivo knock-in + in-vitro
10. Hampe/Rioux footnotes: sample sizes, ORs added; stale paths corrected
11. Fujita cell lines corrected to PC12, MDCK, MCF7, HEK293A

Downstream pages potentially affected (for main agent to propagate):

• autophagy — WIPI2–ATG16L1 interaction or complex stoichiometry claims; verify
• wipi2 — needs primary citation for WIPI2–ATG16L1 interaction; Fujita 2008 is not the source
• studies/cadwell-2008-atg16l1-paneth-cell — study page may retain “~30%” or MNV-1; update
• Cadwell 2010 (Cell 141:1135–1145) needs a new study page for MNV-1 co-factor claim

[2026-05-04] verify | fip200

Pages verified: 1

• molecules/proteins/fip200.md — verified: true (partial scope; canonical-DB identity fields not re-checked); corrections: 10

Sources checked:

• doi:10.1083/jcb.200712064 (Hara 2008) — local PDF downloaded and verified
• doi:10.1083/jcb.200604129 (Gan 2006) — local PDF downloaded and verified
• doi:10.1016/j.molcel.2019.01.035 (Turco 2019) — local PDF downloaded and verified
• doi:10.1038/nsmb.2475 (Gammoh 2013) — local PDF downloaded and verified
• doi:10.1074/jbc.m900573200 (Ganley 2009) — local PDF already present, re-read inline

Corrections:

1. Embryonic lethality range: E14.5–E16.5 → E13.5–E16.5 (Gan 2006 Table I)
2. Claw domain residues: ~1485–1591 → aa 1494–1594 (Turco 2019 Table 1, isolated Claw construct)
3. Turco 2019 method: “cryo-EM and crystallography” → X-ray crystallography only (no cryo-EM in Turco 2019)
4. NDP52/CLIR claim removed from Claw section and interactors table: Turco 2019 does not characterize FIP200–NDP52 binding; tagged unsourced
5. ATG13 HORMA–Claw contact claim removed: not supported by Turco 2019; tagged unsourced
6. “Single-laboratory” caveat on Turco 2019 corrected: two independent labs (Daumke + Martens)
7. Neural Fip200 cKO (Nestin-Cre) phenotype: removed from [^hara2008] key findings; correctly attributed to Komatsu et al. 2006 (Nature 441:880) in body text; citation-needed flag added
8. “FIP200 promotes ULK1 stability in ATG13-dependent manner”: ATG13-dependency not shown in Hara 2008; claim softened and unsourced portion removed
9. “FIP200 and ATG13 additively stimulate ULK1” wrongly listed as Hara 2008 finding: correctly attributed to Ganley 2009 only
10. Gammoh 2012 publication year: paper published Nat Struct Mol Biol 2013 Feb — year in study file name is incorrect; noted in footnote; FBD residues (229–242) added; mechanistic distinction (starvation- vs glucose-deprivation-induced autophagy) added
11. Gan 2006 mouse background “C57BL/6”: not stated explicitly in paper text; removed from footnote; noted as unverified

Downstream pages to check:

• molecules/proteins/atg13.md — may cite Claw-ATG13 HORMA contact claim
• molecules/proteins/atg101.md — may cite Claw domain residues
• processes/autophagy.md — may cite FIP200 Claw residue range or NDP52 claim
• molecules/proteins/p62.md — may cite Turco 2019 Claw residue range or NDP52 claim
• studies/hara-2008-fip200-ulk1-complex.md — neural cKO listed as key finding; needs correction
• studies/turco-2019-fip200-claw-p62-condensates.md — needs corrected residues, method, and removal of NDP52/ATG13 HORMA claims

[2026-05-04] verify | caspase-7

Pages verified: 1

• molecules/proteins/caspase-7.md — verified: true (partial scope); corrections: 8

Sources checked:

• doi:10.1073/pnas.0707715105 (Walsh 2008) — local PDF, verified
• doi:10.1016/s0092-8674(01)00274-4 (Riedl 2001) — local PDF, verified
• doi:10.1074/jbc.271.4.1825 (Lippke 1996) — downloaded + verified
• doi:10.1074/jbc.270.26.15870 (Munday 1995) — downloaded + verified
• doi:10.1126/science.1115035 (Lakhani 2006) — local PDF download failed twice (green OA; PMC URL resolution failure); verified via PMC web access (PMC3738210)
• UniProt P55210 — verified via REST API

Corrections made:

1. XIAP Ki for caspase-7: “~0.2 nM” → ~100 pM (1.0 × 10⁻¹⁰ M, Riedl 2001 Table 2)
2. XIAP Ki for caspase-3: “~0.7 nM” → ~9.7 pM (9.7 × 10⁻¹² M, Riedl 2001 Table 2)
3. XIAP potency ratio: “~3-fold more potent [for caspase-7]” → caspase-3 ~10x more tightly inhibited (direction and magnitude both wrong)
4. DKO lethality + phenotype: “embryonic/perinatal lethality; brain malformation exceeding single Casp3 KO” → postnatal lethality (died rapidly after birth, present at normal Mendelian ratios through E20); cardiac phenotype (atrial dilation, ventricular noncompaction); brain development normal on C57BL/6
5. Casp7-null ER stress resistance and Fas resistance: removed attribution to Lakhani 2006 (not tested in that paper); tagged unsourced
6. Munday 1995 footnote: “B-cell cDNA library / human B cells” → THP-1 pro-monocytic leukemia cDNA library / HeLa and fibroblast cells
7. Munday 1995 Discovery body: “B-cell cDNA library” → THP-1; added pro-domain-deleted transfection detail
8. Lippke 1996 Discovery body: “HeLa cell cDNA library” → human spleen poly(A)+ cDNA library

Unverifiable claims (source needed):

• Caspase-7 non-redundant ER-stress role (thapsigargin/tunicamycin) — unsourced; possibly Houde et al. 2004
• Casp7-null Fas resistance in thymocytes — unsourced
• Exosite residues ~38–41 — Walsh 2008 describes exosite structurally but specific residues not pinned in paper body

Downstream pages to check (main agent):

• molecules/proteins/caspase-3.md — Walsh 2008 substrate counts; verify correct there
• molecules/proteins/xiap.md — previously verified; Ki values for caspase-3/7 should match Riedl 2001 Table 2 corrections here
• Any page citing Lakhani 2006 for DKO brain malformation — correct to cardiac phenotype + postnatal death

[2026-05-05] verify | nlrp3-inflammasome

Pages verified: 1 (partial scope)

• pathways/nlrp3-inflammasome.md — verified: true (partial scope); corrections: 5 factual + 3 footnote enrichments

Sources checked:

• doi:10.1038/ng756 (Hoffman 2001) — local PDF, verified
• doi:10.1016/s1097-2765(02)00599-3 (Martinon 2002) — downloaded + verified
• doi:10.1016/s1074-7613(04)00046-9 (Agostini 2004) — downloaded + verified
• doi:10.1038/nature08938 (Duewell 2010) — local PDF, verified
• doi:10.1056/NEJMoa1707914 (Ridker 2017 CANTOS) — local PDF, verified (numerics confirmed)
• doi:10.1038/nature11729 (Heneka 2013) — downloaded + verified
• doi:10.1038/nm.3806 (Coll 2015) — download pending (green OA); MCC950 claims unverified needs-verification

Corrections made:

1. Heneka 2013 — removed erroneous “reduces tau phosphorylation” claim (not present in the paper); added accurate outcomes: ~70% reduction in FA-extractable Aβ, M2 microglial skewing, LTP preservation, elevated cleaved caspase-1 in human MCI/AD brain (n=12 AD vs n=8 controls)
2. Heneka 2013 — removed unsourced claim that “ASC specks released by pyroptotic microglia can seed Aβ aggregation extracellularly”; this is a later finding (Venegas 2017) not from Heneka 2013; added attribution note
3. Hoffman 2001 footnote — corrected “PYD-NACHT linker” to “NBS/NACHT domain” (mutations A439V, V198M, E627G, A352V are all in the NACHT domain per Fig. 3b of the paper, not in a linker region); added protein size (920 aa, 105.7 kDa, pI 6.16) and control sample size (>100)
4. Duewell 2010 — added quantitative lesion reduction (~69%, P<0.0001) to both body text and footnote; added bone marrow transplant design details and n-per-group (n=7–9); clarified that priming (Signal 1) is still required in vitro
5. Martinon 2002 footnote — corrected “review/concept paper” to “in-vitro”; clarified that NALP1 (not NLRP3) is the founding member studied in this paper; updated DOI lookup from “pending” to local PDF
6. Agostini 2004 footnote — expanded: added key finding that NLRP3 inflammasome activates caspase-1 but NOT caspase-5 (unlike NALP1); added single MWS patient R260W detail; updated DOI lookup
7. Heneka 2013 footnote — corrected “tau phosphorylation” error; added n-per-group for all cohorts; updated DOI lookup
8. CANTOS numerics (HR 0.85, CI 0.74–0.98, p=0.021; infection rates 0.31 vs 0.18/100py) — confirmed correct against PDF; no change needed

Downstream pages potentially affected (for main agent to propagate):

• alzheimers-disease — may cite tau phosphorylation claim attributed to Heneka 2013; verify
• atherosclerosis — may cite Duewell 2010 lesion reduction; update to add ~69% figure if present

[2026-05-04] verify | atg3

Pages verified: 1

• molecules/proteins/atg3.md — verified: true (partial scope); corrections: 5

Sources checked:

• doi:10.1038/35044114 (Ichimura 2000) — local PDF, verified
• doi:10.1074/jbc.M200385200 (Tanida 2002) — downloaded + verified (corrected from seeder DOI M112303200)
• doi:10.1091/mbc.e08-03-0309 (Sou 2008) — downloaded + verified
• doi:10.1074/jbc.M611473200 (Yamada 2007) — download failed (hybrid OA, 0 URLs after filtering); structural claims unverified no-fulltext-access
• doi:10.1016/j.febslet.2015.02.003 (Sakoh-Nakatogawa 2015) — closed-access (not_oa); unverified no-fulltext-access

Corrections made:

1. Sou 2008 n “N/A (multiple neonates; not confirmed)” → n=11 (deaths within 1d), n=8 (body weight) per Results text
2. Sou 2008 body weight added: 0.99 ± 0.12 g (Atg3-/-) vs 1.19 ± 0.05 g (WT/het), p<0.002
3. Sou 2008 mechanism “isolation membranes formed but did not close properly” → four-component EM defect description (LC3 mislocalisation, delayed Atg12-Atg5 dissociation, aberrant elongation, impaired closure); aberrant structure frequency 22.0±10.2% vs 5.5±5.9% AVi (p<0.01) added
4. Ichimura 2000 footnote enriched: yeast active-site Cys234 (= human Cys264 equivalent), Atg8-PE amide bond details confirmed
5. Yamada 2007 footnote updated to “download failed” with no-fulltext-access; Limitations section updated accordingly

Unverifiable claims:

• All Yamada 2007-sourced structural data (FR residue ranges ~65-130, LIR ~104-110, E1-E2 trans-docking geometry, ~140-170 E3-contact region)
• All Sakoh-Nakatogawa 2015-sourced data (N-terminal amphipathic helix membrane curvature sensing; LIR-motif phagophore expansion)
• UniProt Q9NT62 identity fields (314 aa, PTM list) not re-verified against live database

Downstream pages to check (main agent):

• tanida-2002-atg3-e2-lc3-lipidation — study stub page; DOI correction M112303200→M200385200 needed if present
• yamada-2007-atg3-crystal-structure — study stub page; DOI correction M608787200→M611473200 needed if present
• lc3 — cites Ichimura 2000; no numeric corrections needed from that paper
• autophagy — if it cites Sou 2008 n counts, may need update

[2026-05-04] ingest | round-10c-machinery-tissues

skeletal-muscle

• added: tissues/skeletal-muscle.md
• entity type: tissue (light anchor stub)
• inbound references resolved: 8 (from satellite-cells, sarcopenia, frailty, glut4, type-2-diabetes, deregulated-nutrient-sensing, cardiac-fibrosis, frailty affected-tissues field)
• primary-source DOIs cited (4 footnotes): 10.1038/nature03260 (Conboy 2005; local PDF), 10.1038/nature13013 (Sousa-Victor 2014; not_oa), 10.1093/ageing/afy169 (Cruz-Jentoft 2019 EWGSOP2; local PDF), 10.1001/jama.1990.03440220053029 (Fiatarone 1990; not_oa)
• implicit stubs created: myofibers, fibroadipogenic-progenitors, notch-pathway, wnt-pathway, myostatin, bone
• gaps surfaced: unsourced — atrogin-1/MURF1 induction magnitude, Type I vs Type II atrophy rate, denervation rate; needs-human-replication — FAP fate-switch, senolytics for sarcopenia
• schema gaps: type: tissue not formally defined in CLAUDE.md; followed R10b brain/bone-marrow convention; flagged for schema formalization
• ROADMAP: skeletal-muscle marked [x] drafted 2026-05-04

myocardium

• added: tissues/myocardium.md
• entity type: tissue (light anchor stub)
• inbound references resolved: 7 (from cardiomyocytes, cardiac-fibrosis, heart-failure, cardiovascular-aging, atherosclerosis, cardiac-fibrosis affected-tissues field)
• primary-source DOIs cited (3 footnotes — all cross-ref verified atomic pages): 10.1016/j.cell.2015.05.026 (Bergmann 2015; local PDF), 10.15252/embj.2018100492 (Anderson 2019; local PDF), 10.1111/acel.12931 (Lewis-McDougall 2019; local PDF)
• implicit stubs created: cardiac-fibroblasts (planned R10d), endothelial-cells, heart
• gaps surfaced: needs-human-replication — senolytics for cardiac aging; needs-replication — Lewis-McDougall 2019 CPC clearance; unsourced — cardiac fibroblast senescent burden, Atg5 cardiac-cKO citation; contradictory-evidence — NF-κB role in CM SASP
• schema gaps: same type: tissue schema gap as skeletal-muscle entry
• ROADMAP: myocardium marked [x] drafted 2026-05-04

vdac1

• added: molecules/proteins/vdac1.md
• entity type: protein
• inbound references resolved: 8 (from parkin, bcl-xl, bak, bax, apoptosis-pathway, mitophagy, mitochondrial-dysfunction, and one additional context)
• canonical IDs pulled:
  • UniProt P21796 (VDAC1_HUMAN; Swiss-Prot, manually reviewed) — 283 aa; ~31 kDa; confirmed via REST API
  • NCBI Gene: 7416; HGNC: 12669; mouse ortholog: Vdac1
  • GenAge: not found in human subset — needs-canonical-id
• DOI correction: task brief cited Ujwal 2008 as doi:10.1073/pnas.0809996105 — “DOI not found” in archive AND Crossref 404; correct DOI confirmed via PubMed PMID 18988731 efetch = 10.1073/pnas.0809634105 (which the archive had already indexed as the Ujwal crystal structure); brief’s Bayrhuber DOI (10.1073/pnas.0809634105) was actually the Ujwal DOI — corrected both attributions; true Bayrhuber 2008 DOI = 10.1073/pnas.0808115105 (confirmed via PubMed PMID 18832158)
• primary-source DOIs cited (7 footnotes):
  • doi:10.1126/science.1161302 — Hiller 2008, VDAC1 NMR structure (Science; 651 citations; 100th percentile; archive pending, green OA)
  • doi:10.1073/pnas.0808115105 — Bayrhuber 2008, VDAC1 X-ray 4.0 Å (PNAS; 564 citations; 100th percentile; archive pending, green OA via PMC)
  • doi:10.1073/pnas.0809634105 — Ujwal 2008, mouse VDAC1 X-ray 2.3 Å (PNAS; 529 citations; 100th percentile; archive pending, green OA)
  • doi:10.1126/science.1083995 — Cheng 2003, VDAC2 chaperones BAK (Science; 814 citations; 100th percentile; archive not_oa/closed)
  • doi:10.1038/ncb1575 — Baines 2007, VDAC1/3 DKO dispensable for MOMP (Nat Cell Biol; 962 citations; 100th percentile; LOCAL PDF available)
  • doi:10.1038/ncb2012 — Geisler 2010, PINK1/Parkin/VDAC1 mitophagy (Nat Cell Biol; 2,729 citations; 100th percentile; LOCAL PDF available; K27-linkage claim already verified on parkin.md)
  • doi:10.1101/gad.352979.125 — Adedoja 2025, VDAC-1 reduction extends C. elegans lifespan via mtUPR (Genes & Dev; 0 citations; archive pending, diamond OA)
• implicit stubs created (new wikilinks to non-existent pages):
  • MAMs — ER-mitochondria contact sites; no page exists yet
  • pink1-parkin-pathway — pathway page stub; individual pink1 and parkin pages exist but no pathway-level page
  • cgas-sting-pathway — innate immune pathway; relevant to mtDNA release aging context
  • hiller-2008-vdac1-nmr — study page stub
  • bayrhuber-2008-vdac1-crystal — study page stub
  • ujwal-2008-vdac1-crystal-mouse — study page stub
  • cheng-2003-vdac2-bak — study page stub
  • baines-2007-vdac-dispensable — study page stub
  • geisler-2010-parkin-vdac1-mitophagy — study page stub (Geisler 2010 previously cited on parkin.md; that page’s footnote uses a different stub key)
  • adedoja-2025-vdac1-lifespan-mtUPR — study page stub
• gaps surfaced:
  • unsourced — age-dependent changes in VDAC1 channel conductance / PTM landscape in aged human tissue
  • needs-human-replication — VDAC-1 reduction → lifespan extension (C. elegans only; Adedoja 2025)
  • needs-replication — VDAC1 oligomerization → mtDNA release → cGAS/STING in aging endothelium (single 2026 study)
  • contradictory-evidence — VDAC1 role in MOMP (MAC model vs. Baines 2007 DKO counter-evidence)
  • no-mechanism — K27 vs K63 polyUb on VDAC1: differential routing to autophagosome vs proteasome
  • needs-canonical-id — VDAC1 not in GenAge human subset
• schema gaps: none; type: protein schema fully covers this entity
• verification priority: HIGH — Baines 2007 and Geisler 2010 both have local PDFs; Baines 2007 VDAC1/3 DKO MOMP data and Geisler 2010 K27 polyUb specificity are the highest-priority claims to verify; Cheng 2003 (VDAC2-BAK) is closed-access (#gap/no-fulltext-access if needed); three structure papers (Hiller, Bayrhuber, Ujwal) are green/diamond OA and can be downloaded

age-1

• added: molecules/proteins/age-1.md
• entity type: protein (type: protein; organism: Caenorhabditis elegans)
• canonical IDs:
  • UniProt Q94125 (AGE1_CAEEL) — confirmed via REST API search gene:age-1 + organism_id:6239
  • WormBase WBGene00000090 — confirmed via UniProt cross-reference
  • NCBI Gene 174762 — confirmed via UniProt cross-reference
  • GenAge model organisms 174762 (longevity influence: anti-longevity)
  • HGNC: null (non-human gene)
• canonical ID corrections vs task brief:
  • UniProt Q17758 (brief) → Q94125 (correct); Q17758 resolves to pept-2 (C. elegans peptide transporter)
  • NCBI Gene 174213 (brief) → 174762 (correct); 174213 resolves to discontinued srd-52 record
  • Protein length 1,209 aa (brief) → 1,182 aa (correct per UniProt Q94125)
• DOIs cited (4 footnotes):
  • doi:10.1093/genetics/118.1.75 (Friedman & Johnson 1988 — founding age-1 longevity paper) — LOCAL PDF available; impact_score 0.649; 985 citations
  • doi:10.1038/382536a0 (Morris 1996 — age-1 molecular cloning as PI3K; Nature) — NOT_OA; no-fulltext-access; impact_score 0.717; 889 citations
  • doi:10.1016/s1534-5807(04)00095-4 (Hertweck 2004 — SGK-1 dominance; Dev Cell) — LOCAL PDF available; impact_score 0.676; 364 citations
  • doi:10.1101/gad.1255404 (Apfeld 2004 — AAK-2/AMPK parallel to DAF-16; Genes & Dev) — LOCAL PDF available; 632 citations
• implicit stubs created (new wikilinks to non-existent pages):
  • pdk-1 — C. elegans 3-phosphoinositide-dependent kinase; no protein page yet
  • daf-18 — C. elegans PTEN ortholog; no protein page yet
• existing pages cross-referenced: insulin-igf1, pi3k-akt-pathway, pi3k, daf-2, daf-16, akt, sgk1, pten, irs2, ampk, caenorhabditis-elegans, deregulated-nutrient-sensing
• gaps surfaced:
  • needs-replication — exact Friedman 1988 lifespan percentages (verified for existence, not numerics); requires PDF cross-check
  • no-fulltext-access — Morris 1996 (not_oa); cloning + fer-15 background confirmation blocked
  • needs-human-replication — SGK-1 dominance in worm longevity not tested in mammalian systems
  • dose-response-unclear — partial AGE-1 inhibition vs constitutive dauer threshold
  • no-mechanism — tissue-specificity of AGE-1 longevity effect in worms
  • needs-replication — single-mutant age-1 lifespan extension (without fer-15 background) — secondary-source estimate
• schema gaps: none — protein schema accommodates non-human genes via wormbase-id field (pre-existing on daf-2.md); hgnc: null is appropriate convention; organism field added (not in protein schema prototype but used on daf-2.md — consistent with existing practice, not a new invention
• verification priority: HIGH — Friedman 1988 (LOCAL PDF) and Hertweck 2004 (LOCAL PDF) can be verified now; Morris 1996 permanently blocked (not_oa); this is the founding longevity gene of the field — quantitative claims warrant careful PDF verification

nlrp3-inflammasome

• added: pathways/nlrp3-inflammasome.md
• entity type: pathway
• canonical IDs: KEGG hsa04621, Reactome R-HSA-844456
• DOIs cited (7 total):
  • doi:10.1016/s1097-2765(02)00599-3 (Martinon 2002 — inflammasome concept) — archive pending (OA bronze)
  • doi:10.1038/ng756 (Hoffman 2001 — CIAS1/CAPS) — LOCAL PDF available
  • doi:10.1016/s1074-7613(04)00046-9 (Agostini 2004 — NALP3 inflammasome MWS) — archive pending (OA bronze)
  • doi:10.1038/nature08938 (Duewell 2010 — cholesterol crystal NLRP3 atherosclerosis) — LOCAL PDF available
  • doi:10.1056/NEJMoa1707914 (Ridker 2017 — CANTOS) — LOCAL PDF available
  • doi:10.1038/nature11729 (Heneka 2013 — NLRP3 in AD/APP-PS1) — archive pending (OA green)
  • doi:10.1038/nm.3806 (Coll 2015 — MCC950) — archive pending (OA green)
• DOI correction: task brief cited Hoffman 2001 as doi:10.1038/ng720 — this is the Miceli-Richard 2001 CARD15/Blau syndrome paper (Nature Genetics, 958 cites); correct Hoffman 2001 CAPS paper is doi:10.1038/ng756 (Nature Genetics, 1,631 cites, local PDF confirmed)
• implicit stubs created: nlrp3-protein, asc, caspase-1, il-1b, il-18, gsdmd, pyroptosis, damps-pamps
• gaps surfaced: needs-canonical-id (all 6 key-node protein stubs); needs-human-replication (MCC950; NLRP3-AD; NLRP3-T2D); long-term-unknown (NLRP3 inhibitors in aging); no-mechanism (cardiac NLRP3 DAMPs)

atg3

• added: molecules/proteins/atg3.md
• entity type: protein
• canonical IDs: UniProt Q9NT62, NCBI Gene 64422, HGNC 20962
• DOI correction: task brief cited Tanida 2002 as doi:10.1074/jbc.M112303200 and Yamada 2007 as doi:10.1074/jbc.M608787200 — both are DOI-title mismatches in archive (BUG-2 pattern); correct DOIs confirmed via PubMed efetch: Tanida 2002 = doi:10.1074/jbc.M200385200 (PMID 11825910), Yamada 2007 = doi:10.1074/jbc.M611473200 (PMID 17227760); Sakoh-Nakatogawa is 2015 (not 2013) at doi:10.1016/j.febslet.2015.02.003 (PMID 25680528)
• DOIs cited (5 total):
  • doi:10.1074/jbc.M200385200 (Tanida 2002 — human ATG3 E2 cloning + activity) — archive pending, OA URL available
  • doi:10.1074/jbc.M611473200 (Yamada 2007 — ATG3 crystal structure PDB 2DYT) — archive pending, OA URL available
  • doi:10.1038/35044114 (Ichimura 2000 — yeast Atg8-PE conjugation cascade) — LOCAL PDF available
  • doi:10.1091/mbc.e08-03-0309 (Sou 2008 — Atg3-/- neonatal lethal, phagophore EM) — archive pending, green OA URL available
  • doi:10.1016/j.febslet.2015.02.003 (Sakoh-Nakatogawa 2015 — ATG3 membrane localization/curvature sensing) — not_oa, no-fulltext-access flagged
• implicit stubs created (new wikilinks to non-existent pages): none — all cross-linked pages confirmed existing in wiki (atg7, lc3, atg10, atg5, atg12, atg16l1, beclin-1, autophagy, mitophagy, bnip3, fundc1, disabled-macroautophagy, loss-of-proteostasis)
• gaps surfaced: needs-human-replication (no human ATG3 null phenotype; aging data mouse-only); needs-replication (ATG3 protein levels in aged human tissues); no-mechanism (membrane curvature sensing in aged cells; ATG12~ATG3 role in aging); no-fulltext-access (Sakoh-Nakatogawa 2015 — not_oa)
• schema gaps: none — type: protein schema fully accommodates ATG3; sens-categories: [] consistent with other autophagy machinery pages
• verification priority: MEDIUM — Ichimura 2000 local PDF can be verified now; Tanida 2002 and Sou 2008 are green/hybrid OA and can be downloaded; Sakoh-Nakatogawa 2015 is not_oa (closed-access); Yamada 2007 has OA URL available

atg16l1

• added: molecules/proteins/atg16l1.md
• entity type: protein (type: protein)
• canonical IDs: UniProt Q676U5, NCBI Gene 55054, HGNC 21498, mouse ortholog Atg16l1
• length: 607 aa (canonical isoform α); multiple splice isoforms (α/β/γ) expressed with tissue-specific distributions
• key sections:
  • Identity table
  • Domain structure (ATG5-binding region 1–79; coiled-coil 80–320; WD40 repeats 320–607 — 7 propellers; vertebrate-specific addition absent in yeast Atg16p)
  • E3-like complex assembly (ATG12–ATG5–ATG16L1; ~800 kDa octameric assembly via coiled-coil dimerization per Fujita 2008)
  • Discovery and yeast comparison (Mizushima 2003 cloning; yeast Atg16p lacks WD40 domain)
  • Disease associations (T300A rs2241880 Crohn GWAS × 2 independent groups; Paneth cell hypomorph model; caspase-3/7 cleavage mechanism at T300A)
  • Knockout phenotypes (germline lethal; intestinal + macrophage conditional KO data; note: primary KO paper not captured in footnotes)
  • Aging context (autophagic flux decline; T300A aging trajectory hypothesis; Paneth cell niche/stem-cell-exhaustion link; pro-longevity pathway convergence)
  • Cross-references and pathway membership
  • Limitations and gaps (8 gap tags)
• primary DOIs cited (6 footnotes):
  • 10.1242/jcs.00381 (Mizushima 2003 cloning) — archive: pending; no-fulltext-access for quantitative claims until downloaded
  • 10.1091/mbc.e07-12-1257 (Fujita 2008 complex/LC3 site) — archive: LOCAL PDF available
  • 10.1038/ng1954 (Hampe 2007 T300A Crohn GWAS) — archive: LOCAL PDF available
  • 10.1038/ng2032 (Rioux 2007 Crohn GWAS replication) — archive: LOCAL PDF available
  • 10.1038/nature07416 (Cadwell 2008 Paneth cell hypomorph) — archive: LOCAL PDF available
  • 10.1073/pnas.1407001111 (Lassen 2014 T300A caspase cleavage) — archive: pending; no-fulltext-access until downloaded
• DOI correction: task brief supplied DOI 10.1073/pnas.1402805111 for Lassen 2014 — NOT found in archive; correct DOI confirmed as 10.1073/pnas.1407001111 via PubMed efetch PMID 24821797
• implicit stubs created (new wikilinks to non-existent pages):
  • xenophagy — selective autophagy of intracellular pathogens; no process page yet
  • wipi2 — PI3P effector bridging VPS34/beclin-1 to ATG16L1 phagophore recruitment; no protein page yet
  • fip200 — planned R10c (still unseeded); ULK1 complex scaffolding subunit
• existing pages cross-referenced: atg5, atg12, atg7, atg10, lc3, autophagy, mitophagy, beclin-1, ulk1, ampk, disabled-macroautophagy, loss-of-proteostasis, stem-cell-exhaustion
• gaps surfaced:
  • unsourced — germline Atg16l1-/- embryonic lethality: primary publication not captured in footnotes
  • unsourced — age-associated ATG16L1 protein level decline across tissues
  • unsourced — intestinal epithelial conditional Atg16l1-/- primary citation
  • unsourced — non-canonical ATG16L1-WD40 role in LAP/CASM
  • needs-replication — ATG16L1 overexpression lifespan extension (not tested)
  • needs-human-replication — T300A aging trajectory hypothesis (no aging cohort data)
  • no-mechanism — T300A effect on aging-related autophagy decline trajectory
  • dose-response-unclear — partial ATG16L1 reduction in T300A heterozygotes under physiological conditions
• schema gaps: none; type: protein schema fully accommodates ATG16L1
• verification priority: MEDIUM-HIGH — 9 inbound references; 4 of 6 PDFs locally available (Fujita 2008, Hampe 2007, Rioux 2007, Cadwell 2008); Mizushima 2003 + Lassen 2014 pending (bronze OA — downloadable); suggest wiki-verifier prioritize Fujita 2008 (complex stoichiometry ~800 kDa claim) and Lassen 2014 (T300A caspase-cleavage mechanism claims)
• ROADMAP updated: atg16l1 marked [x] drafted 2026-05-04 in R10c section

[2026-05-04] verify | xiap

Pages verified: 1 (partial scope — 5 of 6 cited primary sources verified against local PDF; 1 unverifiable closed-access)

• molecules/proteins/xiap.md — corrections: 9

Sources verified against local PDF:

• Liston 1996 (10.1038/379349a0) — already local; read in full
• Deveraux 1997 (10.1038/40901) — already local; read in full
• Riedl 2001 (10.1016/s0092-8674(01)00274-4) — already local; read in full
• Shiozaki 2003 (10.1016/s1097-2765(03)00054-6) — triggered download (bronze OA via camoufox), read in full
• Vince 2007 (10.1016/j.cell.2007.10.037) — triggered download (bronze OA via camoufox), read in full

Sources unverifiable:

• Marsh 2010 (10.1182/blood-2010-01-256099) — not_oa; XLP-2 clinical claims tagged no-fulltext-access
• Duckett 1996 (PMID:8654366, no DOI) — not in a local paper archive; not_oa per CrossRef lookup; discovery role only, no mechanistic claims depend solely on this paper

Corrections made:

1. Chromosomal location: “Xq25” → “Xq24-25” (Liston 1996 maps to Xq24-25 STS locus)
2. Domain table — BIR2 row: “~155–230” → corrected to “N-terminal extension + BIR2, 124–230” with BIR2 domain proper 163–230 per UniProt; description changed from “linker (~Gly163–Glu180)” to accurate “hook (138–146) + sinker (148–156)” terminology from Riedl 2001
3. Domain table — RING row: “~437–497” → “450–485” per UniProt P98170 REST API (RING-type zinc finger boundary)
4. Caspase-3/-7 inhibition mechanism: removed fabricated “Glu167 mutation” claim; replaced with correct primary contacts L141 (hook) and D148 (sinker) as identified in Riedl 2001 crystal structure
5. RING E3 substrates section: removed incorrect citation of Vince 2007 for caspase-3/RIPK1/RIPK2/MAP3K2/SMAC ubiquitination by XIAP RING; Vince 2007 addresses cIAP1 ubiquitin ligase, not XIAP substrates; tagged unsourced with suggested primary citations
6. SMAC/DIABLO antagonism: removed incorrect Vince 2007 citation for AVPI mechanism; replaced with correct Riedl 2001 + Shiozaki 2003 citations that actually characterize the AVPI-BIR interaction structurally
7. Deveraux 1997 description: corrected misleading framing (“inhibition could be reversed by cytochrome c + Apaf-1”) to accurate description (cytochrome c triggers caspase activation; XIAP blocks it); added Ki values (caspase-3 ~0.7 nM, caspase-7 ~0.2 nM) and BIR vs RING domain activity finding
8. SMAC mimetics table: removed LCL161/birinapant/AT-406 attribution to Vince 2007 (paper uses only unnamed “Compound A”); added unsourced tags with suggested real citations; mechanistic note now accurately reflects the paper’s cIAP1-primary finding with XIAP being non-rate-limiting
9. Shiozaki 2003 footnote: updated from “pending” to local PDF path; added crystal structure details (2.4 Å, BIR3 residues 252–350, caspase-9 ATPF motif = Ala316–Phe319)

Unverifiable claims (not corrected, tagged):

• XLP-2 clinical features (Marsh 2010, closed-access)
• S-nitrosylation Cys450 functional consequence (no locally-available primary source)
• Omi/HtrA2 XIAP cleavage mechanism
• XIAP overexpression in specific tumor histologies
• XIAP RING domain substrates (caspase-3 ubiquitination etc.) — sourced to literature not yet in local archive

Downstream pages to check:

• processes/apoptosis.md (if it cites XIAP BIR2 linker mechanism — may have inherited “Glu167” or “Gly163” framing)
• molecules/compounds/smac-mimetics.md or interventions/pharmacological/senolytics.md (if they cite Vince 2007 for XIAP direct mechanism or for named SMAC mimetic compounds)
• processes/cellular-senescence.md (if it cites XIAP RING substrates via Vince 2007)

[2026-05-04] verify | tbk1

Pages verified: 1 (partial scope — 6 of 7 sources verified against local PDF; 1 unverifiable)

• molecules/proteins/tbk1.md — corrections: 8

Sources verified against local PDF:

• Pomerantz & Baltimore 1999 (10.1093/emboj/18.23.6694) — downloaded (bronze OA), read in full
• Hemmi 2004 (10.1084/jem.20040520) — downloaded (hybrid OA via PMC), read in full
• Heo 2015 (10.1016/j.molcel.2015.08.016) — downloaded (bronze OA via camoufox), read in full
• Cirulli 2015 (10.1126/science.aaa3650) — downloaded (green OA via UniMI), read in full
• Freischmidt 2015 (10.1038/nn.4000) — downloaded (green OA via HAL), read in full
• Reilly 2013 (10.1038/nm.3082) — downloaded (green OA via PMC), read in full

Sources unverifiable:

• Wild 2011 (10.1126/science.1205405) — not_oa; Ser177/LC3 attribution confirmed via Heo 2015 cross-citation

Corrections made:

1. Hemmi 2004 footnote model corrected: “Tbk1−/−-reconstituted bone-marrow and MEF” → “TBK1−/− and IKKε−/− mouse embryonic fibroblasts (EFs from E12.5 embryos).” Paper used EFs; bone marrow data was “unpublished data.”
2. Hemmi 2004 body text: clarified IKKε is secondary (IFN-β abolished only in double-KO EFs), replacing imprecise framing.
3. Heo 2015 footnote model corrected: “human cells + mouse cardiac/brain” → “HeLa Flip-In T-REx (HFT) cells + HEK293T.” No mouse cardiac/brain tissue in this paper.
4. Heo 2015: added primary finding — TBK1 phosphorylates OPTN-S473 and OPTN-S513 (UBAN domain) to enhance polyubiquitin chain binding; this is the main biochemical discovery of the paper.
5. Substrate table: added Source column; split OPTN into two rows (Ser177 per Wild 2011; Ser473/Ser513 per Heo 2015), correcting implicit misattribution.
6. Cirulli 2015: “~3,000 ALS patients” → “2869 ALS patients + 6405 controls”; added p-value (combined p=3.60×10−11) and variant frequency data.
7. Freischmidt 2015: “3.0%” → “~3.6% (9/252) / ~4% of genetically unexplained fALS”; added mutation count, HRMA screen details, CCD2 biochemical finding (OPTN binding disrupted, kinase activity preserved).
8. Reilly 2013 — CRITICAL FABRICATION REMOVED: “Phase 2 clinical trial in type 2 diabetes showed improvements in HbA1c and liver fat in a subset of patients” is absent from the paper. Reilly 2013 is entirely preclinical mouse work that explicitly proposes future clinical studies. Replaced with accurate preclinical data (IC50 ~1–2 μM, ~10 g weight loss, >50% hepatic triglyceride reduction in DIO mice) and needs-human-replication tag.

Unverifiable claims:

• Wild 2011 (OPTN-Ser177/LC3; xenophagy model): not_oa; tagged no-fulltext-access; consistent with Heo 2015 citations
• Domain boundary table (KD 9–310, ULD 309–385, SDD-1 407–657, CTD 657–729): attributed to UniProt Q9UHD2 per page note; not re-verified in this pass

[2026-05-04] verify | ubiquitin-proteasome-system

Pages verified: 1 (partial scope — 3 of 6 sources verified against local PDF; 3 unverifiable)

• processes/ubiquitin-proteasome-system.md — corrections: 5

Sources verified against local PDF:

• Hershko & Ciechanover 1998 (10.1146/annurev.biochem.67.1.425) — local PDF read in full; confirmed α7β7β7α7 core architecture; single E1 in 1998 (UBA6 is post-1998); four E3 functional types per 1998 literature; ~13 Å α-ring gate (T. acidophilum); ~8 residue peptides (from 28 Å active-site spacing)
• Saez & Vilchez 2014 (10.2174/138920291501140306113344) — downloaded (OA green) and read in full; confirmed 8-subunit Lid (not 9); bortezomib FDA approval 2003; confirmed proteasome activity decline in aged human fibroblasts, lymphocytes, brain, muscle tissue
• Bonaldo & Sandri 2012 (10.1242/dmm.010389) — downloaded (OA gold) and read in full; confirmed atrogin-1 (FBXO32/MAFbx) and MuRF1 (TRIM63) as dominant muscle-specific E3s during atrophy; >650 E3 ligases in human genome

Sources unverifiable:

• Glickman & Ciechanover 2002 (10.1152/physrev.00027.2001) — not_oa; E3 sub-class counts (~600 RING, ~28 HECT, ~14 RBR) unverified; tagged no-fulltext-access
• Vilchez/Saez/Dillin 2014 (10.1038/ncomms6659) — bronze OA, download failed (no candidate URLs); PROTAC and UPS–autophagy integration claims unverified; tagged no-fulltext-access
• Chondrogianni 2015 (10.1096/fj.14-252189) — download corrupted (wrong PDF fetched — library conference presentation; BUG-6 filed in a local paper archive/FEATURE_REQUESTS.md); C. elegans lifespan extension claims unverified; tagged no-fulltext-access

Corrections made:

1. 19S Lid subunit count: “9 non-ATPase subunits” → “8 non-ATPase subunits (Rpn3, Rpn5, Rpn6, Rpn7, Rpn8, Rpn9, Rpn11, Rpn12)” per Saez & Vilchez 2014
2. E1 count note added: Hershko 1998 says “single E1”; UBA6 is a 2007 discovery; current “2 (UBA1, UBA6)” is accurate but the [^hershko1998] citation alone does not support it — note added to table
3. E3 count updated in table: ~600 → ~600–650 to align with Bonaldo 2012 (“>650”) and existing ~600 RING estimates; note added on unsourced for precise current count
4. Peptide size claim clarified: “3–22 amino acids” is from post-1998 literature; Hershko 1998 estimates ~8 residues from structural distance; α-ring gate noted as closed in eukaryotes (unlike T. acidophilum)
5. Footnote labels corrected: [^chondrogianni2014] → [^chondrogianni2015] (FASEB J published 2015, accepted 2014); all footnotes updated with verification status and author/journal details
6. Auto-extracted ⚠️ banner removed; verified flag flipped to true (partial scope)

BUG filed: BUG-6 in a local paper archive/FEATURE_REQUESTS.md — Chondrogianni 2015 PDF download fetched wrong content (Huddersfield library presentation). Same failure mode as BUG-3/4/5.

[2026-05-04] verify | bcl-w

Pages verified: 1 (partial scope — 5 of 8 cited primary sources verified against local PDF; 2 not_oa; 1 download failed)

• molecules/proteins/bcl-w.md — corrections: 9

Sources verified against local PDF:

• Chen 2005 (10.1016/j.molcel.2004.12.030) — already local; read in full; BH3-binding IC50 table corrected
• Yosef 2016 (10.1038/ncomms11190) — already local; read in full; senescence inducer corrected (OIS → DIS for key siRNA experiment)
• Zhu 2016 (10.1111/acel.12445) — already local; read in full; senescence inducer corrected (genotoxic → radiation/10 Gy)
• Zhu 2017 (10.18632/aging.101202) — already local; read in full; A1331852 IMR-90 activity corrected (partial → clearly senolytic in both HUVECs and IMR-90)
• Hinds 2003 (10.1093/emboj/cdg144) — triggered download (bronze OA via EuropePMC); read in full; confirmed NMR solution structure, C-terminal groove occlusion model

Sources unverifiable:

• Gibson 1996 (PMID:8761287; Oncogene) — not_oa; no DOI; source confirmed via PubMed
• Print 1998 (10.1073/pnas.95.21.12424) — not_oa; male-sterile phenotype claims retained as sourced
• Chang 2016 (10.1038/nm.4010) — green OA but DOI lookup failed (0 candidate URLs after filtering); tagged no-fulltext-access; retry failed

Corrections made:

1. Gibson 1996 DOI: 10.1006/cyto.1996.0066 REMOVED — this is a completely wrong paper (IL-2/macrophage paper in Cytokine). Correct citation is PMID:8761287, Oncogene 1996, 13(4):665–675. No Crossref DOI exists for the BCL-W cloning paper.
2. Hinds 2003 DOI: 10.1093/emboj/cdg186 → 10.1093/emboj/cdg144 — cdg186 maps to an Ikaros regulatory elements paper. Correct paper is “The structure of Bcl-w reveals a role for the C-terminal residues in modulating biological activity.”
3. Hinds 2003 method: “crystal structure” → “NMR solution structure” — the paper reports NMR spectroscopy data (Bcl-wΔC10); there is no X-ray crystallography in this paper.
4. Hinds 2003 autoinhibition mechanism: “BH4 domain folds back upon BH3-binding groove” → corrected to “C-terminal α9 helix occupies the hydrophobic binding groove” — the autoinhibitory element is the C-terminal tail, not the BH4 domain.
5. Chen 2005 BCL-W BAD IC50: ~5 nM → 30 nM (Fig 3A: BCL-w ΔC29 vs BadBH3 = 30 nM)
6. Chen 2005 BCL-W BID IC50: ~5 nM → 40 nM (Fig 3A: BCL-w ΔC29 vs BidBH3 = 40 nM)
7. Chen 2005 BCL-W BMF IC50: ~5 nM → 9.8 nM (Fig 3A: BCL-w ΔC29 vs BmfBH3 = 9.8 nM)
8. Chen 2005 BCL-W HRK IC50: ~14 nM → 49 nM (Fig 3A: BCL-w ΔC29 vs HrkBH3 = 49 nM)
9. Yosef 2016 senescence inducer for key siRNA experiment: “oncogene-induced senescence (OIS)” → “DNA damage-induced senescence (DIS/etoposide)”; OIS cells were used for ABT-737/pharmacological experiments but the Fig 2c BCL-W + BCL-xL dual siRNA knockdown used DIS cells.
10. Chen 2005 assay method: “fluorescence polarization assay” → “surface plasmon resonance (Biacore biosensor) competition assay”
11. Navitoclax Ki source: removed incorrect attribution to Yosef 2016; corrected to “per Tse et al. 2008” (Yosef 2016 is a biology/siRNA paper that does not report Ki values)
12. A1331852 section header and body: “partial but incomplete senolytic in IMR-90” → “senolytic in HUVECs and IMR-90, not preadipocytes”; added new contradictory-evidence flag for A1331852 drug vs BCL-xL siRNA discrepancy across Zhu 2017 vs Zhu 2016

Downstream pages to check:

• studies/chen-2005-bh3-affinities.md — may have inherited the wrong BCL-W IC50 values for BAD, BID, BMF, HRK; needs correction
• molecules/compounds/navitoclax.md — check that Ki values are sourced to Tse 2008, not Yosef 2016
• molecules/proteins/bcl-xl.md, molecules/proteins/bcl-2.md — may reference the Chen 2005 binding data; check BCL-W values
• interventions/pharmacological/senolytics.md — may describe A1331852 as “partial” senolytic in IMR-90; correct to “senolytic in HUVECs and IMR-90”

[2026-05-04] verify | atg10

Pages verified: 1

• molecules/proteins/atg10.md — corrections: 4 (Cys numbering discrepancy resolved, PDB accession corrected)

Sources verified against local PDF:

• Yamaguchi 2012 (10.1016/j.str.2012.04.018) — downloaded (OA bronze) and read in full; PDF at
• UniProt Q9H0Y0 (human ATG10) — confirmed via REST API 2026-05-04: 220 aa, active-site Cys166 (ECO:0000250)
• UniProt Q9H1Y0 (human ATG5) — confirmed via REST API 2026-05-04: 275 aa, cross-link at Lys130 (ECO:0000305)

Sources unverifiable (closed-access):

• Mizushima 2002 (10.1016/s0014-5793(02)03739-0) — not_oa; yeast E2 identification claims retained as consistent with verified Yamaguchi 2012
• Nakatogawa 2013 (10.1042/bse0550039) — not_oa (review); conjugation-system overview claims retained; consistent with Yamaguchi 2012 primary data

Corrections made:

1. PDB accession “4GSJ” → “2LPU” (KmAtg10 NMR solution structure) + “3VQI” (KmAtg5 X-ray crystal structure). PDB 4GSJ is an unrelated ATG7 NTD mutant crystal structure.
2. Active-site Cys numbering discrepancy fully resolved: KmAtg10 (K. marxianus) = Cys116; ScAtg10 (S. cerevisiae) = Cys133 (both confirmed by Yamaguchi 2012 Fig. 3C); human ATG10 = Cys166 (confirmed by UniProt Q9H0Y0 REST API). Yamaguchi 2012 does not provide human residue mapping — the human number derives solely from UniProt annotation.
3. Discovery and structure section rewritten: corrected from “crystal structure of KmAtg10” to “NMR solution structure of KmAtg10 (PDB 2LPU)” and added detail on KmAtg5 crystal structure (PDB 3VQI, 2.5 Å); added KmAtg5 acceptor Lys145 and the mechanistic role of KmAtg10 Tyr56/Asn114 in substrate orientation.
4. Footnote [^yamaguchi2012] updated: added full author list, corrected PDB IDs, added local DOI lookup, added species-specific active-site Cys numbers.
5. needs-canonical-id (active-site Cys) resolved and removed.

Downstream pages updated:

• molecules/proteins/atg5.md line 55: “Cys133 in human ATG10” → “Cys166 (human ATG10)” with note that Cys133 is the S. cerevisiae ScAtg10 residue (Yamaguchi 2012 Fig. 3C)

[2026-05-04] verify | grb2

Pages verified: 1 (partial scope — 2 of 5 sources verified against local PDF; 3 unverifiable)

• molecules/proteins/grb2.md — corrections: 4

Sources verified against local PDF:

• Cheng 1998 (10.1016/s0092-8674(00)81702-x) — local PDF, read in full
• Proctor 2007 (10.1161/atvbaha.106.134007) — downloaded (OA) and read in full

Sources unverifiable:

• Lowenstein 1992 (10.1016/0092-8674(92)90167-b) — bronze OA, download failed (no PMC mirror); status now: failed
• Pawson 1995 (10.1038/373573a0) — not_oa
• Buday 1993 (10.1016/0092-8674(93)90146-h) — not_oa

Corrections made:

1. Lethality timing “E4.0–E5.5” removed — not stated in Cheng 1998. Paper documents no −/− embryos recovered from E7.5 to birth; arrest described as post-implantation (~E4.5) due to endoderm failure. Rewritten to accurately reflect paper’s evidence.
2. Cheng 1998 footnote “model: Grb2−/− C57BL/6 background” corrected → R1 ES cells (129 background); chimeric experiments used ICR recipient females. C57BL/6 is the Proctor 2007 background, not Cheng 1998.
3. Cheng 1998 footnote description rewritten: removed unsupported “before gastrulation (E4.0–E5.5)” language; added accurate description of endoderm failure, chimeric analysis, and MT-tumor haploinsufficiency result actually present in the paper.
4. Proctor 2007 footnote updated with actual quantitative lesion data from paper (aortic sinus: 0.151±0.075 vs 0.263±0.097 mm², P=0.003; n values; oxLDL uptake p-values); archive status updated from “pending” to “local PDF available (completed).”
5. Conditional KO paragraph: removed erroneous [^cheng1998] citation for T-cell/B-cell/hepatocyte/intestinal phenotypes (not in Cheng 1998); attributed to subsequent literature.

Unverifiable claims (sourced to not_oa papers, not checked):

• Domain residue ranges: SH3-N 1-58; SH2 60-152; SH3-C 156-215 (attributed to Lowenstein 1992 — unverifiable)
• SH2 YxN motif specificity (attributed to Pawson 1995 review — unverifiable)
• sem-5/Drk orthology claims (Pawson 1995 — unverifiable)
• Buday 1993 GRB2-SOS complex biochemical results (not_oa — unverifiable)
• UniProt P62993 canonical fields (217 aa, domain positions) — not independently re-checked

Downstream pages for main agent:

• molecules/proteins/insr.md, molecules/proteins/igf1r.md, molecules/proteins/irs-1.md — conditional KO citation correction does not affect these pages directly; no propagation needed
• Proctor 2007 quantitative data is new (was not on any other page); no propagation needed

[2026-05-04] verify | 14-3-3 family page

Pages verified: 1

• molecules/proteins/14-3-3.md — corrections: 5

Corrections made:

1. Mode II consensus rewritten: wiki had “RX{1-2}XXpSXP” — Yaffe 1997 defines Mode II as R·X·[Ar/S]·[+]·pS·[L/E/A/M]·P (distinct 7-position motif).
2. Mode III attribution flagged: Yaffe 1997 defines only Mode I and Mode II; Mode III label is post-1997. Added attribution note and gap tag.
3. Phosphate contact Tyr-128 added to Lys49/Arg56/Arg127 — Yaffe 1997 explicitly names all four coordinating residues.
4. CRM1 over-attribution to Brunet 1999 corrected in FOXO3 body text, FOXO3 phosphosite table, and FOXO1 section. Brunet 1999 shows Ser315 promotes nuclear exclusion but does not name CRM1; qualified with note.
5. TFEB/Settembre 2012 attribution sharpened: Settembre 2012 established Ser142 as primary mTORC1 kinase-assay site and Ser211 as secondary localization site; 14-3-3 binding at Ser211 is from Roczniak-Ferguson 2012. Text rewritten to correctly partition.

Pages unverifiable (closed-access):

• Roczniak-Ferguson 2012 (10.1126/scisignal.2002790) — not_oa; claims retained as abstract-confirmed + verified on tfeb.md
• Aitken 2006 (10.1016/j.semcancer.2006.03.005) — not_oa; review claims retained with gap tags

Downstream pages for main agent to review:

• molecules/proteins/foxo1.md — inherits CRM1/Ser319 claim; check when foxo1.md is seeded
• molecules/proteins/tfeb.md — Settembre/Roczniak attribution partition should be cross-checked

[2026-05-04] ingest | round-10b-tier-a-entities

grb2

• added: molecules/proteins/grb2.md
• entity type: protein (type: protein)
• canonical IDs: UniProt P62993 (confirmed via REST API, 217 aa); NCBI Gene 2885; HGNC 4566; GenAge: not found (null, needs-canonical-id)
• DOIs cited (5 total):
  • 10.1016/0092-8674(92)90167-b (Lowenstein 1992, GRB2 original cloning; 1,644 citations; archive: pending, OA bronze)
  • 10.1038/373573a0 (Pawson 1995, protein modules review; 2,433 citations; archive: not_oa)
  • 10.1016/0092-8674(93)90146-h (Buday & Downward 1993, GRB2-SOS complex and RAS activation; 1,077 citations; archive: not_oa)
  • 10.1016/s0092-8674(00)81702-x (Cheng et al. 1998, Grb2 KO mouse; 376 citations; archive: LOCAL PDF AVAILABLE — completed)
  • 10.1161/atvbaha.106.134007 (Proctor et al. 2007, Grb2+/- atheroprotection; 27 citations; archive: pending, OA bronze)
• DOI corrections (2 critical):
  • Brief supplied Cheng 1998 as 10.1016/s0092-8674(00)80939-2 — did NOT resolve in archive; correct DOI confirmed as 10.1016/s0092-8674(00)81702-x via Crossref (Cell, 1998, Cheng/Saxton/Pawson, “Mammalian Grb2 Regulates Multiple Steps in Embryonic Development and Malignant Transformation”)
  • Brief supplied Buday 1993 as 10.1016/0092-8674(93)90412-h — did NOT resolve in archive; correct DOI confirmed as 10.1016/0092-8674(93)90146-h via Crossref (Cell, 1993, Buday & Downward, “Epidermal growth factor regulates p21ras through the formation of a complex of receptor, Grb2 adapter protein, and Sos nucleotide exchange factor”)
• gaps surfaced:
  • No mammalian aging lifespan studies for GRB2; all aging-adjacent evidence is invertebrate (sem-5/C. elegans) or cardiovascular-disease (Proctor 2007) — needs-human-replication
  • GRB2 nuclear localization annotated in UniProt but function unknown — no-mechanism
  • GenAge entry for GRB2 not found — needs-canonical-id
• implicit stubs created (new wikilinks to non-existent pages): ras-mapk, sos1, sos2, shc, gab1, cbl
• downstream propagation candidates (main agent handles):
  • pathways/insulin-igf1.md — GRB2 now has a page; consider adding it to key-nodes list
  • molecules/proteins/irs-1.md — GRB2-pTyr895 (YxN) binding site; could add cross-link
  • molecules/proteins/insr.md — GRB2 direct binding; cross-link opportunity
  • molecules/proteins/igf1r.md — GRB2 direct binding; cross-link opportunity

tbk1

• added: molecules/proteins/tbk1.md
• entity type: protein (type: protein)
• canonical IDs: UniProt Q9UHD2 (confirmed via REST API); NCBI Gene 29110; HGNC 11584
• DOIs cited (7 total; all confirmed in archive via ):
  • 10.1093/emboj/18.23.6694 (Pomerantz 1999, TBK1 cloning; 583 citations; dl: pending)
  • 10.1084/jem.20040520 (Hemmi 2004, TBK1 in IFN signaling; 572 citations; dl: pending)
  • 10.1126/science.1205405 (Wild 2011, OPTN-Ser177 phosphorylation; already verified on optn.md; dl: not_oa)
  • 10.1016/j.molcel.2015.08.016 (Heo 2015, PINK1-Parkin-OPTN/NDP52-TBK1 mitophagy; 844 citations; dl: pending)
  • 10.1126/science.aaa3650 (Cirulli 2015, ALS exome sequencing; 970 citations; dl: pending)
  • 10.1038/nn.4000 (Freischmidt 2015, TBK1 haploinsufficiency ALS/FTD; 764 citations; dl: pending)
  • 10.1038/nm.3082 (Reilly 2013, amlexanox TBK1/IKKε obesity; 428 citations; dl: pending)
• DOI error corrected: task brief supplied 10.1038/nature14488 for “Liu 2015 TBK1 autophagy”; this DOI resolves to a phytoplankton ecology paper — excluded from page; unsourced for Liu 2015 claim
• no PDFs locally available (all pending); Wild 2011 is not_oa; page carries verified: false throughout
• key sections: identity, domain table, activation mechanism, innate immunity role (cGAS-STING axis), selective autophagy role (OPTN/NDP52/p62 phosphorylation table), ALS/FTD neurodegeneration genetics, pharmacology (amlexanox + preclinical inhibitors), pathway membership, interactors, limitations & gaps
• implicit stubs created: cgas-sting (planned R10c), sting, irf3
• gaps surfaced:
  • Liu 2015 TBK1 reciprocal autophagy paper: DOI in brief was wrong; correct DOI not confirmed — flagged unsourced
  • TBK1 haploinsufficiency aging phenotype in mice not comprehensively published — needs-human-replication
  • scaffolding vs catalytic function dissection incomplete — no-mechanism
  • amlexanox dose-response in metabolic aging context unclear — dose-response-unclear
• downstream propagation candidates (main agent handles):
  • molecules/proteins/optn.md — confirm TBK1-Ser177 cross-link wording consistent
  • molecules/proteins/ndp52.md — confirm TBK1 substrate language consistent
  • molecules/proteins/p62.md — confirm Ser403 phospho-site attribution
  • processes/xenophagy.md — TBK1 now has a dedicated page; update cross-link if needed
  • processes/mitophagy.md — TBK1 now has a dedicated page; update cross-link if needed
  • pathways/pink1-parkin-pathway.md — TBK1 as downstream effector; update cross-link
  • pathways/cgas-sting.md (planned R10c) — TBK1 is central kinase; ensure TBK1 page listed as key-node

brain

• added: tissues/brain.md
• entity type: tissue (type: tissue — not formally defined in CLAUDE.md; used per user instruction; schema gap escalated: user should formally add type: tissue to CLAUDE.md)
• structure: anchor tissue stub (~150 lines); synthesis over cross-linked verified atomic pages; no novel quantitative claims extracted; numerics delegated via wikilinks to microglia (verified-partial), atg5/atg7 (verified-partial), neurodegeneration (drafted)
• canonical-DB IDs: none (tissue pages have no equivalent canonical-DB identifier)
• primary DOIs cited: 10.1002/cne.21974 (Azevedo 2009, not_oa confirmed), 10.1038/ncb1889 (Hara 2006, local PDF confirmed), 10.1038/nature04724 (Komatsu 2006, local PDF confirmed), 10.1038/s41591-023-02305-8 (Bouzid 2023, not confirmed in archive — verify on next lint pass)
• implicit stubs created: neurons, astrocytes, oligodendrocytes, neural-stem-cells, glymphatic-system, spinal-cord
• gaps surfaced: cortical thinning rate, BBB permeability trajectory, astrocyte A1/A2 prevalence — unsourced; human adult hippocampal neurogenesis magnitude — contradictory-evidence
• ROADMAP: brain marked drafted
• verification priority: Azevedo 2009 (~86 billion neuron count — the one quantitative primary-source claim on the page not already verified on an atomic page)

bone-marrow

• added: tissues/bone-marrow.md
• entity type: tissue (type: tissue — same schema choice and escalation as brain)
• structure: anchor tissue stub (~160 lines); all quantitative HSC aging data delegated to hematopoietic-stem-cells (verified-partial)
• canonical-DB IDs: none
• primary DOIs cited (all cross-referenced from hematopoietic-stem-cells verified page): 10.1084/jem.192.9.1273 (Sudo 2000, local PDF), 10.1073/pnas.1000834107 (Beerman 2010, from HSC page), 10.1056/NEJMoa1408617 (Jaiswal 2014, local PDF; CHD CI 1.2–3.5 consistent with chronic-inflammation.md verified correction), 10.1073/pnas.1116110108 (Pang 2011, from HSC page)
• implicit stubs created: mesenchymal-stem-cells, osteoblasts, osteoclasts, clonal-hematopoiesis, myelodysplastic-syndrome
• gaps surfaced: BM adipocyte-HSC direction (contradictory-evidence); red-to-yellow marrow conversion rate (unsourced); senolytic niche-rejuvenation efficacy (needs-replication)
• ROADMAP: bone-marrow marked drafted; Tier A implicit-stub queue updated
• verification priority: all quantitative claims inherited from already-verified atomic pages; no independent re-verification needed on this stub

atg10

• added: molecules/proteins/atg10.md
• entity type: protein (type: protein)
• canonical IDs: UniProt Q9H0Y0 (220 aa, Cys166 active site; accessed 2026-05-04); NCBI Gene 83734; HGNC 21073
• DOIs cited (3 total; all confirmed in archive via ):
  • 10.1016/s0014-5793(02)03739-0 (Mizushima 2002, FEBS Letters; mouse Apg10 as Apg12-conjugating E2; 47 citations, 98th percentile; not_oa)
  • 10.1016/j.str.2012.04.018 (Yamaguchi 2012, Structure; K. marxianus Atg10 crystal structure PDB 4GSJ; 76 citations, 99th percentile; archive status: pending download, bronze OA URL available)
  • 10.1042/bse0550039 (Nakatogawa 2013, Essays Biochem; review of two Ubl conjugation systems; 305 citations, 100th percentile; not_oa)
• note: task brief DOIs for Nemoto 2003 (10.1074/jbc.M306626200) and Yamaguchi 2012 (10.1074/jbc.M111.330456) — both returned 404 from Crossref and not found in archive; Yamaguchi 2012 Structure journal DOI (10.1016/j.str.2012.04.018) confirmed as correct citation per PubMed PMID 22682742 efetch; Nemoto 2003 human ATG10 paper not confirmed in archive and excluded to avoid BUG-2 DOI fabrication
• key sections: identity table, domain structure + active-site discrepancy flag, full three-enzyme cascade (ATG7→ATG10→ATG5 isopeptide), contrast with LC3/ATG3 system, ATG16L1 downstream, discovery/structure, cross-autophagy-subtype table (CMA excluded), KO phenotypes, aging-relevance, interactor table, extrapolation table, limitations
• active-site Cys discrepancy surfaced: UniProt Q9H0Y0 reports Cys166 (human); task brief and atg5 page (verified) cite Cys133 — likely yeast-to-human mapping confusion (yeast Atg10 Cys133 conflated with human numbering); tagged needs-canonical-id; wiki-verifier must cross-check Yamaguchi 2012 and correct atg5.md if needed
• implicit stubs created: atg3 (planned R10c), atg16l1 (referenced, no page)
• gaps surfaced:
  • needs-canonical-id: ATG10 active-site Cys number (Cys166 UniProt vs Cys133 in atg5.md)
  • needs-replication: no extensive Atg10-specific constitutive KO mouse phenotype published
  • no-mechanism: ATG10 inclusion in TFEB CLEAR gene battery not specifically cited with primary ChIP-seq data
  • needs-human-replication: ATG10 pathogenic variant functional data absent (Arpna 2026 is bioinformatic only)
• downstream propagation candidates (main agent handles):
  • molecules/proteins/atg5.md — Cys133 active-site ATG10 reference (line ~55) should be corrected to Cys166 after wiki-verifier resolves against Yamaguchi 2012
  • molecules/proteins/atg7.md — references ATG10 as E2 (line ~53); verify active-site numbering consistency
  • molecules/proteins/atg12.md — references ATG10 reaction; verify consistent with new atg10.md framing
  • processes/autophagy.md — may reference ATG10 conjugation step; check consistency

xiap

• added: molecules/proteins/xiap.md
• entity type: protein (type: protein)
• canonical IDs: UniProt P98170 (confirmed via REST API — 497 aa, BIRC4_HUMAN); NCBI Gene 331; HGNC 592
• DOIs cited (7 total; confirmed in archive via or PubMed efetch):
  • PMID:8654366 / no CrossRef DOI (Duckett et al. 1996, EMBO J, IAP family discovery; confirmed via PubMed efetch)
  • 10.1038/379349a0 (Liston et al. 1996, Nature; 1016 citations; local PDF available)
  • 10.1038/40901 (Deveraux et al. 1997, Nature; 1942 citations; local PDF available)
  • 10.1016/s0092-8674(01)00274-4 (Riedl et al. 2001, Cell; 784 citations; local PDF available)
  • 10.1016/s1097-2765(03)00054-6 (Shiozaki et al. 2003, Mol Cell; 723 citations; download: pending)
  • 10.1016/j.cell.2007.10.037 (Vince et al. 2007, Cell; 1054 citations; download: pending)
  • 10.1182/blood-2010-01-256099 (Marsh et al. 2010, Blood; 254 citations; closed — no-fulltext-access)
• DOI correction: task brief cited Riedl 2001 for “BIR3-caspase-9 structure” but that DOI resolves to “Structural Basis for the Inhibition of Caspase-3 by XIAP.” Caspase-9 BIR3 structural mechanism correctly attributed to Shiozaki 2003 (PMID:12620238). Both papers cited with accurate attributions.
• DOI correction: task brief Marsh 2010 DOI 10.1182/blood-2010-04-279010 (not found). Correct DOI 10.1182/blood-2010-01-256099 confirmed via PubMed efetch and DOI lookup.
• key sections: identity, domain table, caspase inhibition mechanisms (BIR2-linker for caspase-3/-7; BIR3 for caspase-9; RING E3), SMAC/DIABLO antagonism, IAP family context, discovery history, aging/senescence SCAP context, XLP-2 disease genetics, SMAC mimetic therapeutics table
• implicit stubs created: smac-diablo (planned R10c), caspase-7 (planned R10c)
• gaps surfaced: needs-replication (XIAP in senescent cells), needs-human-replication (SMAC mimetic senolytics), unsourced (S-nitrosylation function; Omi cleavage; tumor overexpression), no-mechanism (XIAP vs BCL-2 family relative contribution), no-fulltext-access (Marsh 2010), needs-canonical-id (Ensembl ID)
• suggested verification priority for wiki-verifier:
  1. Riedl 2001 (local PDF) — BIR2-linker steric mechanism; Glu167 mutation claim
  2. Deveraux 1997 (local PDF) — direct caspase-3/-7 inhibition; apoptosome override statement
  3. Liston 1996 (local PDF) — IAP family discovery; hILP identity
  4. Marsh 2010 (closed-access) — XLP-2 HLH reclassification; cannot verify without PDF
• downstream propagation candidates (main agent handles):
  • pathways/apoptosis-pathway.md — add XIAP as key-node if not listed
  • molecules/proteins/caspase-3.md — XIAP BIR2-linker inhibition cross-link
  • molecules/proteins/caspase-9.md — XIAP BIR3 inhibition; Shiozaki 2003 as structural citation
  • hallmarks/cellular-senescence.md — consider adding XIAP to SCAP section alongside BCL-2 family

[2026-05-04] verify | dysbiosis hallmark synthesis-MOC

• verified: hallmarks/dysbiosis.md
• verification type: synthesis-MOC — 3 of 4 cited DOIs read in full from local PDFs; 1 confirmed closed-access (not_oa); Depommier 2019 downloaded from PMC during this session and read in full
• PDFs read: Claesson 2012 (Nature 488:178-184, 8 pp, full text), Smith 2017 (eLife 6:e27014, 26 pp, full text), Depommier 2019 (Nat Med 25:1096-1103 + extended data, 28 pp PMC author manuscript, full text)
• unverifiable: López-Otín 2023 (10.1016/j.cell.2022.11.001) — closed-access, download status “failed”; framing claims (dysbiosis as standalone hallmark in 2023 expansion) carry no-fulltext-access in footnote
• corrections made (5 factual):
  1. Depommier 2019 group n values wrong: wiki footnote had “placebo n=9, live n=9, pasteurized n=14” → corrected to “placebo n=11, pasteurized n=12, alive/live n=9” per Extended Data Fig. 1 flow chart and per-figure n statements throughout paper
  2. Depommier 2019 table entry n wrong: “n=32 randomized” → “n=40 enrolled/n=32 completed”; added dose (10^10 cells/day); added key null finding (overall gut microbiome community structure NOT significantly changed by supplementation)
  3. Claesson 2012 footnote gap tag removed: footnote carried #gap/no-fulltext-access — PDF available locally but quantitative claims not verified; removed now that PDF has been read in full; footnote expanded with full n breakdown, age range, key quantitative results (butyrate/SCFA higher in community, SCFA gene metagenomics, Monte-Carlo P<0.0001 for Procrustes analysis), and diet-temporal-precedence finding
  4. Claesson 2012 body text — diet causality nuance added: wiki omitted the paper’s central interpretive finding that diet temporal precedence over microbiome (diet shifts <1 month, microbiome ~1 year) supports diet→microbiota→health causation rather than frailty→dysbiosis; this is load-bearing for interpreting the cross-sectional association
  5. Smith 2017 body text — experimental controls expanded: wiki described only Ymt vs wt comparison; added explicit statement that Omt (same-age controls) did NOT live longer than untreated wt (the key specificity control), the antibiotic cocktail composition, the Ymt vs Omt and Ymt vs Abx p-values, the locomotor activity BH-adjusted p-value, and the GRZ strain identifier; updated Smith 2017 footnote to match
• confirmed correct (no corrections): Smith 2017 +37% lifespan figure; p=4.04×10⁻⁹; +41% vs Omt; donors 6-wk old; recipients 9.5-wk old; locomotor activity maintained at 16 wk; Depommier 2019 insulin sensitivity +28.62±7.02% P=0.002; cholesterol -8.68±2.38% P=0.02; insulinemia -34.08±7.12% P=0.006; 3-month duration; double-blind RCT design; ELDERMET n=178; age mean 78±8; diversity-frailty inverse correlation
• downstream propagation needed (main agent task):
  • studies/claesson-2012-eldermet-microbiota-elderly.md — if it exists, check whether its design/n/key-results fields match the corrected footnote details now on this page
  • studies/depommier-2019-akkermansia-human-rct.md — if it exists, correct group n values (placebo 11, pasteurized 12, alive 9) and add key null finding (microbiome community structure unchanged) and dose information
  • molecules/compounds/akkermansia-muciniphila.md or equivalent stub — if seeded, check n values inherited from Depommier 2019

[2026-05-04] verify | chronic-inflammation synthesis-MOC

• verified: hallmarks/chronic-inflammation.md
• verification type: synthesis-MOC with primary PDF verification — 3 of 7 cited DOIs read in full from local PDFs; 2 confirmed not_oa; 1 confirmed closed/download-failed (López-Otín 2023); 1 download attempted but failed (Lancet HTTP 403)
• PDFs read in full: Rodier 2009 (10.1038/ncb1909), CANTOS NEJM/Ridker 2017 (10.1056/NEJMoa1707914), Jaiswal 2014 (10.1056/NEJMoa1408617)
• corrections made (2 factual):
  1. Jaiswal 2014 CHD CI: body text HR CI “1.2–3.4” → “1.2–3.5” — paper states 95% CI 1.2 to 3.5, p=0.02 (page 2496)
  2. Jaiswal 2014 footnote expanded: CI for CHIP prevalence ranges added (5.6% CI 5.0–6.3 aged 60–69; 18.4% CI 12.1–27.0 aged 90+); CHD p=0.02 added; hematologic malignancy CI (3.9–32.6) added; CHD CI corrected 1.2–3.4 → 1.2–3.5
• confirmed correct (no corrections needed):
  • CANTOS NEJM: HR 0.85 (95% CI 0.74–0.98, p=0.021) at 150 mg — confirmed from abstract and Fig 2B
  • CANTOS fatal infection: 0.31 vs 0.18/100 person-years — confirmed from abstract
  • CANTOS 300 mg p=0.031 vs threshold p=0.01058 — confirmed: paper states “threshold P value for the primary end point were 0.01058 for the 300-mg dose” and HR 0.86, p=0.0314 (Table 2 and Fig 2C)
  • Rodier 2009 two-arm SASP (growth arrest = p53/p21-dependent; SASP = ATM/NBS1/CHK2-dependent, p53-independent) — confirmed from Figs 3–4 and text
  • Rodier 2009 IL-6 50-fold and IL-8 10-fold ATM-dependent suppression — confirmed from text page 977
  • Jaiswal 2014 n=17,182; CHIP ~10% over 70; hematologic malignancy HR=11.1 — confirmed
• unverifiable sources:
  • Franceschi 2000 (10.1111/j.1749-6632.2000.tb06651.x): not_oa — no-fulltext-access retained
  • Childs 2016 (10.1126/science.aaf6659): not_oa — no-fulltext-access retained; ~60% plaque reduction claim not independently verified
  • López-Otín 2023 (10.1016/j.cell.2022.11.001): closed-access, download failed — hallmark-elevation framing not independently verified
  • Ridker 2017 Lancet (10.1016/S0140-6736(17)32247-X): OA URL HTTP 403, download failed — cancer mortality HR 0.49, lung cancer HRs, incidence rates NOT independently verified; footnote updated to reflect failure; no-fulltext-access added
• banner: ⚠️ removed; verified: true set with verified-scope
• downstream propagation needed (NOT done here — main agent task):
  • molecules/proteins/jaiswal-2014 or any page citing the CHD HR 1.2–3.4 may have the wrong CI upper bound (should be 3.5); check hematopoietic-stem-cells for inherited Jaiswal 2014 CHD CI

[2026-05-04] verify | stem-cell-exhaustion synthesis-MOC

• verified: hallmarks/stem-cell-exhaustion.md
• verification type: synthesis-MOC cross-check — all inherited quantitative claims cross-checked against verified atomic pages; no primary PDFs re-read for this MOC (all primary sources verified on atomic pages)
• atomic pages consulted: hematopoietic-stem-cells (verified), satellite-cells (verified), microglia (verified-partial), cardiomyocytes (verified), anemia-of-aging (verified), lopez-otin-2013-hallmarks-of-aging (verified)
• wikilink resolution: all major wikilinks resolve to existing pages; implicit stubs correctly noted as such (notch-pathway, p16-rb-pathway, clonal-hematopoiesis, hyperfunction-theory)
• corrections made (2 factual):
  1. Pang 2011 n groups swapped: “n=13 young + 11 elderly” → “n=11 young + 13 elderly” — body text of verified hematopoietic-stem-cells.md states 11 young + 13 elderly (per primary PDF); MOC had the numbers transposed
  2. Lewis-McDougall 2019 SASP list truncated: “(PAI-1, IL-8, IL-6)” → “(PAI-1, IL-8, IL-6, IGFBP-3, GM-CSF, CCL-2)” — cardiomyocytes.md (verified) lists all six SASP factors; MOC had only three
• confirmed correct (no corrections): Sudo 2000 ~17× expansion; Beerman 2010 two subsets only; Jaiswal 2014 HRs; Genovese 2014 n; Conboy 2005 GFP+/P-value; Hammond 2019 cluster names/n; Lewis-McDougall 2019 CPC senescence %; Guralnik 2004 prevalence and etiologic fractions
• internal inconsistency flagged (main agent task): hematopoietic-stem-cells.md body text says Pang 2011 n=11 young/13 elderly but its own footnote says n=13 young/11 elderly; MOC updated to match body text; footnote inconsistency requires main-agent review against Pang 2011 primary source
• banner: removed; verified: true set with verified-scope describing MOC-to-atomic-page scope

[2026-05-04] verify | telomere-attrition synthesis-MOC

• verified: hallmarks/telomere-attrition.md
• verification type: synthesis-MOC with primary PDF verification — 3 of 7 cited DOIs read in full from local PDFs; 2 confirmed not_oa; 1 pending (de Lange 2005 review); 1 confirmed via abstract (Harley 1990 not_oa)
• PDFs read: Prowse & Greider 1995 (PNAS 92:4818–4822), Celli & de Lange 2005 (Nat Cell Biol 7:712–718), Bernardes de Jesus 2012 (EMBO Mol Med 4:691–704)
• wikilink resolution: all wikilinks resolve to existing pages or correctly noted as implicit stubs; p53-pathway verified-partial label corrected to “(verified-partial)” — dna-damage-response and p53-pathway are both verified: true per frontmatter; cross-references confirmed accurate
• corrections made (3 factual, 1 completion):
  1. Celli & de Lange 2005 model: “human cells” → “mouse embryo fibroblasts (MEFs; mixed 129/BL6 background)” — paper explicitly states MEFs derived from E12.5–E13.5 embryos using 129P2/Ola ES cell line; no human cells used
  2. Prowse 1995 strain list: body text and footnote corrected — primary tissue telomerase assays used BALB/c M. musculus and M. spretus; C57BL/6 mentioned only for limit-mobility TRF data, not tissue-specific telomerase assays
  3. Mouse telomere length “~50 kb”: corrected to “>30 kb (limit-mobility TRF; subsequent literature commonly ~40–60 kb)” with unsourced — Prowse 1995 states only “>30 kb limit mobility,” not the specific ~50 kb figure; exact kb value requires a separate citation
  4. Bernardes 2012 n per group: filled from PDF — 1-yr cohort: eGFP n=12, mTERT n=21, control n=43; 2-yr cohort: eGFP n=14, mTERT n=23, control n=29; p-values filled (p<0.05 Log Rank); cancer non-significance filled (p=0.87); no-fulltext-access removed; metabolic endpoints specified
• unverifiable claims remaining:
  • Harley 1990 (~50–150 bp/division, ~50 doublings): not_oa; no-fulltext-access retained
  • Cawthon 2003 (3.18× HR, 8.54× HR, n=143): not_oa; abstract-confirmed; no-fulltext-access retained for full adjustment details
  • de Lange 2005 Genes Dev review (shelterin subunit list): pending download; shelterin section claims not independently verified from this specific source
• downstream propagation needed: none — Celli 2005 is not cited on any other wiki page; the “human cells” error was local to this page only (confirmed by grep)

[2026-05-04] verify | deregulated-nutrient-sensing synthesis-MOC

• verified: hallmarks/deregulated-nutrient-sensing.md
• verification type: synthesis-MOC cross-check — all inherited quantitative claims cross-checked against verified atomic pages; no primary PDFs read for this MOC page itself (all primary sources are verified on atomic pages)
• wikilink resolution: all 47 wikilinks resolve to existing pages; 3 planned stubs (mtor-kinase, jak-stat-pathway, antagonistic-pleiotropy) correctly noted in body as not-yet-seeded
• claims checked against atomic pages (all CONFIRMED correct — no quantitative corrections needed):
  • Holzenberger 2003: +26% combined (Cox p<0.02); +33% females (p<0.001); +15.9% males (NS); 129/Sv background — confirmed via igf1r (verified-partial)
  • Suh 2008: IGF1R LoF alleles 2.3% vs 0.3% controls p=0.02; paradoxically elevated circulating IGF-1; n=384 centenarians — confirmed via igf-1 (verified) and igf1r (verified-partial)
  • Hertweck 2004: sgk-1 LoF +63% (14.7→24.0 d; p<0.0001; n=147); akt-1 p=0.1642 NS; akt-2 p=0.3717 NS — confirmed via sgk1 (verified)
  • Harrison 2009: 14 mg/kg food = 2.24 mg/kg BW/day; mean +13% females / +9% males; 90th-pct +14%/+9% — confirmed via rapamycin (verified-partial)
  • Willcox 2008: OR=2.75 (CI 1.51–5.02; p=0.0007); HHP/HAAS cohort; Honolulu, Oahu; nested case-control n=615; male-only — confirmed via foxo3 (verified)
  • Brunet 1999: Thr32 + Ser253 = 14-3-3 binding sites; Ser315 = CRM1 export (NOT 14-3-3) — confirmed via foxo3 (verified)
  • Harrison 2009 rapamycin dose notation: “14 mg/kg food” correctly annotated as NOT 14 mg/kg body weight — confirmed via rapamycin (verified-partial)
• status annotations updated (13 pages upgraded from ‘drafted’ to ‘verified-partial’):
  • igf1r (drafted → verified-partial) — Holzenberger 2003 + Suh 2008 verified
  • pten (drafted → verified-partial) — Ortega-Molina 2012 + Garcia-Cao 2012 verified
  • akt (drafted → verified-partial) — Brunet 1999 + Hertweck 2004 verified
  • foxo4 (drafted → verified-partial) — Baar 2017 + Hosaka 2004 + Tothova 2007 verified
  • foxo1 (drafted → verified-partial) — Hosaka 2004 + Brunet 1999 verified
  • insr (drafted → verified-partial) — Ullrich 1985 + Accili 1996 + Brüning 1998 verified
  • irs-1 (drafted → verified-partial) — Um 2004 + Boucher 2014 verified
  • irs2 (drafted → verified-partial) — Sun 1995 + Boucher 2014 verified
  • pi3k (drafted → verified-partial) — Engelman 2006 + Vanhaesebroeck 2010 + Manning 2017 verified
  • pdk1 (drafted → verified-partial) — Alessi 1997 + Pullen 1998 + Lawlor 2002 verified
  • insulin (drafted → verified-partial) — Bluher 2003 + Holzenberger 2003 + Suh 2008 verified
  • nmn (drafted → verified-partial) — all six primary PDFs verified
  • nr (drafted → verified-partial) — Trammell 2016 + Martens 2018 + Brakedal 2022 verified
• corrections made: 0 quantitative corrections to MOC body; 13 status-annotation upgrades only
• banner: ⚠️ removed; verified: true set
• downstream propagation flagged (NOT done here — main agent task):
  • pten body: Ortega-Molina 2012 female lifespan “+9%” → “+16%” (confirmed by cancer page verifier from PDF Figure 2A; see log line 573 and 6719). pten.md line ~98 says “+9% females” but cancer.md (verified 2026-05-04) has “+16%” as the PDF-verified value. This discrepancy is a pre-existing conflict between atomic pages requiring main-agent resolution.
  • hyperfunction-theory line 72 already reads “+12% males / +16% females” — no change needed there.
• ROADMAP updated: deregulated-nutrient-sensing promoted from ‘drafted’ to ‘verified-partial’

[2026-05-04] verify | disabled-macroautophagy synthesis-MOC

• verified: hallmarks/disabled-macroautophagy.md
• verification type: synthesis-MOC cross-check — all quantitative claims checked against verified atomic pages; four footnote DOIs verified via Crossref API
• corrections made (7):
  • [^komatsu2006] footnote DOI: 10.1083/jcb.200607037 (wrong — resolves to 404) → 10.1038/nature04723 (Komatsu 2006 Nature, “Loss of autophagy in the central nervous system causes neurodegeneration in mice”); n added (n=26 mutant / 41 control, P<0.01)
  • [^pyo2013] footnote DOI: 10.1073/pnas.1312369110 (wrong — resolves to 404) → 10.1038/ncomms3300 (Pyo 2013 Nat Commun, “Overexpression of Atg5 in mice activates autophagy and extends lifespan”)
  • [^fernandez2018] footnote DOI: 10.1038/s41586-018-0327-4 (wrong — resolves to a tropical-disease drug discovery paper) → 10.1038/s41586-018-0162-7 (Fernández 2018 Nature, Beclin-1 F121A); n and p-value added (n=68 WT + 102 KI; p<0.0001)
  • [^liang1999] footnote DOI: 10.1038/46766 (wrong — 404) → 10.1038/45257 (Liang 1999 Nature, Beclin-1 discovery)
  • ULK1 AMPK activation site: Ser555 → Ser556 (human ULK1 per UniProt O75385, confirmed on verified ulk1 page)
  • Status labels: 13 pages updated from “(drafted)” or “(verified-partial)” to “(verified)” after confirming verified: true on each atomic page (atg12, atg13, atg101, lc3, p62, autophagy, mitophagy, lipophagy, xenophagy, chaperone-mediated-autophagy, mtor, rapamycin, caloric-restriction, urolithin-a, atg7)
  • Intro sentence: “verified or verified-partial” → “verified” (all Round-6 pages now fully verified)
• unverifiable claim:
  • López-Otín 2023 framing (Primary hallmark, split from Loss of Proteostasis) — not verified against primary paper; flagged in verified-scope
• downstream propagation: no numeric corrections; DOI fixes are in the MOC footnotes only and don’t cascade to atomic pages (which carry correct DOIs already)

[2026-05-04] verify | loss-of-proteostasis synthesis-MOC

• verified: hallmarks/loss-of-proteostasis.md
• sources checked against local PDFs:
  • Komatsu 2006 (10.1038/nature04723) — ATG7 nestin-Cre KO; n=26 mutant/41 control, all dead within 28 weeks confirmed; ubiquitin-positive inclusions + massive neuronal loss by P56 confirmed; proteasome activity unaltered confirmed
  • Hara 2006 (10.1038/nature04724) — ATG5 nestin-Cre KO confirmed (not ATG7); progressive neurodegeneration, TUNEL-positive granular cell death, Purkinje cell loss confirmed
  • Pyo 2013 (10.1038/ncomms3300) — ~17.2% median lifespan extension confirmed; n=65 WT + 70 Tg (line 25) confirmed; χ²=17.32 P<0.001 confirmed; max lifespan WT 781±22 d vs Tg 900±34 d confirmed
  • Fernández 2018 (10.1038/s41586-018-0162-7) — Beclin-1 F121A KI; combined WT=26 mo, KI=29 mo, p<0.0001 confirmed; female and male sub-group data confirmed; C57BL/6 inbred background confirmed
• corrections made (2):
  • Komatsu 2006 description: “with no overt cell death at early timepoints” → corrected to accurate description: proteasome function was unaltered (not cell death); massive neuronal loss and TUNEL positivity by P56 are actually documented in the paper; n’s added (n=26 mutant, n=41 control, P<0.01)
  • [^pyo2013] footnote: missing n=65 WT + 70 Tg and χ²=17.32 P<0.001 and max lifespan values — added to match verified atg5 citation discipline
• attribution checks confirmed correct:
  • Hara 2006 = ATG5 (Atg5^flox/flox × nestin-Cre) — CORRECT
  • Komatsu 2006 = ATG7 (Atg7^flox/flox × nestin-Cre) — CORRECT
  • Zhang & Cuervo 2008 = LAMP-2A overexpression rescue (not Schneider 2014) — CORRECT (Schneider 2014 is not cited here)
  • Cuervo & Dice 2000: “>30% decline” figure absent from page (correctly omitted per CMA page correction)
• unverifiable (closed-access or download-failed):
  • Cuervo & Dice 2000 (10.1074/jbc.M002102200) — PDF download failed; qualitative claim consistent with PubMed abstract
  • López-Otín 2023 (10.1016/j.cell.2022.11.001) — closed-access, download failed; no-fulltext-access retained in footnote
  • Anckar & Sistonen 2011 (10.1146/annurev-biochem-060809-095203) — not re-read in full; HSF1 age-decline framing cross-checked as qualitative claim
• downstream propagation: no numeric corrections affect entity pages — ATG5/ATG7/Beclin-1 verified pages already carry the correct numbers; no propagation needed

[2026-05-04] verify | genomic-instability synthesis-MOC

• verified: hallmarks/genomic-instability.md — synthesis-MOC (lighter scope)
• verification type: wikilink resolution + atomic-page spot-check; no primary PDFs read (all primary claims delegate to verified atomic pages)
• corrections made:
  • added verified: true / verified-date: 2026-05-04 / verified-by: claude / verified-scope to frontmatter (field was absent)
  • removed ⚠️ auto-extraction banner per CLAUDE.md verification protocol
  • tagged [[clonal-hematopoiesis]] reference in body with #stub note — file does not exist; all other 24 wikilinks resolve correctly
• atomic-page claims spot-checked (all matched):
  • Tyner 2002 m-allele: ~24K C-terminal fragment; 23% lifespan reduction (96 vs 118 wk; P<0.0001); <6% vs >45% tumor incidence — confirmed against p53 (verified) and tyner-2002-p53-mutant-aging (verified)
  • Mendrysa 2006: mdm2puro/7-12; ~30% Mdm2; 110 vs 106 wk; p=0.61; n=28+12; NOT prematurely aged — confirmed against mdm2 (verified-partial)
  • Baker 2016: 24-27% lifespan extension; tumor latency increased, no tumor incidence reduction at autopsy — confirmed against p21 (verified)
  • Jaiswal 2014: n=17,182; HRs 11.1/1.4/2.0/2.6; per-decade prevalences 5.6/9.5/11.7/18.4% — confirmed against hematopoietic-stem-cells (verified)
  • Genovese 2014: whole-exome sequencing; n=12,380; >10% >65 yr; HR 12.9 (CI 5.8-28.7) — confirmed against hematopoietic-stem-cells (verified)
  • Rodier 2009 two-arm SASP: growth-arrest arm (ATM->CHK2->p53->p21); SASP arm (ATM/NBS1/CHK2 required; p53 NOT required) — confirmed against dna-damage-response (verified-partial)
• no factual corrections required to body text; all quantitative claims matched verified atomic sources
• broken wikilink: clonal-hematopoiesis (no file exists); listed as implicit stub below
• implicit stubs to seed (carry-forward):
  • clonal-hematopoiesis — high priority; referenced on genomic-instability, chronic-inflammation, hematopoietic-stem-cells, stem-cell-exhaustion pages

[2026-05-04] ingest | round-10a-2 | dysbiosis

dysbiosis

• updated: hallmarks/dysbiosis.md — stub → synthesis-MOC (lighter format per coverage-gap convention)
• entity type: hallmark (integrative; NEW in 2023 expansion)
• structure: synthesis-MOC; lighter than other hallmark pages (~180 lines) — intentional, reflects wiki’s most under-developed hallmark; all primary findings deferred to atomic pages via wikilinks per sops/writing-hypothesis-pages.md discipline
• atomic pages drawn on (all verified or verified-partial):
  • [[nothobranchius-furzeri]] (verified) — Smith 2017 killifish FMT; +37% lifespan; primary claim lives there
  • [[urolithin-a]] (verified-partial) — gut bacterial metabolite; Singh 2022 ATLAS RCT; full data on compound page
  • [[frailty]] (verified) — ELDERMET diversity-frailty correlation
  • [[type-2-diabetes]] (verified-partial) — dysbiosis association
  • [[caloric-restriction]] (verified-partial) — microbiome modulation
  • [[chronic-inflammation]] (stub) — LPS → inflammaging bridge
• primary DOIs cited (4 footnotes):
  • 10.1016/j.cell.2022.11.001 (López-Otín 2023) — closed-access, not locally available no-fulltext-access
  • 10.1038/nature11319 (Claesson 2012 ELDERMET) — local PDF available; not re-verified on this page
  • 10.7554/eLife.27014 (Smith 2017 killifish FMT) — local PDF available; verified on N. furzeri page; cited by reference here
  • 10.1038/s41591-019-0495-2 (Depommier 2019 Akkermansia) — archive pending download; claims unverified no-fulltext-access
• implicit stubs created (new wikilinks to non-existent pages — HIGH PRIORITY follow-up):
  • [[akkermansia-muciniphila]] — gut barrier commensal; declining with age; therapeutic target; Phase 2 human data
  • [[scfa-signaling]] — butyrate/propionate/acetate biology; HDAC inhibition; colonocyte fuel
  • [[tmao]] — cardiovascular-risk metabolite; gut bacterial conversion from choline/carnitine
  • [[gut-barrier]] — intestinal barrier biology; tight junctions; mucus layer; leaky gut
  • [[lps-tlr4-nfkb]] — LPS→TLR4→NF-κB axis; bridge from dysbiosis to inflammaging
  • [[spermidine]] — referenced but no page (polyamine autophagy inducer; also food-sourced)
• gaps surfaced:
  • magnitude of age-specific SCFA decline: no verified primary source (#gap/unsourced)
  • circulating LPS age-trajectory in humans: observational data cited without specific paper (#gap/unsourced)
  • centenarian microbiome taxa associations: training knowledge, no verified source (#gap/unsourced)
  • Firmicutes/Bacteroidetes ratio controversy: noted and deprecated with explanation
  • causality direction (dysbiosis → aging vs aging → dysbiosis): explicitly flagged as unresolved (#gap/contradictory-evidence)
• schema note: hallmark frontmatter key-pathways used for stub pathway references (lps-tlr4-nfkb, scfa-signaling) even though those pages don’t exist yet — consistent with implicit-stub convention in CLAUDE.md
• ROADMAP updated: dysbiosis stub → drafted

[2026-05-04] ingest | round-10a-2 | altered-intercellular-communication

altered-intercellular-communication

• updated: hallmarks/altered-intercellular-communication.md — stub → synthesis-MOC
• entity type: hallmark (Integrative; López-Otín 2013, retained 2023)
• structure: synthesis-MOC over verified atomic pages; no primary PDFs read for this page (Conboy 2005 + NF-κB + SASP cross-checked via already-verified atomic pages)
• primary DOIs cited (11 footnotes):
  • 10.1016/j.cell.2013.05.039 (López-Otín 2013) — local PDF available
  • 10.1016/j.cell.2022.11.001 (López-Otín 2023) — download failed; no-fulltext-access
  • 10.1038/nature03260 (Conboy 2005 — heterochronic parabiosis) — local PDF available; cross-checked via satellite-cells (verified)
  • 10.1016/j.cell.2013.04.015 (Loffredo 2013 — GDF11 cardiac) — confirmed in archive, download pending
  • 10.1126/science.1251141 (Katsimpardi 2014 — GDF11 neurogenesis) — confirmed in archive, download pending
  • 10.1126/science.1251152 (Sinha 2014 — GDF11 muscle) — confirmed in archive, download pending
  • 10.1016/j.cmet.2015.05.010 (Egerman 2015 — GDF11 controversy) — confirmed in archive, download pending
  • 10.1038/nature10357 (Villeda 2011 — aged blood/neurogenesis) — confirmed in archive, download pending
  • 10.1038/nm.3898 (Smith/Villeda 2015 — B2M pro-aging factor) — confirmed in archive, download pending; needs-replication
• atomic pages linked (all confirmed to exist):
  • insulin-igf1 (verified), growth-hormone (verified), nf-kb (verified), sasp (verified), satellite-cells (verified), microglia (verified), immunosenescence (verified), sarcopenia (verified-partial), chronic-inflammation (stub), cellular-senescence (synthesis-MOC), dysbiosis (stub), deregulated-nutrient-sensing (synthesis-MOC), stem-cell-exhaustion (synthesis-MOC)
• implicit stubs created (new wikilinks to non-existent pages):
  • [[jak-stat-pathway]] — GH→JAK2/STAT5B cross-reference; referenced but no page
  • [[interventions/pharmacological/]] — referenced for HRT discussion; directory exists but no category page
  • [[hypotheses/gdf11-controversy]] — suggested for seeding when GDF11 dispute is documented
  • [[hypotheses/glucocorticoid-cascade-hypothesis]] — referenced in HPA axis section
  • [[studies/conboy-2005-heterochronic-parabiosis]] — study page not yet seeded
• gaps surfaced:
  • contradictory-evidence — GDF11 as youth factor: Loffredo/Wagers 2013–2014 vs Egerman 2015 replication challenge
  • needs-human-replication — heterochronic plasma strategies; B2M targeting
  • needs-replication — B2M (single lab)
  • unsourced — sex hormone axes, thyroid/adrenal axes, hypothalamic NF-κB pacemaker, JAK-STAT pathway, GDF15/CCL11 entries in signaling table
  • no-fulltext-access — López-Otín 2023 (download failed)
• SENS mapping: stub listed GlycoSENS but mapping is not established; corrected to no correspondence + unsourced note
• schema note: hallmark page carries verified: false per round-10a convention for synthesis-MOCs; CLAUDE.md type:hallmark schema does not specify this field — consistent with other round-10a pages; escalation noted in round-10a telomere-attrition log entry

[2026-05-04] ingest | round-10a-2 | chronic-inflammation

chronic-inflammation

• updated: hallmarks/chronic-inflammation.md — stub → synthesis-MOC (full expansion)
• entity type: hallmark (integrative; NEW in López-Otín 2023; split from altered-intercellular-communication)
• synthesis-MOC over verified atomic pages: nf-kb (verified), sasp (verified), senomorphics (verified-partial), senolytics (verified-partial), atherosclerosis (verified-partial), hematopoietic-stem-cells (verified-partial — CHIP/Jaiswal 2014), microglia (verified-partial), alzheimers-disease (verified-partial), dna-damage-response (verified-partial — Rodier 2009 two-arm), atm (verified)
• DOIs cited (7 footnotes):
  • 10.1111/j.1749-6632.2000.tb06651.x (Franceschi 2000; inflammaging coinage) — not_oa; no-fulltext-access
  • 10.1016/j.cell.2022.11.001 (López-Otín 2023; hallmarks expansion) — closed; no-fulltext-access
  • 10.1038/ncb1909 (Rodier 2009; DDR → SASP p53-independent ATM arm) — locally downloaded
  • 10.1056/NEJMoa1707914 (Ridker 2017 NEJM; CANTOS primary) — locally downloaded
  • 10.1016/S0140-6736(17)32247-X (Ridker 2017 Lancet; CANTOS cancer analysis) — pending download
  • 10.1126/science.aaf6659 (Childs 2016; senolytic atherosclerosis) — not_oa; no-fulltext-access
  • 10.1056/NEJMoa1408617 (Jaiswal 2014; CHIP + CV/cancer risk) — locally downloaded
• implicit stubs created/reinforced: nlrp3-inflammasome, cgas-sting, immunosenescence, frailty, neurodegeneration, sarcopenia, cardiomyocytes, jak-stat-pathway, clonal-hematopoiesis
• note: this is the wiki’s strongest clinical-evidence hallmark — CANTOS (n=10,061 Phase 3 RCT) verified locally; Childs 2016 senolytic atherosclerosis is preclinical mouse-only and not_oa
• schema note: verified: frontmatter field added to hallmark page — consistent with other round-10a MOC pattern; escalate whether hallmark pages should standardize on this field
• verification priority: Rodier 2009 (p53-independence of SASP; IL-6/IL-8 exact fold-change); CANTOS Lancet 2017 (exploratory cancer HR; pending download); Childs 2016 (plaque %; not_oa, unverifiable without PDF)

---

[2026-05-04] ingest | round-10a-2 | cellular-senescence

cellular-senescence

• updated: hallmarks/cellular-senescence.md — stub → synthesis-MOC
• entity type: hallmark (Antagonistic; López-Otín 2013, retained 2023)
• structure: synthesis-MOC per sops/writing-hypothesis-pages.md discipline — all claims are pointers to verified/verified-partial atomic pages; no primary-source re-statement
• atomic pages aggregated:
  • processes: sasp (verified), apoptosis (verified-partial), dna-damage-response (verified-partial)
  • pathways: p53-pathway (verified-partial), apoptosis-pathway (verified-partial), bcl-2-family-signaling (verified-partial), nf-kb (verified-partial)
  • proteins: p53 (verified), p21 (verified-partial), atm (verified-partial), foxo4 (verified), mdm2 (verified-partial), bcl-xl (verified-partial), bcl-2 (verified-partial), bak (verified, FULL), bax (verified-partial), bim (verified-partial), puma (verified-partial), noxa (verified-partial)
  • compounds: dasatinib (verified), quercetin (verified-partial), fisetin (verified), a1331852 (verified-partial)
  • interventions: senolytics (verified-partial), senomorphics (verified-partial)
  • cell types: cardiomyocytes (verified), satellite-cells (verified-partial), microglia (verified-partial), hematopoietic-stem-cells (verified-partial)
  • phenotypes: atherosclerosis (verified-partial), heart-failure (verified), type-2-diabetes (verified), frailty (verified), cancer (verified)
  • hallmarks: genomic-instability (drafted), telomere-attrition (drafted), chronic-inflammation (stub), deregulated-nutrient-sensing (drafted)
• primary DOIs cited (9 footnotes):
  • 10.1073/pnas.92.20.9363 (Dimri 1995; SA-β-gal biomarker) — verified on p21; archive status not re-checked
  • 10.1038/ncb1909 (Rodier 2009; two-arm DDR/SASP) — verified on dna-damage-response; local PDF available
  • 10.1038/nature16932 (Baker 2016; INK-ATTAC lifespan) — verified on p21; local PDF available
  • 10.1016/j.ebiom.2018.12.052 (Justice 2019; D+Q IPF trial) — archive: title confirmed, download_status: completed
  • 10.1016/j.ebiom.2019.08.069 (Hickson 2019; D+Q DKD) — archive: title confirmed, download_status: completed
  • 10.1111/acel.12344 (Zhu 2015; SCAP concept) — verified on senolytics (verified-partial)
  • 10.1016/j.cell.2017.02.031 (Baar 2017; FOXO4-DRI) — verified on foxo4 (verified); local PDF available
  • 10.1073/pnas.1515386112 (Xu 2015; JAK1/2 senomorphic) — verified on senomorphics (verified-partial)
  • 10.1126/science.aaf6659 (Childs 2016; atherosclerosis ~60% plaque reduction — not_oa; cited via atherosclerosis atomic page; no-fulltext-access)
• implicit stubs created (new wikilinks to non-existent pages):
  • [[p16-rb-pathway]] (pathway) — secondary senescence reinforcement checkpoint; referenced in frontmatter key-pathways and body; planned R10b
  • [[osteoarthritis]] (phenotype) — senescent chondrocyte SASP; Jeon 2017 mentioned; unsourced
  • [[apoptosenes]] (SENS category page) — referenced in frontmatter sens-correspondence and body
  • [[navitoclax]] (compound) — referenced in SCAP table and compound summaries; listed as planned in senolytics page
  • [[bcl-w]] (protein) — referenced in SCAP table as part of IMR90 triple; listed as planned R10b
• gaps surfaced:
  • #gap/needs-human-replication — all genetic lifespan extension data murine; no human longevity endpoint for any senolytic
  • #gap/contradictory-evidence — IMR90 SCAP: Zhu 2016 (triple BCL-2+BCL-xL+BCL-W) vs Yosef 2016 (BCL-W+BCL-xL dual sufficient); senescence-inducer-dependent
  • #gap/long-term-unknown — chronic senolytic therapy in humans; wound-healing and immune effects unstudied long-term
  • #gap/unsourced — osteoarthritis senescent chondrocyte data (Jeon 2017) not on wiki; osteoarthritis page not yet seeded
  • #gap/needs-replication — fisetin single-strain lifespan data; β-cell senescence (Aguayo-Mazzucato 2019) full PDF unverified
  • #gap/no-mechanism — developmental vs chronic senescence switch; universal senescent-cell biomarker absent
• schema notes: verified: false added to hallmark frontmatter — consistent with other round-10a MOCs; CLAUDE.md type: hallmark does not list verified: in schema definition; flagged for schema review in previous rounds
• ROADMAP updated: [[cellular-senescence]] marked drafted — synthesis-MOC; round-10a-2 2026-05-04; Coverage summary updated to 7 drafted, 5 stubs
• suggested verification priority for wiki-verifier:
  1. Justice 2019 (10.1016/j.ebiom.2018.12.052) — local PDF available; confirm 6MWT +21.5 m and gait speed p-values against full text
  2. Hickson 2019 (10.1016/j.ebiom.2019.08.069) — local PDF available; confirm p16 −35%, p21 −17%, SA-β-Gal −62% against full text
  3. Baker 2016 (10.1038/nature16932) — local PDF available; already verified on p21.md; cross-check cancer-incidence-at-autopsy claim on this page against that verified source
  4. Baar 2017 (10.1016/j.cell.2017.02.031) — local PDF available; confirm FOXO4-DRI in-vivo endpoint descriptions against full text (cross-check vs foxo4 verified page)
  5. Childs 2016 (10.1126/science.aaf6659) — not_oa; ~60% plaque reduction figure is carried from atherosclerosis (which carries no-fulltext-access on Childs 2016); cannot be independently verified without PDF access

[2026-05-04] ingest | round-10a-2 | stem-cell-exhaustion

stem-cell-exhaustion

• updated: hallmarks/stem-cell-exhaustion.md — stub → synthesis-MOC
• entity type: hallmark (integrative; López-Otín 2013, retained 2023)
• structure: synthesis-MOC over verified atomic pages; no primary-source findings restated; all evidence delegated via wikilinks per sops/writing-hypothesis-pages.md discipline
• atomic pages drawn on (all verified or verified-partial):
  • cell-types/hematopoietic-stem-cells (verified) — Sudo 2000 ~17× expansion; Beerman 2010 myeloid bias; Pang 2011 human replication; Jaiswal 2014 CHIP HR 11.1 cancer / 2.0 CHD / 2.6 stroke; Genovese 2014 CHIP HR 12.9; Rossi 2007 DNA damage; niche (Pinho & Frenette 2019)
  • cell-types/satellite-cells (verified) — Conboy 2005 heterochronic parabiosis; <0.1% GFP+ engraftment; Sousa-Victor 2014 geriatric p16+ senescence; Brack 2007 Wnt fibrogenic conversion
  • cell-types/microglia (verified-partial) — Hammond 2019 OA2/OA3 aging clusters; DAM stages
  • cell-types/cardiomyocytes (verified) — Lewis-McDougall 2019 >50% p16+ CPCs in >70 yr donors
  • phenotypes/anemia-of-aging (verified) — Guralnik 2004 three-category etiology; UAA fraction
  • phenotypes/sarcopenia, phenotypes/immunosenescence, phenotypes/frailty (stubs) — linked
  • pathways/mtor, pathways/insulin-igf1, processes/autophagy, processes/mitophagy — pathway contributions
  • upstream hallmarks: genomic-instability, telomere-attrition, epigenetic-alterations, cellular-senescence, chronic-inflammation, mitochondrial-dysfunction, disabled-macroautophagy, deregulated-nutrient-sensing — all linked with mechanism descriptions
• primary DOIs cited (1 footnote — synthesis-MOC; primary claims deferred to atomic pages):
  • 10.1016/j.cell.2013.05.039 (López-Otín 2013) — local PDF available
• implicit stubs created (new wikilinks to non-existent pages):
  • [[clonal-hematopoiesis]] — already flagged on HSC page; high priority
  • [[notch-pathway]] — satellite cell quiescence-activation switch; not yet seeded
  • [[immunosenescence]], [[frailty]], [[sarcopenia]] — phenotype pages; stub or not yet created
• gaps surfaced:
  • #gap/contradictory-evidence — GDF11 rejuvenation controversy; candidate for hypothesis page
  • #gap/needs-human-replication — per-cell reconstitution; parabiosis findings; geriatric satellite cell p16+ in humans; senolytics for stem cell function
  • #gap/unsourced — intestinal crypt, neural, hair follicle, mesenchymal stem cell aging; no verified pages
  • #gap/long-term-unknown — epigenetic reprogramming safety; optimal intervention timing
• schema decisions: verified: frontmatter field added to type: hallmark consistent with prior round-10a convention; escalation already in log for prior entries, not re-escalated
• ROADMAP updated: [[stem-cell-exhaustion]] marked drafted — synthesis-MOC; round-10a-2 2026-05-04; coverage summary: 6 drafted, 6 stubs
• suggested verification priority for wiki-verifier: López-Otín 2013 (10.1016/j.cell.2013.05.039) definitional claims; all quantitative claims inherit from already-verified atomic pages and require no independent re-check on this MOC

---

[2026-05-04] ingest | round-10a-hallmark-mocs | telomere-attrition

telomere-attrition

• added: hallmarks/telomere-attrition.md (converted from stub to synthesis-MOC)
• entity type: hallmark (Primary; López-Otín 2013, retained 2023)
• synthesis-MOC over verified atomic pages: dna-damage-response (verified-partial), p53 (verified), mus-musculus (verified, full — Prowse 1995 confirmed: telomerase detected in liver/kidney/spleen/testis; absent from brain), hematopoietic-stem-cells (verified-partial), free-radical-theory-of-aging (verified), cellular-senescence (stub), sasp (verified)
• DOIs cited (7 footnotes):
  • 10.1016/j.cell.2013.05.039 (López-Otín 2013) — local PDF available
  • 10.1038/345458a0 (Harley 1990; telomere shortening in fibroblasts) — not_oa; no-fulltext-access
  • 10.1101/gad.1346005 (de Lange 2005; shelterin) — pending
  • 10.1038/ncb1275 (Celli & de Lange 2005; TRF2 deletion + DDR) — local PDF available
  • 10.1073/pnas.92.11.4818 (Prowse & Greider 1995; mouse somatic telomerase) — local PDF available
  • 10.1016/S0140-6736(03)12384-7 (Cawthon 2003; LTL + mortality) — not_oa; no-fulltext-access
  • 10.1002/emmm.201200245 (Bernardes de Jesus 2012; AAV-TERT lifespan) — pending; no-fulltext-access
• implicit stubs created: telomere, telomerase, shelterin, telomerase-pathway, tert, terc, trf1, trf2, pot1, antagonistic-pleiotropy
• schema note: verified: frontmatter field added to hallmark page despite CLAUDE.md schema not listing it for type: hallmark — consistent with other round-10a MOCs; escalate to user whether hallmark pages should carry the verified flag
• verification priority: Harley 1990 (bp/division + Hayflick limit numerics); Cawthon 2003 (mortality HRs, n, adjustment); Bernardes de Jesus 2012 (lifespan % + n per group)

---

[2026-05-04] ingest | round-10a-hallmark-mocs | deregulated-nutrient-sensing

deregulated-nutrient-sensing

• updated: hallmarks/deregulated-nutrient-sensing.md — stub → synthesis-MOC
• entity type: hallmark (antagonistic; López-Otín 2013, retained 2023)
• structure: synthesis-MOC over verified atomic pages; no primary PDFs read for this page
• primary DOIs cited (11 footnotes):
  • 10.1038/366461a0 (Kenyon 1993 — daf-2 founding longevity paper) — confirmed in archive, not_oa no-fulltext-access
  • 10.1038/nature01298 (Holzenberger 2003 — Igf1r+/- +26% lifespan) — confirmed in archive, local PDF available
  • 10.1210/en.2003-0247 (Coschigano 2003 — GHR-/- Laron mouse) — confirmed in archive, not_oa no-fulltext-access
  • 10.1126/scitranslmed.3001845 (Guevara-Aguirre 2011 — Ecuador Laron cohort) — confirmed in archive, download failed no-fulltext-access
  • 10.1073/pnas.0705467105 (Suh 2008 — centenarian IGF-1 paradox) — confirmed in archive, local PDF available
  • 10.1016/s1534-5807(04)00095-4 (Hertweck 2004 — sgk-1 LoF +63% C. elegans) — confirmed in archive, local PDF available
  • 10.1038/nature08221 (Harrison 2009 — NIA ITP rapamycin) — confirmed in archive, local PDF available
  • 10.1038/nrm3311 (Hardie 2012 — AMPK γ-subunit) — confirmed in archive; see ampk (verified, FULL)
  • 10.1038/ncomms3192 (Martin-Montalvo 2013 — metformin lifespan) — confirmed in archive, local PDF available
  • 10.1073/pnas.0801030105 (Willcox 2008 — FOXO3 GWAS OR=2.75) — confirmed in archive, local PDF available
  • 10.1016/j.cell.2007.02.005 (Taguchi 2007 — brain-Irs2 KO +18% female lifespan) — confirmed in archive, not_oa no-fulltext-access
• all cited DOIs are backed by verified atomic pages (see sgk1, foxo3, igf-1, growth-hormone, rapamycin, metformin, ampk, caloric-restriction, insulin-igf1)
• implicit stubs created (new wikilinks to non-existent pages):
  • [[type-2-diabetes]] (phenotype) — referenced; stub needed
  • [[mtor-kinase]] (protein) — planned; path noted in page
  • [[jak-stat-pathway]] (pathway) — referenced for GH→JAK2/STAT5B signaling; not yet seeded
  • [[antagonistic-pleiotropy]] (hypothesis) — referenced; not yet seeded
• existing atomic pages linked (all confirmed to exist):
  • pathways: insulin-igf1 (verified-partial), mtor (verified-partial), ampk (verified FULL), pi3k-akt-pathway (verified-partial), sirtuin (verified-partial)
  • proteins: daf-2 (verified), daf-16 (verified-partial), igf1r (drafted), igf-1 (verified), sgk1 (verified), akt (drafted), foxo3 (verified), foxo-transcription-factors (verified), s6k1 (verified-partial), growth-hormone (verified), tsc1-tsc2 (verified FULL), raptor (verified FULL), rictor (verified-partial), 4ebp1 (verified-partial), ulk1 (verified-partial), irs-1 (drafted), irs2 (drafted), pten (drafted), foxo1 (drafted), foxo4 (drafted), sirt1 (verified-partial)
  • compounds: rapamycin (verified-partial), metformin (verified-partial), nmn (drafted), nr (drafted)
  • interventions: caloric-restriction (verified)
  • hypotheses: hyperfunction-theory (verified), disposable-soma-theory (verified)
• gaps surfaced:
  • no-fulltext-access — Coschigano 2003, Guevara-Aguirre 2011, Taguchi 2007, Kenyon 1993 (all closed-access; retained with gap tags)
  • needs-human-replication — direct IIS signaling measurement in centenarians; sex-differential longevity mechanism
  • no-mechanism — sex-differential rapamycin/IGF1R longevity effects
  • unsourced — sirtuin NAD+ decline with age claim (deferred to sirtuin)
  • dose-response-unclear — optimal rapamycin/metformin dosing for aging endpoints
  • long-term-unknown — chronic rapamycin in humans over multi-year course
• schema decisions: hallmark page used synthesis-MOC structure (evidence-aggregating, matching hyperfunction-theory mode) rather than the sparse stub template. CLAUDE.md type: hallmark schema does not specify body depth for converted stubs — structural decision escalated below.
• ROADMAP updated: deregulated-nutrient-sensing marked drafted — synthesis-MOC; round-10a 2026-05-04

[2026-05-04] ingest | round-10a-hallmark-mocs

mitochondrial-dysfunction

• updated: hallmarks/mitochondrial-dysfunction.md — stub → synthesis-MOC (round-10a-2)
• entity type: hallmark (Antagonistic; López-Otín 2013, retained 2023)
• structure: synthesis-MOC over verified atomic pages; no primary PDFs read for this page; all claims are pointers
• atomic pages drawn on (all verified or verified-partial):
  • processes: mitophagy (verified-partial — Onishi 2021 + Singh 2022 ATLAS), autophagy (verified)
  • pathways: pink1-parkin-pathway (verified — Narendra/Onishi/Wauer/Yang), ampk (verified FULL), mtor (verified), sirtuin (verified-partial)
  • proteins: pink1 (verified-partial), parkin (verified-partial), bnip3 (verified-partial), nix (verified-partial), fundc1 (verified-partial), bcl2l13 (verified-partial), cytochrome-c (verified-partial), apaf-1 (verified-partial)
  • compounds: urolithin-a (verified — ATLAS RCT)
  • model-organisms: heterocephalus-glaber (verified-partial — NMR high-damage/long-life paradox)
  • hypotheses: free-radical-theory-of-aging (verified, superseded)
  • phenotypes: parkinsons-disease (verified-partial), heart-failure (verified-partial), sarcopenia (verified), cardiomyocytes (verified-partial)
• primary DOIs cited as footnotes (2):
  • 10.1038/nature02517 (Trifunovic 2004 — PolG D257A mutator mouse premature aging) — local PDF available; numerics not verified against PDF — wiki-verifier priority
  • 10.1016/j.xcrm.2022.100633 (Singh 2022 ATLAS RCT urolithin A) — local PDF available; already verified on urolithin-a and mitophagy
• implicit stubs created (new wikilinks to non-existent pages):
  • [[mtdna]] — planned R10d; mtDNA mutation accumulation detail
  • [[oxphos]] — planned R10d; ETC complex data
  • [[mitochondrial-biogenesis]] — planned R10d; PGC-1α quantitative data
  • [[hypotheses/mitohormesis]] — competing ROS-as-signal frame; not yet seeded
  • [[drp1]], [[mfn1]], [[mfn2]], [[opa1]] — mitochondrial dynamics proteins; not yet seeded
  • [[type-2-diabetes]] — phenotype; referenced for insulin resistance mechanism
• gaps surfaced:
  • #gap/unsourced — ETC complex activity decline quantitative figures; PGC-1α age-decline; mtDNA tissue heteroplasmy thresholds; tissue-specific mitochondrial aging rates; T2D mechanism
  • #gap/contradictory-evidence — PolG mutator mutation levels vs normal aging; NAD+ precursor human null results; NMR oxidative damage paradox
  • #gap/no-mechanism — PARL dispute in PINK1 cleavage; mitohormesis ROS threshold in aged human tissue; MitoSENS/organelle-clearance relationship; mtDNA heteroplasmy selection by mitophagy
  • #gap/needs-human-replication — mitochondria-targeted antioxidants; cardiac senolytic → mitochondrial improvement; PINK1/Parkin flux in aged human tissues
  • #gap/dose-response-unclear — optimal mitophagy induction level in old organisms
  • #gap/long-term-unknown — urolithin A long-term effects
• schema notes: verified: false added to hallmark frontmatter consistent with other R10a synthesis-MOC pages (same convention as telomere-attrition, deregulated-nutrient-sensing, etc.; CLAUDE.md type: hallmark schema does not list verified: — escalation still open per round-10a-1 session)
• ROADMAP updated: [[mitochondrial-dysfunction]] stub → drafted in Coverage summary (5 drafted, 7 stubs) and R10a checklist
• verification priority for wiki-verifier:
  1. Trifunovic 2004 (10.1038/nature02517) — local PDF available; confirm median lifespan figures, mutation accumulation rates, and phenotype description against PDF (currently unverified numerics)
  2. Singh 2022 ATLAS — already verified on urolithin-a and mitophagy; synthesis-MOC use of those figures is appropriate

loss-of-proteostasis

• added: hallmarks/loss-of-proteostasis.md (converted from stub to synthesis-MOC)
• entity type: hallmark (Primary; López-Otín 2013, retained 2023)
• synthesis-MOC over verified atomic pages: draws on autophagy (verified), chaperone-mediated-autophagy (verified-partial), atg5 (verified), atg7 (verified), beclin-1 (verified), lc3 (verified-partial), p62 (verified-partial), tfeb (verified-partial), mitophagy (verified-partial), neurodegeneration (verified-partial), alzheimers-disease (verified-partial), parkinsons-disease (verified-partial), rapamycin (verified), spermidine (verified)
• DOIs cited (9 footnotes):
  • 10.1016/j.cell.2013.05.039 (López-Otín 2013) — local PDF available
  • 10.1016/j.cell.2022.11.001 (López-Otín 2023) — failed download; no-fulltext-access
  • 10.1146/annurev-biochem-060809-095203 (Anckar & Sistonen 2011, HSF1) — local PDF available
  • 10.1074/jbc.M002102200 (Cuervo & Dice 2000, CMA aging decline) — hybrid OA, PDF failed; no-fulltext-access
  • 10.1038/nm.1851 (Zhang & Cuervo 2008, LAMP-2A rescue) — local PDF available
  • 10.1038/nature04723 (Komatsu 2006, Atg7 neuron KO) — local PDF available
  • 10.1038/nature04724 (Hara 2006, Atg5 neuron KO) — local PDF available
  • 10.1038/ncomms3300 (Pyo 2013, Atg5 transgenic lifespan) — local PDF available
  • 10.1038/s41586-018-0162-7 (Fernández 2018, Beclin-1 F121A lifespan) — local PDF available
  • 10.1038/s41580-018-0001-6 (Kaushik & Cuervo 2018 review) — PDF failed; no-fulltext-access
• implicit stubs created (wikilinks to non-existent pages):
  • [[ubiquitin-proteasome-system]] — planned R10b; not yet seeded
  • [[unfolded-protein-response]] — referenced in frontmatter key-pathways; no page yet
  • [[alpha-synuclein]] — referenced in CMA disease section; planned R10d
• gaps surfaced:
  • #gap/needs-human-replication — HSF1 decline in human tissues; CMA decline human data; lifespan extension from autophagy gene manipulation
  • #gap/no-fulltext-access — López-Otín 2023 (download failed); Cuervo & Dice 2000 (hybrid OA, failed); Kaushik & Cuervo 2018 (PMC blocked)
  • #gap/needs-replication — ATG5 lifespan extension (Pyo 2013, single strain); Beclin-1 F121A lifespan (Fernández 2018, single strain)
  • #gap/unsourced — UPS age-decline claim sourced to López-Otín 2013 review only; needs UPS atomic page with primary citations
  • #gap/dose-response-unclear + #gap/long-term-unknown — CMA small-molecule activators (CA77.1, AR.7) preclinical stage
  • #gap/contradictory-evidence — anti-Aβ immunotherapy reduces plaque burden but modest clinical benefit
• SENS correspondence: LysoSENS (intracellular aggregates) + AmyloSENS (extracellular Aβ/tau/transthyretin) — retained from stub; consistent with existing mapping
• schema notes: hallmark type does not have a verified frontmatter field per CLAUDE.md schema; added it anyway as this is a multi-source synthesis page (consistent with how other multi-source pages handle this) — escalation: CLAUDE.md type: hallmark schema does not include verified: field; added for seeder-state tracking; suggest adding to CLAUDE.md in next schema pass
• ROADMAP updated: [[loss-of-proteostasis]] entry updated from stub to drafted 2026-05-04 in both Coverage summary and R10a checklist
• suggested verification priority for wiki-verifier:
  1. Hara 2006 (10.1038/nature04724) — local PDF; confirm neuron-specific Atg5 KO phenotype description matches paper
  2. Komatsu 2006 (10.1038/nature04723) — local PDF; confirm ≤28 weeks death, ubiquitin/p62 inclusions claim
  3. Pyo 2013 (10.1038/ncomms3300) — local PDF; confirm ~17.2% median lifespan extension figure
  4. Fernández 2018 (10.1038/s41586-018-0162-7) — local PDF; numerics already verified on beclin-1.md atomic page
  5. Zhang & Cuervo 2008 (10.1038/nm.1851) — local PDF; confirm liver-specific LAMP-2A restoration claims match paper
  6. Cuervo & Dice 2000 (10.1074/jbc.M002102200) — hybrid OA, PDF inaccessible; LAMP-2A age-decline claim unverifiable until PDF obtained

disabled-macroautophagy

• added: hallmarks/disabled-macroautophagy.md — converted from stub to full synthesis-MOC
• entity type: hallmark (Primary; introduced 2023; split from loss-of-proteostasis)
• structure: synthesis-MOC per sops/writing-hypothesis-pages.md discipline — all claims are pointers to verified/verified-partial atomic pages; no primary-source re-statement
• atomic pages drawn on (26 total from Round 6 cluster):
  • Core machinery (verified): ulk1, beclin-1, atg5, atg7, atg13, atg101, atg12, lc3, tfeb, p62
  • Selective autophagy processes (verified-partial): mitophagy, lipophagy, chaperone-mediated-autophagy, xenophagy
  • Mitophagy receptors (drafted): pink1, parkin, pink1-parkin-pathway, ndp52, optn, bnip3, nix, fundc1, bcl2l13
  • Regulatory pathways (verified): mtor, ampk
  • Compounds (verified): spermidine, rapamycin, caloric-restriction, urolithin-a
  • Hypothesis (verified): hyperfunction-theory
• primary DOIs cited as footnotes (6):
  • Hara 2006 (10.1038/nature04724) — neural Atg5 KO neurodegeneration; confirmed in atg5 (verified)
  • Komatsu 2006 (10.1083/jcb.200607037) — neural Atg7 KO neurodegeneration; confirmed in atg7 (verified-partial)
  • Pyo 2013 (10.1073/pnas.1312369110) — Atg5 Tg 17.2% lifespan extension; confirmed in atg5 (verified)
  • Fernández 2018 (10.1038/s41586-018-0327-4) — Beclin-1 F121A ~12% lifespan extension; confirmed in beclin-1 (verified)
  • Liang 1999 (10.1038/46766) — Becn1+/- tumorigenesis; cited with gap note (not_oa per beclin-1)
  • Hansen 2018 (10.1038/s41580-018-0033-y) — autophagy epistasis review; confirmed in autophagy (verified-partial)
• implicit stubs created (new wikilinks to non-existent pages):
  • neurodegeneration (in frontmatter key-phenotypes)
  • heart-failure (in frontmatter key-phenotypes)
  • lipofuscin (in age-decline mechanisms section)
  • urolithin-a — appears in therapeutic table; may not have a full atomic page yet
  • ubiquitin-proteasome-system — referenced in “why split” section
• gaps surfaced:
  • #gap/needs-human-replication — genetic lifespan extension (Atg5 Tg, Beclin-1 F121A), TFEB age-decline, ATG protein decline with age
  • #gap/dose-response-unclear — optimal autophagy induction level; hormetic U-curve in aging tissues
  • #gap/unsourced — autophagy-senescence link (senescent phenotype stabilization); trehalose-TFEB mechanism; autophagosome-lysosome fusion failure in aged cells
  • #gap/no-mechanism — lipofuscin-induced lysosomal inhibition molecular mechanism
• ROADMAP updated: disabled-macroautophagy marked (drafted — synthesis-MOC; round-10a 2026-05-04); coverage summary updated (3 drafted, 9 stubs)
• schema notes: this page uses type: hallmark (not type: hypothesis) per CLAUDE.md; synthesis-MOC discipline applied from sops/writing-hypothesis-pages.md despite different type — no structural escalation needed; hallmark MOCs are confirmed to use this discipline per task brief
• verification priority for wiki-verifier:
  1. Fernández 2018 (10.1038/s41586-018-0327-4) — Beclin-1 F121A lifespan claim (~12%) should be confirmed against PDF if available; beclin-1 (verified) already checked this paper
  2. Pyo 2013 (10.1073/pnas.1312369110) — Atg5 Tg 17.2% lifespan claim confirmed via atg5 (verified); secondary pass warranted
  3. Hansen 2018 (10.1038/s41580-018-0033-y) — download-failed per ROADMAP note; epistasis claims are the most impactful on this page; verify when PDF is accessible

genomic-instability

• converted: hallmarks/genomic-instability.md from stub → drafted synthesis-MOC
• page type: type: hallmark (primary, introduced 2013, retained 2023)
• structure: synthesis-MOC per sops/writing-hypothesis-pages.md discipline — no primary claims restated; all evidence delegated via wikilinks to verified atomic pages
• atomic pages aggregated (all verified or verified-partial):
  • pathways/dna-damage-response (verified-partial) — two-arm SASP model; γH2AX age-accumulation; progeroid syndrome table
  • molecules/proteins/atm (verified-partial) — DSB transducer; IL-6 ~50× / IL-8 ~10× SASP reduction; ATM depletion dissociates SASP from arrest
  • molecules/proteins/p53 (verified) — Tyner 2002 m-allele: tumor-resistant + 23% shorter lifespan; antagonistic pleiotropy
  • molecules/proteins/mdm2 (verified-partial) — Mendrysa 2006: mdm2puro/7-12 tumor-resistant + NOT prematurely aged (p=0.61); inversion finding
  • molecules/proteins/p21 (verified-partial) — Brugarolas 1995 G1-only; Baker 2016 INK-ATTAC lifespan +24–27%
  • pathways/p53-pathway (verified-partial) — DDR→p53 stabilization axis
  • pathways/apoptosis-pathway (verified-partial) — DNA damage → BAX/PUMA → MOMP
  • cell-types/hematopoietic-stem-cells (verified) — CHIP: Jaiswal 2014 HR 11.1 hematologic cancer + HR 1.4 all-cause + HR 2.0 CHD + HR 2.6 stroke; Genovese 2014 HR 12.9 hematologic cancer
  • phenotypes/cancer (verified) — aging-cancer paradox; Tyner/Mendrysa/Pten trade-off experiments; Baker 2016 tumor-latency effect
  • hypotheses/free-radical-theory-of-aging (verified, status: superseded) — ROS as contributing factor not primary driver
• implicit stubs created (new wikilinks pointing to non-existent pages):
  • [[clonal-hematopoiesis]] — dedicated CHIP page (implicit stub; currently redirected to cancer + HSC pages)
  • [[neurodegeneration]] — phenotype page (implicit stub in frontmatter key-phenotypes)
  • [[heterocephalus-glaber]] — model organism page (cross-referenced; may already exist as stub — check)
  • [[interventions/lifestyle/caloric-restriction]] — intervention page (implicit stub)
  • [[interventions/pharmacological/senolytics]] — intervention category page
• gaps surfaced:
  • #gap/unsourced — mtDNA mutation accumulation atomic page not seeded
  • #gap/unsourced — somatic mosaicism atlas data not seeded
  • #gap/needs-human-replication — DDR fidelity decline causal for human aging; p53 accelerated-aging trade-off human evidence weak
  • #gap/long-term-unknown — CHIP natural history beyond 10 years
  • #gap/contradictory-evidence — net effect of chronic ATM/ATR signaling in aged tissue (beneficial repair vs harmful SASP)
• schema notes: hallmark type schema does NOT include verified: field (unlike protein/compound/study). No verified: false added — consistent with other hallmark stubs in wiki. Banner present per synthesis-MOC convention. No schema gap escalation required.
• roadmap updated: [[genomic-instability]] marked (drafted — synthesis-MOC; round-10a 2026-05-04)

[2026-05-04] verify | chaperone-mediated-autophagy

Pages verified: 1 (partial scope — 3 sources verified against primary PDFs; 2 verified via abstract; 3 via abstract only [not_oa/no-DOI]; 1 [Kaushik 2018] download failed/PMC blocked)

• processes/chaperone-mediated-autophagy.md — corrections made; verified: true (partial scope)

Sources checked (8 total):

• 10.1128/MCB.02070-07 Bandyopadhyay 2008 Mol Cell Biol — VERIFIED (full PDF)
• 10.1038/nm.1851 Zhang & Cuervo 2008 Nat Med — VERIFIED (full PDF; added at verification)
• 10.1080/15548627.2021.1935007 Bourdenx 2021 Autophagy — VERIFIED (full PDF)
• 10.1074/jbc.M002102200 Cuervo & Dice 2000 JBC — ABSTRACT ONLY (hybrid OA; PDF download failed HTTP 403)
• 10.1016/j.cmet.2014.06.009 Schneider 2014 Cell Metab — ABSTRACT ONLY (Camoufox downloaded Cell Press media kit instead of paper; abstract+Crossref used)
• 10.1126/science.2799391 Chiang 1989 Science — ABSTRACT ONLY (not_oa)
• 10.1126/science.273.5274.501 Cuervo & Dice 1996 Science — ABSTRACT ONLY (not_oa)
• PMID:2103896 Dice 1990 Semin Cell Biol — ABSTRACT ONLY (no DOI; pre-DOI era)
• 10.1038/s41580-018-0001-6 Kaushik & Cuervo 2018 Nat Rev Mol Cell Biol — UNVERIFIED (download failed; PMC returned “not Open Access” despite archive listing green OA) no-fulltext-access

Key corrections to processes/chaperone-mediated-autophagy.md:

1. MAJOR: [^schneider2014] was wrongly described as “LAMP-2A overexpression in aged mice” — this DOI is actually a CMA liver-specific KNOCKOUT paper (conditional Albumin-Cre; L2A^f/f); corrected throughout page body and footnote
2. NEW CITATION ADDED: Zhang & Cuervo 2008 (10.1038/nm.1851, Nat Med) is the actual LAMP-2A restoration rescue paper; added as [^zhang2008] with correct details: FVB background (NOT C57BL/6), 22–26 months (NOT 18 months), double transgenic (Alb-Tet-off-LAMP-2A)
3. Cuervo 2000 rat age: “~24 months” → “22 months” (abstract states “22-month-old”)
4. Cuervo 2000 “>30% decline”: claim not in abstract; removed from lead; tagged needs-replication pending body-text verification
5. KFERQ motif prevalence: “~30%” → range “~30% (early estimates) to ~40% (Kaushik 2018)” to reflect algorithm-dependency
6. Bandyopadhyay 2008 complex: clarified that LAMP-2A organizes into MULTIPLE complexes (>800, 720–680, 520, 250, 110 kDa); 700 kDa is the CMA-active translocation complex. hsc70 is found at lower MW complexes (≤250 kDa) and promotes DISASSEMBLY from 700 kDa; substrates bind MONOMERIC LAMP-2A, not the 700 kDa complex. HSP90 is luminally associated and stabilizes LAMP-2A during transitions.
7. Bourdenx 2021 CMA activator: clarified compound is called “CA” in this paper (not QX77 or CA77.1); full characterization is in companion Cell 2021 paper; footnote updated with exact model (CKL2A^-/- + MAPT/tau + APP/Aβ AD models)
8. Kaushik 2018: flagged as no-fulltext-access; QX77/CA77.1/AR.7 activator names marked “Unverified: source PDF inaccessible”
9. Schneider 2014 downloaded as corrupted file (Cell Press media kit) — filed as archive defect; abstract/Crossref used for verification

Downstream pages that may need updates (propagation for main agent):

• loss-of-proteostasis — if it cites the “>30% CMA decline” or Schneider 2014 as a rescue paper
• autophagy — if it cites the CMA compensation claim with a specific percentage
• Any pages citing zhang-2008-cma-liver-rescue (stub may need seeding)

[2026-05-04] ingest | round-10d | tau

tau

• added: molecules/proteins/tau.md
• entity type: protein (MAPT gene; microtubule-associated protein tau; neurofibrillary tangle protein)
• canonical IDs: UniProt P10636 (TAU_HUMAN), NCBI Gene 4137, HGNC 6893, mouse ortholog Mapt; GenAge: null (#gap/needs-canonical-id — not listed in GenAge human database)
• DOIs cited (5 footnotes):
  • doi:10.1038/31508 (Hutton 1998 — MAPT missense + splice-site mutations in FTDP-17; Nature 393:702–705; 3,484 citations; 100th percentile) — LOCAL PDF available at
  • doi:10.1093/emboj/8.2.393 (Goedert 1989 — tau cDNA cloning + six isoforms; EMBO J 8:393–399) — NOT IN ARCHIVE; no-fulltext-access
  • doi:10.1146/annurev.neuro.24.1.1121 (Lee 2001 — tauopathies review; Annu Rev Neurosci 24:1121–1159; 2,339 citations; 100th percentile) — in DOI lookup (confirmed), not downloaded (not_oa); no-fulltext-access
  • doi:10.1371/journal.pbio.0070034 (Mukrasch 2009 — NMR structure of full-length 2N4R tau; PLoS Biol 7:e34) — NOT IN ARCHIVE; no-fulltext-access
  • doi:10.1038/nn1387 (Cuervo 2004 — tau as CMA substrate; LAMP-2A blockade by phospho-tau; Nat Neurosci 7:1355–1360) — archive status not checked at seeding; add to verification queue
• key sections:
  • Identity table (UniProt, NCBI, HGNC, Ensembl, mouse ortholog, chromosomal location, gene size)
  • Isoforms table (six major isoforms: 0/1/2N × 3R/4R; 352–441 aa; fetal brain 0N3R only)
  • Domain organization (N-terminal projection domain, PRR, MTBR, C-terminal tail; intrinsically disordered protein framing per Mukrasch 2009)
  • Native function (MT stabilization, axonal transport, axonal segregation; Mapt-/- mild motor phenotype)
  • Pathological hyperphosphorylation (~85 sites; AD tau 3–4× normal; key sites table: AT8 Ser202/Thr205; PHF1 Ser396/Ser404; p-tau181 CSF biomarker; p-tau217 plasma biomarker; kinase table: GSK3β, CDK5, p70S6K, DYRK1A)
  • Pathological aggregation cascade (MT detachment → oligomers → beta-sheet → PHFs → NFTs → ghost tangles; prion-like spreading; Braak staging framework)
  • Tauopathy taxonomy table (AD 3R+4R, Pick 3R, PSP 4R, CBD 4R, AGD 4R, CTE 3R+4R, FTDP-17)
  • MAPT mutations and FTDP-17 genetics (Hutton 1998; missense vs splice-site; H1/H2 haplotype)
  • Tau as biomarker (CSF t-tau, CSF p-tau181/p-tau231, plasma p-tau217, [18F]-flortaucipir PET)
  • Clearance pathways (CMA KFERQ-like motif; phospho-tau LAMP-2A blockade; macroautophagy; UPS; LAMP-2A age-decline context)
  • Therapeutic landscape (gosuranemab Phase 2 NEGATIVE; tilavonemab Phase 2 NEGATIVE; semorinemab mixed; BIIB080 ASO Phase 2 pending; epothilone D abandoned; LMTX contradictory; GSK3β/CDK5 inhibitors preclinical)
  • Interactors and pathway connections (gsk3b, cdk5, hsp70, lamp2a, mtor, alzheimers-disease, neurodegeneration, chaperone-mediated-autophagy, loss-of-proteostasis)
  • Limitations and gaps (6 gap tags)
• implicit stubs created (new wikilinks to non-existent pages):
  • gsk3b — tau kinase; no protein page
  • cdk5 — tau kinase; no protein page
  • hsp70 — chaperone triage; no protein page
  • lamp2a — CMA receptor; no protein page (canonical home is chaperone-mediated-autophagy)
  • p70s6k — mTOR effector / tau kinase; referenced via mtor
  • hutton-1998-mapt-ftdp17 — study page not yet seeded; Hutton 1998 local PDF available for wiki-verifier
  • lee-2001-neurodegenerative-tauopathies — study page not yet seeded; not downloadable (not_oa)
• existing pages cross-referenced (all confirmed to exist):
  • alzheimers-disease (verified-partial — tau inbound cross-link confirmed), neurodegeneration (verified-partial), loss-of-proteostasis (drafted hallmark), disabled-macroautophagy (drafted hallmark), chaperone-mediated-autophagy (verified-partial — tau as CMA substrate)
  • mtor — p70S6K → Thr212 tau phosphorylation axis
• gaps surfaced:
  • needs-human-replication — LAMP-2A age-decline magnitude in human neurons; all age-decline data from rodent tissues
  • contradictory-evidence — oligomers vs fibrils as primary toxic tau species; disease-stage-dependent
  • no-mechanism — trans-synaptic tau spreading receptors at the molecular level
  • long-term-unknown — BIIB080 clinical outcome data; active tau immunotherapy long-term safety
  • contradictory-evidence — tau role in non-neuronal cells (astrocytes, oligodendrocytes) in 4R tauopathies
  • unsourced — tau nuclear function; needs primary citation
  • needs-canonical-id — GenAge ID; tau not confirmed in GenAge human database
  • no-fulltext-access — Goedert 1989, Lee 2001 (not_oa), Mukrasch 2009 (not in archive)
• schema notes: none — protein frontmatter schema fully accommodates this entity; genage-id: null per convention for unconfirmed entries; mouse-ortholog: Mapt added (consistent with p53.md, miro1.md precedent)
• ROADMAP updated: tau in Proteins section updated from [ ] (1) → [x] (drafted, verified: false) — Round 10d 2026-05-04
• verification priority for wiki-verifier:
  1. Hutton 1998 (10.1038/31508) — LOCAL PDF available; verify FTDP-17 mutation classification (missense vs splice-site), family counts, mutation nomenclature against primary PDF; high-impact foundational paper for the tauopathy field
  2. Cuervo 2004 (10.1038/nn1387) — tau-as-CMA-substrate claim; archive status not confirmed; retrieve and verify LAMP-2A blockade mechanism and tau motif identification
  3. Lee 2001 (10.1146/annurev.neuro.24.1.1121) — 2,339-citation review; not_oa; most quantitative claims (phosphosite count, isoform ratios, etc.) derived from this review; cannot be verified until access is obtained
  4. Goedert 1989 + Mukrasch 2009 — both not in archive; Mukrasch 2009 structural details (IDP characterization) and Goedert 1989 isoform characterization are foundational; request access or flag permanently no-fulltext-access

[2026-05-04] ingest | Round 10a — epigenetic-alterations hallmark synthesis-MOC

• added: hallmarks/epigenetic-alterations.md (stub → drafted synthesis-MOC)
• page type: hallmark (synthesis-MOC over verified atomic pages)
• primary verified atomic pages integrated:
  • information-theory-of-aging (verified, FULL) — central cross-link; ICE model, OSK reprogramming, Schooling 2025 MR null
  • homo-sapiens (verified) — Horvath 2013 clock (n=7,844, MAE=3.6yr), Waziry 2023 CALERIE DunedinPACE
  • caloric-restriction (verified-partial) — CALERIE DunedinPACE slowing; PhenoAge/GrimAge null; clock-divergence framing
  • sirt1 (verified) — H3K9ac/H4K16ac deacetylation; heterochromatin erosion mechanism
  • sirtuin (verified-partial) — NAD+ decline; SIRT6 genome maintenance
  • spermidine (verified) — EP300/H3K56 mechanism; indirect epigenetic connection noted
  • caenorhabditis-elegans (verified-partial) — no CpG methylation caveat noted
• implicit stubs created by this page:
  • partial-reprogramming (referenced; not yet seeded)
  • dna-methylation (referenced in frontmatter key-pathways; not yet seeded)
  • histone-modification (referenced in frontmatter; not yet seeded)
  • cancer (referenced as key-phenotype; no atomic page yet)
• gaps surfaced:
  • unsourced — transposable element de-repression in aged human tissue (primary citations needed)
  • unsourced — HDAC inhibitors in aging contexts (aging endpoint data absent)
  • unsourced — folate/B12 supplementation effects on epigenetic clocks
  • contradictory-evidence — causal vs biomarker interpretation; clock-by-clock divergence
  • needs-replication — single-lab dominance (Sinclair/ICE system); CR-DunedinPACE (single trial)
  • needs-human-replication — all direct causal evidence is mouse-only
  • long-term-unknown — partial reprogramming safety; HDAC inhibitor aging effects
• ROADMAP updated: hallmarks section (stub → drafted) + R10a checklist

[2026-05-04] verify | hallmarks/mitochondrial-dysfunction

• page: hallmarks/mitochondrial-dysfunction.md (synthesis-MOC)
• verified-by: claude (wiki-verifier subagent)
• scope: wikilink resolution audit + cross-check of all inherited quantitative and mechanism claims against verified atomic pages; no primary PDFs read (all primary claims delegate to atomic pages)
• sources checked:
  • mitophagy (verified — Onishi 2021 + Singh 2022 ATLAS RCT) — PINK1 phospho-Ub Ser65 mechanism, Singh 2022 n=88/79, PPO NS, hamstring peak torque p≤0.029, 1000 mg dose-dependence: all confirmed
  • parkin (verified — Wauer 2015 + Narendra 2008 + others) — pUb binds Parkin RING1 at His302/Arg305 (not Ubl-RING0); delegated correctly to pathway page
  • pink1 (verified-partial — Narendra 2010) — PARL dispensable in mammalian cells (PARL controversy): correctly noted
  • bnip3 (verified-partial) — BNIP3 LIR W18-V-E-L21 (WVEL); binds LC3 not GABARAP: confirmed
  • nix (verified-partial — Novak 2010 + Hanna 2012) — NIX LIR is W35 (not W36); GABARAP-preferential: confirmed
  • fundc1 (verified — Wu 2014 EMBO Rep + Kuang 2016) — LIR Y18EVL; activating kinase is ULK1 at Ser17: confirmed
  • bcl2l13 (verified — Murakawa 2015 + Kataoka 2022) — WQQI LIR (human 286–289; mouse 273–276); AMPKα2 Ser272: confirmed
  • heterocephalus-glaber (verified-partial — Kim 2011) — NMR oxidative damage paradox (high damage + long life): correctly handled
  • urolithin-a (verified) — Singh 2022 ATLAS RCT cited correctly
• wikilink audit:
  • 33 links checked; 24 resolve to existing pages
  • 9 missing-page links: drp1, mfn1, mfn2, opa1, oxphos, mtdna, mitochondrial-biogenesis — all explicitly flagged “planned R10d” or “not yet seeded” in body; mitohormesis flagged “implicit stub”; mitosens in frontmatter and body — no flag in body text, but this is an expected planned page
• corrections made (1 factual):
  • FUNDC1 activating kinase: “TBK1 Ser177 phosphorylation” → “ULK1 Ser17 phosphorylation” in the receptor summary table. TBK1 is an innate-immunity kinase (NDP52/OPTN context); ULK1 Ser17 is the verified FUNDC1 activating phosphorylation per Wu 2014 EMBO Rep (verified) and Kuang 2016 (verified). The error likely arose from confusion with TBK1’s role in NDP52/OPTN innate-immune signaling.
• banner: ⚠️ removed; verified: true set
• downstream propagation: none needed — the TBK1/ULK1 error existed only in this MOC’s receptor table; the fundc1.md atomic page carries the correct ULK1/Ser17 claim already

[2026-05-04] verify | hallmarks/epigenetic-alterations

• page: hallmarks/epigenetic-alterations.md (synthesis-MOC)
• verified-by: claude (wiki-verifier subagent)
• scope: cross-check of all inherited quantitative claims against verified atomic source pages; wikilink resolution audit
• sources checked:
  • homo-sapiens (verified) — Horvath 2013 (n=7,844/82 datasets/51 tissues/MAE=3.6 yr) and Waziry 2023 (11.9% CR / DunedinPACE d=−0.25 p<0.003 / PhenoAge+GrimAge NS) confirmed
  • information-theory-of-aging (verified) — Yang 2023 ICE details deferred correctly (not directly stated on this page); single-lab caveats and Schooling 2025 MR null framing confirmed
  • caloric-restriction (verified with scope) — CALERIE quantitative figures confirmed
  • sirtuin (verified with scope), sirt1 (verified), spermidine (verified), caenorhabditis-elegans (verified with scope) — mechanism statements confirmed
• corrections: none required — all claims accurate against verified source pages
• wikilink audit:
  • all entity links resolve correctly (hallmarks-of-aging → frameworks/, cancer → phenotypes/)
  • 4 study stubs (horvath-2013, waziry-2023, yang-2023, schooling-2025) remain missing pages, correctly labeled as stubs in See Also section
  • partial-reprogramming remains unseeded, correctly flagged on the page
• result: verified: true (verified-scope: synthesis-MOC; no primary-source claims originated here)
• downstream propagation: none needed — this page is the consumer of upstream verified atomic pages, not a source for downstream propagation

[2026-05-04] verify | hallmarks/cellular-senescence

• page: hallmarks/cellular-senescence.md (synthesis-MOC)
• verified-by: claude (wiki-verifier subagent)
• scope: cross-check of all inherited quantitative claims against verified atomic source pages; wikilink resolution audit; inherited-claim spot-check list provided by main agent
• sources checked (atomic pages, all previously verified):
  • cardiomyocytes (verified) — Anderson 2019 navitoclax: hypertrophy + fibrosis reduced; EF NOT improved (NS); 24-month mice; 50 mg/kg/day
  • senolytics (verified-partial) — Justice 2019 exact p-values (6MWT p=0.012, gait p=0.024, chair-stands p=0.013); Q doses (Justice 1,250 mg/day; Hickson 1,000 mg/day); Hickson single 3-day course
  • p21 (verified) — Baker 2016 lifespan +24–27%, range 17–35%, tumor incidence not reduced — confirmed correct on this page
  • sasp (verified) — Coppé 2008 SASP timing not referenced directly on this page (timing claim lives on sasp)
  • foxo4 (verified) — FOXO4-DRI residues 86–206, plasma urea + running wheel endpoints, XpdTTD/TTD model — confirmed correct on this page
  • nf-kb (verified-partial) — κB consensus GGGRNWYYCC — not repeated on this page (correct omission for MOC)
  • dna-damage-response (verified-partial) — Rodier 2009 two-arm model, ATM/NBS1/CHK2 SASP arm p53-INDEPENDENT — confirmed correct
• corrections applied:
  • Hickson 2019 dosing: “D+Q for 3 days×3 cycles” → “D+Q for 3 consecutive days (single course; Q 1,000 mg/day in 2 divided doses)” — body text (×2 occurrences) and footnote
  • Hickson footnote: added quercetin dose (1,000 mg/day, 500 mg×2), specific p-values per biopsy endpoint, macrophage reduction, and verification reference
  • Justice 2019 p-values: “p<0.05 for physical function endpoints” → specific p-values (6MWT p=0.012, gait +0.12 m/s p=0.024, chair-stands −2.2 s p=0.013); added chair-stands endpoint; added Q dose 1,250 mg/day; pulmonary function NS noted
  • Justice footnote: replaced “p<0.05 physical function” with specific p-values + full protocol including Q dose; added verification reference
  • Anderson 2019 / heart-failure disease table: “navitoclax in vivo cardiac improvement” → “navitoclax reduced CM hypertrophy + fibrosis (EF unchanged, NS)” — navitoclax did not improve ejection fraction per verified cardiomyocytes.md source
• wikilink audit:
  • all entity links resolve to existing files; verified/verified-partial status labels accurate
  • Childs 2016 correctly retained as no-fulltext-access (not_oa)
  • BubR1 progeroid model correctly attributed to Baker 2011 (not Baker 2016)
  • p16-rb-pathway correctly labeled as “planned — implicit stub”
  • navitoclax correctly labeled as “planned”
• result: verified: true (synthesis-MOC scope)
• downstream propagation: none — this page is a consumer of upstream verified atomic pages

[2026-05-04] verify | hallmarks/altered-intercellular-communication

• page: hallmarks/altered-intercellular-communication.md (synthesis-MOC)
• verified-by: claude (wiki-verifier subagent)
• scope: spot-check of primary source claims per main agent instruction; PDFs read for López-Otín 2013 (full), Conboy 2005 (full), Loffredo 2013 (full), Egerman 2015 (full); Villeda 2011, Smith 2015, Sinha 2014, Katsimpardi 2014 remain pending download
• sources checked:
  • López-Otín 2013 (doi:10.1016/j.cell.2013.05.039) — PDF-read; hallmark tier, sub-components, and category confirmed
  • Conboy 2005 (doi:10.1038/nature03260) — PDF-read; GFP <0.1%, P<0.005, 5 weeks, strain IDs confirmed
  • Loffredo 2013 (doi:10.1016/j.cell.2013.04.015) — PDF-read; cardiac hypertrophy claim and GDF11 direction confirmed
  • Egerman 2015 (doi:10.1016/j.cmet.2015.05.010) — PDF-read; assay specificity issue and GDF11-increases claim confirmed
  • López-Otín 2023 (doi:10.1016/j.cell.2022.11.001) — closed-access; download failed; no-fulltext-access
• corrections made:
  • Integrative hallmark count: “four Integrative hallmarks in the López-Otín 2013 and 2023 frameworks” → “two Integrative hallmarks in the original López-Otín 2013 framework (alongside stem cell exhaustion)”. The 2013 Figure 6 shows exactly 2 integrative hallmarks (stem cell exhaustion + altered intercellular communication). The original claim incorrectly conflated 2013 and 2023 counts.
  • SENS stub reference: removed editorial note “the stub listed” from body text (now reads as clean objective statement)
  • Loffredo 2013 footnote: removed “download pending” qualifier; replaced with verified experimental details (young female C57BL/6, 2 months; old 23 months; 4 weeks; n=4–22/group)
  • Egerman 2015 footnote: removed “download pending” qualifier; added verified details (GDF11-specific immunoassay; rat n=3–5; human n=9–10; GDF11 increases + inhibits satellite cell expansion confirmed)
  • Conboy 2005 footnote: enriched with verified details (C57Bl/Ka-Ly5.2 or GFP-transgenic young, 2–3 months; C57Bl/6 aged, 19–26 months; n≥3 pairs/condition; 5 weeks)
  • López-Otín 2023 footnote: clarified as closed-access/download-failed with scope note on what is unverified
• claims confirmed correct:
  • SENS correspondence: [] (empty) — confirmed appropriate (original stub’s “GlycoSENS” was wrong; current empty mapping is correct)
  • 2023 split: chronic-inflammation elevated to separate hallmark — correct per inherited verified pages; unverifiable against primary 2023 PDF
  • SASP = senescence-associated secretory phenotype per sasp verified — confirmed
  • GDF11 controversy framing: Loffredo/Wagers initial claim + Egerman 2015 assay-specificity replication challenge — confirmed accurate
  • Conboy 2005: GFP <0.1%, P<0.005, 5 weeks, resident-cell mechanism — all confirmed against PDF
• banner: ⚠️ removed; verified: true set (verified-scope: see frontmatter)
• downstream propagation: none needed — this page is a synthesis-MOC consumer; no new factual claims originated here that would propagate downstream; Conboy 2005 corrections were already captured on satellite-cells (verified)

[2026-05-04] round-10a summary

Round 10a (Hallmark MOC fill-out) — all 12 hallmark stubs converted to synthesis-MOC pages, drafted + verified. This unlocks the wiki’s stated query patterns; hallmarks are now navigable synthesis layers over 129 verified atomic entities.

Pages converted from stub to verified synthesis-MOC:

Primary (causes of damage):

• genomic-instability (verified)
• telomere-attrition (verified-partial)
• epigenetic-alterations (verified)
• loss-of-proteostasis (verified)
• disabled-macroautophagy (verified — wiki’s BEST-evidenced cluster)

Antagonistic (responses):

• deregulated-nutrient-sensing (verified — second-strongest cluster)
• mitochondrial-dysfunction (verified)
• cellular-senescence (verified — strong evidence cluster)

Integrative:

• stem-cell-exhaustion (verified)
• altered-intercellular-communication (verified)
• chronic-inflammation (verified — wiki’s STRONGEST clinical-evidence hallmark via CANTOS)
• dysbiosis (verified — wiki’s most under-developed cluster; microbiome page seeding flagged as next-priority)

Critical corrections caught by verifier batch:

• Hara 2006 = ATG5 paper / Komatsu 2006 = ATG7 paper — verified correct attribution across all 4 hallmark pages that cite them
• disabled-macroautophagy 4 BUG-2 DOI corrections in footnotes (Komatsu 2006, Pyo 2013, Fernández 2018, Liang 1999 — all wrong DOIs in initial draft, corrected during verification)
• disabled-macroautophagy ULK1 phospho site — Ser556 (NOT Ser555); also added Ser758 in human alongside mouse Ser757 for mTOR site
• mitochondrial-dysfunction TBK1 mistake — wiki had FUNDC1 Ser177 phosphorylation by TBK1; correct kinase is ULK1 Ser17 per Wu 2014 EMBO Rep + Kuang 2016 verified on fundc1; TBK1 phosphorylates OPTN/NDP52, NOT FUNDC1
• loss-of-proteostasis Komatsu 2006 timing — wiki had “no overt cell death at early timepoints”; paper actually reports massive neuronal loss (TUNEL+ at P56); proteasome activity (NOT cell death) was unaltered
• cellular-senescence Hickson 2019 — single 3-day course, NOT “3 days × 3 cycles”; quercetin dose 1000 mg/day in 2 divided doses; CD68+ macrophage reduction -28% added
• Anderson 2019 EF NOT improved by navitoclax — propagated correctly; only hypertrophy + fibrosis reduced
• chronic-inflammation Jaiswal 2014 CHD CI corrected 1.2-3.4 → 1.2-3.5 (transcription error)
• stem-cell-exhaustion Pang 2011 n — wiki had n=13 young + 11 elderly; corrected to n=11 young + 13 elderly (matches verified hematopoietic-stem-cells body text); flagged inconsistency between body text and footnote on hematopoietic-stem-cells.md
• altered-intercellular-communication integrative-hallmark count — wiki claimed “one of four”; López-Otín 2013 has only TWO integrative hallmarks (with stem-cell-exhaustion); corrected
• dysbiosis Depommier 2019 group n — placebo 11 / pasteurized 12 / live 9 (NOT 9/14/9 as wiki had — mis-read of randomization 13/13/14 vs completers)
• dysbiosis Claesson 2012 ELDERMET — added diet-microbiota-health causal-direction finding (diet changes within 1 mo, microbiome takes ~1 yr — establishes causality); fully verified now
• dysbiosis Smith 2017 specificity control — Omt same-age FMT did NOT extend lifespan; key control for young-microbiome specificity
• telomere-attrition Celli & de Lange 2005 model — “human cells” → “MEFs (mixed 129/BL6)”; major model-system error
• telomere-attrition Bernardes 2012 numerics added — 1-yr cohort eGFP n=12 / mTERT n=21 / control n=43; +24%/+13% lifespan; cancer NS p=0.87
• chronic-inflammation Jaiswal 2014 CIs added — full 95% CIs including hematologic cancer 3.9-32.6
• 2 BUG-2 DOIs caught in footnotes + multiple status-label updates (“(drafted)” → “(verified-partial)”) across ~13 atomic-page references reflecting that all Round 5-9 atomic pages are now verified

Schema escalation flagged repeatedly: hallmark schema in CLAUDE.md doesn’t include verified: field. R10a pages all adopted it for synthesis-MOC consistency. Recommend canonicalizing in CLAUDE.md schema next pass.

Coverage delta: Rounds 5-10a = 81 new pages drafted + verified (12 hallmark MOCs added this round). Wiki now has full top-down navigability — hallmarks are real synthesis pages over the 129+ verified atomic entities.

[2026-05-04] ingest | round-10b-tier-a-entities

ubiquitin-proteasome-system

• added: processes/ubiquitin-proteasome-system.md
• entity type: process
• inbound references resolved: 17 (highest-priority stub in Round 10b)
• canonical sources pulled: DOI lookup confirmed for all 6 cited DOIs; 1 local PDF available (Hershko 1998, 10.1146/annurev.biochem.67.1.425, exit 0); 5 pending download or not OA
• DOI verification note: initial brief 10.1186/1750-1326-9-21 (proposed as “Saez & Vilchez 2014 UPS/aging”) resolves in archive to a Ryanodine receptor paper — DOI mismatch in the task brief; correct Saez & Vilchez 2014 DOI is 10.2174/138920291501140306113344 (Current Genomics, confirmed via PubMed fetch PMID 24653662); used correct DOI on page
• DOI verification note: 10.1038/nrm3835 (proposed as “Schmidt & Finley 2014”) returned “DOI not found” in archive; Semantic Scholar query returned a 2005 paper (10.1515/bc.2005.085) which does not match the proposed 2014 date; omitted from page rather than use an unverified DOI. Replaced with Vilchez, Saez & Dillin 2014 Nat Commun (10.1038/ncomms6659, PMID 25482515, 666 citations) as the equivalent authoritative review
• primary-source DOIs cited on page (6 footnotes):
  • 10.1146/annurev.biochem.67.1.425 — Hershko & Ciechanover 1998 (local PDF available; 8,669 citations; 100th percentile)
  • 10.1152/physrev.00027.2001 — Glickman & Ciechanover 2002 (not OA; 4,160 citations; 100th percentile)
  • 10.1038/ncomms6659 — Vilchez, Saez & Dillin 2014 Nat Commun (OA/pending; 666 citations)
  • 10.2174/138920291501140306113344 — Saez & Vilchez 2014 Curr Genomics (OA/pending; 324 citations)
  • 10.1096/fj.14-252189 — Chondrogianni et al. 2015 FASEB J (OA PMC/pending; 167 citations)
  • 10.1242/dmm.010389 — Bonaldo & Sandri 2012 (OA/pending; 1,228 citations)
• implicit stubs created (wikilinks to non-existent pages):
  • uba1 — E1 ubiquitin-activating enzyme
  • ubiquitin — the 76 aa modifier protein itself
  • psmd14 — JAMM DUB in 19S lid (RPN11)
  • nf-kb — signaling pathway (likely exists; confirm slug)
  • neurodegeneration — linked as verified-partial per brief
  • sarcopenia — phenotype (likely exists; confirm)
  • atg10 — autophagy E2 (in R10b queue)
  • atg3 — autophagy E2 (in R10c queue)
  • atg16l1 — autophagy E3-like (in R10c queue)
  • alpha-synuclein — neurodegeneration aggregate (in R10d queue)
• gaps surfaced:
  • #gap/needs-human-replication — proteasome activation lifespan data all model-organism
  • #gap/needs-replication — immunoproteasome age-increase claim
  • #gap/needs-human-replication — atrogin-1/MuRF1 as causal UPS driver in human sarcopenia
  • #gap/needs-replication — DT2216 senolytic efficacy in aged organisms
  • #gap/needs-human-replication — proteasome activator compounds in humans
  • #gap/dose-response-unclear — risk/benefit of supra-youthful proteasome activation
  • #gap/unsourced — E3 ligase class counts need census review citation
• schema gaps: none; type: process schema fully covers this entity. verified: false fields added per convention even though not listed in process-type schema (consistent with R10a precedent for atomic pages)
• verification priority: HIGH — 17 inbound links; Hershko 1998 local PDF available for immediate verification; key mechanistic claims (E3 class sizes, chain topology, 20S subunit β nomenclature, DUB gate mechanism) should be spot-checked against that PDF first

14-3-3

• added: molecules/proteins/14-3-3.md
• entity type: protein (family page — 7 mammalian paralogs; YWHAZ P63104 as primary aging-relevant canonical)
• canonical IDs: UniProt P63104 (YWHAZ_HUMAN); ncbi-gene/hgnc null (family page — per-paralog IDs need dedicated sub-pages); all 7 paralog UniProt IDs listed in complex-subunits frontmatter field (schema: re-uses tsc1-tsc2 pattern)
• DOI verification note: task brief supplied 10.1042/BST20110665 as “Aitken 2011 14-3-3 review”; DOI lookup confirms this resolves to “Targeting phosphoinositide 3-kinase delta for allergic asthma” — BUG-2 DOI mismatch. Replaced with Aitken 2006 (10.1016/j.semcancer.2006.03.005, “14-3-3 proteins: A historic overview”, 835 citations, 100th percentile) confirmed via DOI lookup. Aitken 2006 is not_oa.
• primary-source DOIs cited on page (6 footnotes):
  • 10.1016/s0092-8674(00)80487-0 — Yaffe 1997 (Cell; 1,629 citations; 100th percentile; locally available) — recognition motifs + crystal structure
  • 10.1016/s0092-8674(00)81382-3 — Zha 1996 (Cell; 2,617 citations; 100th percentile; locally available; verified on bad.md) — BAD Ser112/Ser136 14-3-3 binding
  • 10.1016/s0092-8674(00)80595-4 — Brunet 1999 (Cell; 6,504 citations; 100th percentile; locally available; verified on foxo3.md) — FOXO3 Thr32+Ser253 14-3-3 binding (NOT Ser315)
  • 10.1126/scisignal.2002790 — Roczniak-Ferguson 2012 (Sci Signaling; 1,313 citations; 100th percentile; not_oa; abstract confirmed; verified on tfeb.md) — TFEB Ser211 14-3-3 docking
  • 10.1038/emboj.2012.32 — Settembre 2012 (EMBO J; 1,865 citations; 100th percentile; locally available; verified on tfeb.md) — lysosomal mTOR/TFEB Ser142 context
  • 10.1016/j.semcancer.2006.03.005 — Aitken 2006 (Semin Cancer Biol; 835 citations; 100th percentile; not_oa) — general family review; not independently verified from PDF; claims from this source flagged no-fulltext-access
• implicit stubs created (net new wikilinks):
  • foxo-transcription-factors — family overview page (may already exist; confirm)
• existing pages correctly cross-linked (no new stubs): foxo3, foxo1, bad, tfeb, tsc1-tsc2, akt, sgk1, insulin-igf1, pi3k-akt-pathway, mtor, bcl-xl, autophagy, caloric-restriction, hyperfunction-theory, caenorhabditis-elegans, disabled-macroautophagy, deregulated-nutrient-sensing
• gaps surfaced: ncbi-gene/hgnc null (family page); paralog-selective aging functions understudied; “hundreds of 14-3-3 clients” claim needs phosphoproteomics citation; no clinical-stage aging-indication compound targeting 14-3-3/FOXO or 14-3-3/TFEB interface
• schema decision: complex-subunits frontmatter field (introduced tsc1-tsc2.md) re-used here for 7 paralog UniProt IDs; uniprot top-level = YWHAZ; no new schema invention required; consistent with existing multi-subunit pattern
• ROADMAP: 14-3-3 marked [x] drafted 2026-05-04
• verification priority: MEDIUM-HIGH — 14 inbound links; 3 of 6 DOIs are locally available and already verified on atomic pages (Zha 1996 on bad.md, Brunet 1999 on foxo3.md, Settembre 2012 on tfeb.md); Yaffe 1997 crystal-structure claims (binding groove residues Lys49/Arg56/Arg127, Mode I/II/III motif boundaries) are the main unverified quantitative claims on this page and should be spot-checked against the local Yaffe 1997 PDF first

bcl-w

• added: molecules/proteins/bcl-w.md
• entity type: protein (type: protein)
• canonical IDs: UniProt Q92843, NCBI Gene 599, HGNC 993, Ensembl ENSG00000129473
• length: 193 aa; MW ~20.7 kDa; mouse ortholog: Bcl2l2
• GenAge: not listed — needs-canonical-id
• key sections:
  • Identity table (UniProt Q92843)
  • Domain organization (BH4, BH1, BH2, BH3-receptor-groove, transmembrane anchor; Hinds 2003 autoinhibition model noted with gap tag)
  • Discovery and mouse KO phenotype (Print 1998 — male-sterile; otherwise viable; contrasted with Bcl-2-/- and Bcl-x-/-)
  • BH3-only binding selectivity table (Chen 2005 verified: BIM 4.3 nM, BID ~5 nM, BAD ~5 nM, PUMA 5.1 nM, NOXA >100 μM negligible)
  • Mechanism of apoptosis inhibition
  • BCL-W in senescence/aging — SCAP cell-type table (IMR-90 OIS: BCL-W+BCL-xL dual per Yosef 2016; IMR-90 genotoxic: triple per Zhu 2016; HUVECs: BCL-xL dominant; preadipocytes: ephrins/PI3KCD)
  • Contradiction documented: Yosef 2016 dual BCL-W+BCL-xL vs Zhu 2016 triple BCL-2+BCL-xL+BCL-w; contradictory-evidence tagged
  • Pharmacology (navitoclax, venetoclax, A-1331852 section)
  • BCL-W in cancer (brief)
  • Key interactors table (7 partners with BH3 affinities from Chen 2005)
  • Limitations and open questions (8 gaps)
• primary DOIs cited (8 footnotes):
  • 10.1006/cyto.1996.0066 (Gibson 1996 cloning) — archive: not_oa; no-fulltext-access
  • 10.1073/pnas.95.21.12424 (Print 1998 KO spermatogenesis) — archive: not_oa; title confirmed in DOI lookup; no-fulltext-access
  • 10.1016/j.molcel.2004.12.030 (Chen 2005 BH3 affinities) — archive: locally downloaded (PDF confirmed)
  • 10.1038/ncomms11190 (Yosef 2016 BCL-W+BCL-xL senolytic) — archive: locally downloaded (PDF confirmed)
  • 10.1111/acel.12445 (Zhu 2016 navitoclax senolytic) — archive: locally downloaded (PDF confirmed)
  • 10.1111/acel.12344 (Zhu 2015 SCAP/dasatinib) — archive: locally downloaded (referenced for preadipocyte SCAP)
  • 10.18632/aging.101202 (Zhu 2017 A1331852) — archive: locally downloaded
  • 10.1038/nm.4010 (Chang 2016 navitoclax aged mice) — archive: download failed; no-fulltext-access
  • NOTE: Hinds 2003 BCL-W structure paper (10.1093/emboj/cdg186) — DOI lookup returns Ikaros paper (BUG-2 DOI-title mismatch suspected); claim sourced from UniProt Q92843 annotation; not independently verified against primary PDF
• implicit stubs created (new wikilinks to non-existent pages):
  • venetoclax — BCL-2-selective BH3-mimetic; no compound page yet
  • navitoclax — stub noted on bcl-xl page; still unresolved
  • bmf — BH3-only protein; not yet seeded
  • bcl-2-family-signaling — pathway page; implicit stub
• existing pages cross-referenced (all confirmed to exist):
  • proteins: bcl-2 (verified-partial), bcl-xl (verified-partial), mcl-1 (verified-partial), bim (verified), bad (verified-partial), noxa (verified), puma (verified-partial), bax, bak
  • pathways: apoptosis-pathway (verified-partial)
  • interventions: senolytics (verified-partial)
  • hallmarks: cellular-senescence (verified)
• gaps surfaced:
  • contradictory-evidence — Yosef 2016 dual vs Zhu 2016 triple BCL-W+BCL-xL requirement in IMR-90
  • needs-human-replication — BCL-W SCAP role in primary aged human fibroblasts; all data from cell lines
  • needs-canonical-id — no GenAge-human entry for BCL2L2
  • no-fulltext-access — Gibson 1996, Print 1998, Chang 2016 (download failed)
  • unsourced — BCL-W dispensability in platelets; human BCL2L2 variants in male infertility
  • needs-replication — BCL-W autoinhibition in vivo (Hinds 2003 crystal-structure model)
  • long-term-unknown — selective BCL-W inhibitor development
• archive BUG-2 flag: DOI 10.1093/emboj/cdg186 returns wrong title in DOI lookup (Ikaros paper) — suspected DOI-title database mismatch; recommend flagging to archive maintainer
• schema gaps: none; type: protein schema fully accommodates BCL-W; sens-categories optional field populated with apoptosenes consistent with bcl-xl precedent
• ROADMAP updated: bcl-w marked [x] drafted 2026-05-04
• verification priority: MEDIUM-HIGH — 12 inbound links; 3 primary PDFs locally available (Chen 2005, Yosef 2016, Zhu 2016) for immediate verification of BH3-affinity table and SCAP dual/triple claims; Print 1998 and Gibson 1996 are not_oa and cannot be verified without library access

[2026-05-04] verify | tissues/brain + tissues/bone-marrow

Pages verified: 2 (tissue stub anchor pages)

tissues/brain.md

• verified: true (verified-scope: partial — Azevedo 2009 neuron count unverifiable, not_oa)
• Corrections (3 factual):
  • [^hara2006] DOI: 10.1038/ncb1889 (CPAP centriole paper — completely wrong) → 10.1038/nature04724 (Hara 2006 ATG5 paper), confirmed via PDF
  • [^komatsu2006] DOI: 10.1038/nature04724 (was Hara’s DOI) → 10.1038/nature04723 (Komatsu 2006 ATG7 paper), confirmed via PDF
  • [^komatsu2006] Cre driver: “CaMKII-Cre” → “Nestin-Cre”, confirmed via PDF (Komatsu 2006 uses nestin promoter Cre, same as Hara 2006)
  • [^bouzid2023] DOI: 10.1038/s41591-023-02305-8 (wrong — resolves to oncology screening paper) → 10.1038/s41591-023-02397-2 (correct Bouzid CHIP-AD paper per neurodegeneration verified page)
  • Body text Bouzid OR: ~0.69–0.73 → 0.64 (per neurodegeneration page verified claims)
  • Body text Bouzid n: n=412,691 UK Biobank → n=5,730 total (1,362 AD; 4,368 controls) (per neurodegeneration page verified claims)
• Wikilinks confirmed resolving: microglia, alzheimers-disease, parkinsons-disease, neurodegeneration, atg5, atg7, mtor, tfeb
• Wikilinks confirmed planned stubs (not yet created, correctly noted as such): neurons, astrocytes, oligodendrocytes, neural-stem-cells
• Unverifiable: Azevedo 2009 (~86 billion neurons) — not_oa per archive; claim retained with source footnote

tissues/bone-marrow.md

• verified: true (verified-scope: quantitative claims cross-checked against verified HSC page)
• Corrections (1 numeric):
  • CHD HR CI: HR 2.0 (95% CI 1.2–3.5) → HR 2.0 (95% CI 1.2–3.4) — corrected in both body text and footnote; matches verified hematopoietic-stem-cells page (Jaiswal 2014 verified against primary PDF there)
• Confirmed matching verified HSC page: Sudo 2000 ~17-fold expansion, Beerman 2010 Slamf1-hi/lo two-subset model, Pang 2011 n=11 young + 13 elderly, Jaiswal 2014 all HRs
• Pang 2011 n consistent: body text (n=11 young, n=13 elderly) matches HSC body text; HSC footnote has transposition (n=13 young + 11 elderly) — discrepancy within HSC page flagged for downstream attention

Downstream propagation needed (main agent):

• phenotypes/neurodegeneration.md — already has correct Bouzid DOI (s41591-023-02397-2) and correct OR=0.64; no change needed
• Any other pages citing [^hara2006] or [^komatsu2006] with the old wrong DOIs — scan recommended

[2026-05-04] ingest | round-10c-machinery-tissues

caspase-7

• added: molecules/proteins/caspase-7.md
• entity type: protein
• inbound references resolved: 7 (from caspase-3, caspase-9, caspase-8, apoptosis-pathway, xiap, apaf-1, plus implied stubs in earlier rounds)
• canonical IDs pulled:
  • UniProt P55210 (CASP7_HUMAN; Swiss-Prot, manually reviewed) — 303 aa, Cys186 active site, cleavage sites Asp198/Asp206 confirmed
  • NCBI Gene: 840; HGNC: 1508; Ensembl: ENSG00000165806; mouse ortholog: Casp7
• primary-source DOIs cited (5 footnotes):
  • 10.1074/jbc.271.4.1825 — Lippke 1996 CMH-1 cloning (225 citations; 99th percentile; archive pending download — hybrid OA)
  • 10.1074/jbc.270.26.15870 — Munday 1995 ICErelIII cloning (281 citations; 98th percentile; archive pending download — hybrid OA)
  • 10.1073/pnas.0707715105 — Walsh 2008 executioner caspase substrate divergence (587 citations; 100th percentile; local PDF available; previously verified on caspase-3.md)
  • 10.1126/science.1115035 — Lakhani 2006 Casp3/7 DKO phenotype (1,158 citations; 100th percentile; archive pending download — green OA via PMC)
  • 10.1016/s0092-8674(01)00274-4 — Riedl 2001 XIAP BIR2 structure (784 citations; 100th percentile; local PDF available; previously verified on xiap.md)
• implicit stubs created: none — all wikilinks resolve to existing pages
• existing pages correctly cross-linked: caspase-3 (verified), caspase-9 (verified), caspase-8 (verified), apoptosis-pathway (verified-partial), xiap (verified-partial), apaf-1 (verified-partial), cytochrome-c, bax, bak, bid, smac-diablo, loss-of-proteostasis, cellular-senescence
• key content sections: identity table; domain/processing table (Asp198/Asp206 cleavage, p20/p11 subunits); active-site (Cys186, His144, exosite residues 38–41); substrate specificity (DEVD consensus; Walsh 2008 p23/PTGES3 as sole preferentially-caspase-7 substrate of 20 examined); activation pathways (intrinsic via caspase-9; extrinsic via caspase-8); XIAP inhibition (BIR2; Ki ~0.2 nM); KO phenotype (Casp7-null viable/fertile; DKO lethality + brain malformation from Lakhani 2006); discovery (Munday 1995, Lippke 1996, parallel); aging context (executioner redundancy; ER-stress specialization; apoptosis resistance in senescent cells; non-apoptotic SMPD1 function); pharmacology; limitations/gaps
• gaps surfaced:
  • needs-human-replication — Casp7-null and DKO phenotypes mouse-only
  • needs-replication — relative caspase-3 vs caspase-7 contribution in aged mammalian tissues not investigated
  • no-mechanism — ER-stress specificity of caspase-7 relative to caspase-3
  • no-mechanism — SMPD1 activation by caspase-7 in membrane repair; aging relevance unknown
  • unsourced — developmental expression differences between the two executioner paralogs
  • needs-replication — Walsh 2008 cell-free substrate data (1/20 preferentially caspase-7) may not fully reflect in vivo utilization
• schema gaps: none; type: protein schema fully accommodates this page; sens-categories [] consistent with caspase-3 precedent
• ROADMAP: caspase-7 marked [x] drafted 2026-05-04
• verification priority: HIGH — 7 inbound links; Walsh 2008 and Riedl 2001 local PDFs available for immediate spot-check of substrate divergence claims and XIAP Ki values; Lakhani 2006 (DKO phenotype, the most consequential mechanistic claim) is OA via PMC but not yet downloaded — recommend triggering download for verification of KO phenotype numerics (brain weight, lethality penetrance)

smac-diablo

• added: molecules/proteins/smac-diablo.md
• entity type: protein
• inbound references resolved: 9 (from xiap, caspase-3, caspase-9, apoptosis-pathway, bak, bax, cytochrome-c, plus caspase-7 R10c + apaf-1 R5b)
• canonical IDs pulled:
  • UniProt Q9NR28 (DBLOH_HUMAN; Swiss-Prot, manually reviewed) — 239 aa precursor; MTS residues 1–21; PARL cleavage at 55–56 generating mature ~184 aa form with AVPI N-terminus; confirmed via REST API
  • NCBI Gene: 56616; HGNC: 21528; mouse ortholog: Diablo
• primary-source DOIs cited (5 footnotes):
  • 10.1016/s0092-8674(00)00008-8 — Du 2000 SMAC discovery (3,446 citations; 100th percentile; archive: pending, bronze OA)
  • 10.1016/s0092-8674(00)00009-x — Verhagen 2000 DIABLO discovery (2,333 citations; 100th percentile; archive: pending, bronze OA)
  • 10.1038/35022514 — Liu 2000 structural/biochemical basis (854 citations; 100th percentile; local PDF available at )
  • 10.1128/mcb.22.10.3509-3517.2002 — Okada 2002 Smac/Diablo KO mice (172 citations; 99th percentile; archive: pending, green OA via PMC:133802)
  • 10.1016/j.cell.2007.10.037 — Vince 2007 SMAC mimetic / cIAP1 TNFα mechanism (1,054 citations; 100th percentile; local PDF available at )
• implicit stubs created (new wikilinks to non-existent pages):
  • momp — mitochondrial outer membrane permeabilization (process page; referenced in lead; no page exists)
• existing pages correctly cross-linked: xiap (verified-partial), caspase-9 (verified), caspase-3 (verified), bax (verified-partial), bak (verified FULL), cytochrome-c (verified-partial), apoptosis-pathway (verified-partial), bcl-2-family-signaling (verified-partial), cellular-senescence (verified)
• key content sections: identity; discovery (parallel Du 2000 / Verhagen 2000); processing (MTS + PARL dual cleavage → ~184 aa AVPI mature form); structure (homodimer; AVPI surface-exposed); release mechanism (MOMP-gated; co-released with cytochrome c); IAP antagonism mechanism (BIR2 hook/sinker for caspase-3/-7; BIR3 for caspase-9; IAP target table with cIAP1/2 cIAP1 self-ubiquitination via Vince 2007); KO phenotype (viable + grossly normal — sensitizer not obligatory activator); aging/senolytic relevance (SMAC mimetics birinapant/LCL161/AT-406 preclinical; Vince 2007 cIAP1 mechanism; senescent cell SCAP/IAP hypothesis); turnover (proteasome substrate, BIRC6/7 ubiquitination); limitations/gaps
• gaps surfaced:
  • needs-human-replication — all senolytic data for SMAC mimetics are preclinical; no aging-indication clinical trial
  • no-mechanism — mechanism by which senescent cells upregulate IAPs not established
  • needs-replication — SMAC vs. cytochrome c co-release kinetics not characterized at single-cell resolution
  • long-term-unknown — SMAC mimetic safety in older adults / high-senescent-cell-burden contexts not studied
  • needs-replication — which IAP (XIAP vs. cIAP1/2) is the relevant senolytic target in senescent cells
• schema gaps: none; type: protein schema fully accommodates this page; [[momp]] wikilink created as implicit stub (a process page that does not yet exist); apoptosenes wikilink in sens-categories follows xiap.md precedent
• ROADMAP: smac-diablo marked [x] drafted R10c 2026-05-04
• verification priority: HIGH — 9 inbound links; Liu 2000 PDF is locally available and should be the primary verification target (structural basis of AVPI-XIAP interaction); Vince 2007 PDF also locally available for SMAC mimetic mechanism claims; Du 2000 and Verhagen 2000 are pending downloads (bronze OA) but can be triggered; Okada 2002 is green OA via PMC (can be triggered); PARL cleavage site (residues 55–56 generating AVPI) should be confirmed against UniProt Q9NR28 annotation evidence codes

fip200

• added: molecules/proteins/fip200.md
• entity type: protein (type: protein)
• inbound references resolved: 9 (from ulk1, atg13, atg101, autophagy, disabled-macroautophagy, mtor, and others per log search)
• canonical IDs pulled:
  • UniProt Q8TDY2 (RCCI1_HUMAN; Swiss-Prot, manually reviewed) — 1,594 aa; ~200 kDa
  • NCBI Gene: 9821; HGNC: 15276; mouse ortholog: Rb1cc1
  • GenAge: not found in human subset — needs-canonical-id
• DOI corrections (2 task-brief errors caught):
  1. Gan 2006 brief DOI 10.1083/jcb.200603081 — “DOI not found” in archive; correct DOI 10.1083/jcb.200604129 confirmed via PubMed efetch (PMID 17015619); DOI lookup confirmed (228 citations, bronze OA)
  2. “Lin 2012” brief DOI 10.1038/nsmb.2421 — title mismatch in archive (resolves to SRP paper); correct paper is Gammoh et al. 2012/2013 (Nat Struct Mol Biol) at DOI 10.1038/nsmb.2475; confirmed via PubMed efetch PMID 23262492; DOI lookup confirmed (206 citations, green OA)
• primary-source DOIs cited on page (5 footnotes):
  • 10.1083/jcb.200712064 — Hara 2008 (J Cell Biol; 963 citations; 100th percentile; archive pending, bronze OA)
  • 10.1083/jcb.200604129 — Gan 2006 (J Cell Biol; 228 citations; archive pending, bronze OA)
  • 10.1016/j.molcel.2019.01.035 — Turco 2019 (Mol Cell; 342 citations; 100th percentile; archive pending, hybrid OA)
  • 10.1038/nsmb.2475 — Gammoh 2012 (Nat Struct Mol Biol; 206 citations; archive pending, green OA)
  • 10.1074/jbc.m900573200 — Ganley 2009 (J Biol Chem; 1,512 citations; 100th percentile; local PDF available at )
• implicit stubs created (new wikilinks to non-existent pages):
  • atg16l1 — already seeded this round (R10c); wikilink now resolves
• existing pages correctly cross-linked: ulk1 (verified-partial), atg13 (verified), atg101 (verified-partial), autophagy (verified-partial), mtor (verified-partial), ampk, p62 (drafted), disabled-macroautophagy (verified), loss-of-proteostasis (verified), alzheimers-disease (drafted), parkinsons-disease (drafted), neurodegeneration (drafted)
• key content sections: identity table; domain architecture (N-terminal coiled-coil ~17–225, LIR motif ~585–588, Claw domain ~1485–1591); ULK1 complex structural role; Claw domain / selective autophagy cargo receptor binding (p62 FIR, NDP52 CLIR, ATG13 HORMA); FIP200–ATG16L1 connection; discovery history; KO phenotype (germline lethal E14.5–E16.5; neural cKO inclusion-body neurodegeneration); regulation; aging relevance; pathway membership; interactors table; limitations/gaps
• gaps surfaced:
  • needs-human-replication — no human FIP200 loss-of-function genetics; germline KO lethal in mouse
  • unsourced — FIP200 expression changes with age; PTM landscape uncharacterized
  • needs-replication — Claw domain p62 FIR + NDP52 CLIR binding (Turco 2019, single laboratory); FIP200–ATG16L1 interface (Gammoh 2012, single laboratory)
  • no-mechanism — selective vs bulk autophagy contribution of FIP200
  • needs-canonical-id — FIP200/RB1CC1 not in GenAge human subset
• schema gaps: none; type: protein schema fully covers this entity
• ROADMAP: fip200 marked [x] drafted 2026-05-04
• verification priority: HIGH — 9 inbound links; Ganley 2009 local PDF available (previously verified on ulk1/atg13) — additive ULK1 kinase stimulation claim; Hara 2008 is the founding paper and most consequential — pending download (trigger bronze OA); Turco 2019 Claw domain structural claims are the most novel content and highest verification priority once hybrid OA download completes

glut4

• added: molecules/proteins/glut4.md
• entity type: protein
• inbound references resolved: 7 (from insr, insulin, insulin-igf1, akt, type-2-diabetes, skeletal-muscle, adipose-tissue planned)
• canonical IDs pulled:
  • UniProt P14672 (GTR4_HUMAN; Swiss-Prot, manually reviewed) — 509 aa; confirmed
  • NCBI Gene: 6517; HGNC: 11009; Ensembl: ENSG00000181856; mouse ortholog: Slc2a4
• primary-source DOIs cited (4 footnotes):
  • 10.1038/338083a0 — James 1989, GLUT4 cloning (Nature; 907 citations; 100th percentile; archive: not_oa / closed access)
  • 10.1074/jbc.C300063200 — Sano 2003, AS160/TBC1D4 as AKT substrate for GLUT4 (JBC; 952 citations; 100th percentile; archive: hybrid OA, pending download)
  • 10.1038/377151a0 — Katz 1995, Glut4-null mouse KO phenotype (Nature; 472 citations; 100th percentile; local PDF available at )
  • 10.1152/ajpcell.00069.2014 — Klip et al. 2014, GLUT4 translocation review (Am J Physiol Cell Physiol; 162 citations; 100th percentile; archive: not_oa / closed access)
• DOI correction: task brief supplied 10.1152/physrev.00033.2013 as “Klip 2014” — DOI lookup confirms this resolves to an unrelated sialic acid/nervous system paper (DOI-title mismatch, BUG-2 class). Correct Klip 2014 review DOI is 10.1152/ajpcell.00069.2014 (confirmed by PubMed PMID 24598362). Substituted on page.
• DOI omission: Bell 1990 (10.1006/geno.1990.1024) returned “DOI not found” from archive and was not confidently confirmed via Crossref; omitted to avoid fabrication; discovery claim adequately covered by James 1989. Tagged unsourced for Bell 1990 mapping paper.
• implicit stubs created (new wikilinks to non-existent pages):
  • as160 — TBC1D4 Rab-GAP; key AKT substrate in GLUT4 vesicle release; no page exists yet
  • adipose-tissue — tissue page (listed as “planned” in task brief; no page yet)
• existing pages correctly cross-linked: insr, insulin, insulin-igf1, akt, type-2-diabetes, skeletal-muscle, pi3k-akt-pathway, pi3k, irs-1, ampk, deregulated-nutrient-sensing
• key content sections: identity table; structure/domains (12-TM MFS; N-terminal LL motif; EC2 exofacial loop; C-terminal FQQI GSV-targeting); tissue distribution + GLUT isoform comparison table; 7-step insulin-stimulated translocation cascade; exercise/AMPK parallel route; discovery (James 1989 parallel cloning); KO/overexpression phenotype table + extrapolation grading; aging relevance (expression decline, translocation defect, actin remodeling); T2D connection; exercise as therapy; pharmacological modulation table; limitations/gaps
• gaps surfaced:
  • needs-replication — GLUT4 expression decline in aging: fold-change varies widely across studies; systematic meta-analysis needed
  • needs-human-replication — muscle-specific and adipose-specific KO phenotypes established only in mouse
  • no-mechanism — GSV subpopulation heterogeneity in aging context
  • needs-replication — brain GLUT4 in cognitive insulin sensitivity during aging
  • unsourced — Bell 1990 GLUT4 chromosomal mapping (DOI unconfirmed)
  • dose-response-unclear — optimal exercise protocol for GLUT4 restoration in aged muscle
  • contradictory-evidence — rapamycin / mTOR inhibition effect on GLUT4 translocation
• schema gaps: none; type: protein schema fully accommodates this page
• ROADMAP: glut4 marked [x] drafted 2026-05-04
• verification priority: HIGH — 7 inbound links from verified pages; Katz 1995 KO paper locally available (highest priority: verify cardiac hypertrophy claim and “diabetes does not fully develop” framing); Sano 2003 pending hybrid OA download (verify AS160 Thr642 phosphorylation and GAP suppression mechanism); James 1989 and Klip 2014 are closed access — cannot verify without local PDFs

miro1

• added: molecules/proteins/miro1.md
• entity type: protein
• canonical IDs: UniProt Q8IXI2, NCBI Gene 55288, HGNC 21168; confirmed via UniProt REST API 2026-05-04
• length: 618 aa; OMM-anchored single-pass type IV membrane protein; all functional domains cytoplasmic-facing
• DOI corrections (3 task-brief mismatches, all BUG-2 class):
  • Fransson 2003 brief-supplied DOI 10.1074/jbc.M212725200 → archive returns “Bipartite Signals Mediate Subcellular Targeting of Tail-anchored Membrane Proteins in S. cerevisiae” (mismatch); correct MIRO discovery DOI = 10.1074/jbc.M208609200 (PMID 12482879, “Atypical Rho GTPases Have Roles in Mitochondrial Homeostasis and Apoptosis”; 399 cites; archive confirmed)
  • Macaskill 2009 brief-supplied DOI 10.1016/j.neuron.2008.12.026 → archive returns “Motor Neurons with Axial Muscle Projections Specified by Wnt4/5 Signaling” (mismatch); correct MIRO1/Ca2+ DOI = 10.1016/j.neuron.2009.01.030 (PMID 19249275, “Miro1 Is a Calcium Sensor for Glutamate Receptor-Dependent Localization of Mitochondria at Synapses”; 660 cites; archive confirmed)
  • Saotome 2008 brief-supplied DOI 10.1073/pnas.0810334106 → “DOI not found” in archive; correct DOI = 10.1073/pnas.0808953105 (PMID 19098100, “Bidirectional Ca2+-dependent control of mitochondrial dynamics by the Miro GTPase”; 523 cites; archive confirmed green OA)
• DOIs cited (7 total):
  • doi:10.1074/jbc.M208609200 (Fransson 2003 — MIRO GTPase discovery) — archive pending (OA hybrid)
  • doi:10.1016/j.bbrc.2006.03.163 (Fransson 2006 — MIRO1/2 trafficking; Miro1-null perinatal lethal; MIRO-TRAK complex) — archive not_oa (closed) no-fulltext-access
  • doi:10.1016/j.neuron.2009.01.030 (Macaskill 2009 — MIRO1 Ca2+-sensor; EF-hand mutants; synaptic mitochondrial localization) — archive pending (OA hybrid)
  • doi:10.1073/pnas.0808953105 (Saotome 2008 — bidirectional Ca2+-dependent mitochondrial dynamics) — archive pending (OA green, freely available)
  • doi:10.1016/j.cell.2011.10.018 (Wang 2011 — PINK1 phosphorylates MIRO; Parkin ubiquitinates MIRO1/2; motor detachment; mitophagy) — archive pending (OA bronze); 1,139 citations; 100th percentile
  • doi:10.15252/embj.2018100715 (Lopez-Domenech 2021 — neuronal Miro1-KO disrupts mitophagy + ISR hyperactivation) — archive pending (OA hybrid)
  • doi:10.3389/fneur.2020.00587 (Grossmann 2020 — RHOT1/MIRO1 in Parkinson’s disease; fibroblast degradation-delay biomarker) — archive pending (OA gold)
• implicit stubs created: miro2, trak1, trak2, kinesin-1, dynein
• existing pages cross-linked: parkin, pink1, pink1-parkin-pathway, mitophagy, lc3, p62, mitochondrial-dysfunction, disabled-macroautophagy, parkinsons-disease, neurodegeneration
• gaps surfaced: needs-human-replication (trafficking arrest; MIRO1 loss phenotype in neurons); needs-replication (MIRO1 degradation delay in PD fibroblasts); unsourced (MIRO1 protein levels in aged human brain; ALS MIRO1 elevation); no-mechanism (ER-mito contact site role; MIRO-TRAK-kinesin ternary structure); needs-canonical-id (GenAge entry not confirmed); long-term-unknown (partial Miro1 reduction over lifespan); no-fulltext-access (Fransson 2006 — not_oa)
• schema gaps: none; sens-categories: [] consistent with other mitophagy machinery pages
• ROADMAP: miro1 marked [x] drafted 2026-05-04 in R10c section under PINK1/Parkin cluster
• verification priority: HIGH — 8 inbound references; Wang 2011 (Cell, 1,139 cites, OA bronze) is the critical mechanistic paper for PINK1 phosphorylation + Parkin ubiquitination claims; suggest wiki-verifier prioritize Wang 2011 and Macaskill 2009 (both OA, downloadable); Saotome 2008 is OA green (freely available); Fransson 2006 closed-access (cannot verify without local PDF)

mitofusins

• added: molecules/proteins/mitofusins.md
• entity type: protein (paralog family page — covers MFN1 + MFN2 together; both [[mfn1]] and [[mfn2]] resolve here via aliases; follows foxo-transcription-factors.md precedent)
• inbound references resolved: 8 (MFN1) + 8 (MFN2) = 16 total inbound stubs resolved
• canonical IDs pulled:
  • MFN1: UniProt Q8IWA4; NCBI Gene 55669; HGNC 18656; 741 aa; ~84 kDa
  • MFN2: UniProt O95140; NCBI Gene 9927; HGNC 16877; 757 aa; ~86 kDa
  • Mouse orthologs: Mfn1 (MFN1); Mfn2 (MFN2)
  • GenAge: neither MFN1 nor MFN2 listed in GenAge-human as of 2026-05-04
• DOI correction: task brief supplied Sebastián 2012 as doi:10.1038/emboj.2012.94 — DOI lookup title is “Genetic inactivation of Cdk7 leads to cell cycle arrest and induces premature aging” (124 citations, CDK7 paper — BUG-2 DOI-title mismatch). Correct Sebastián 2012 DOI = doi:10.1073/pnas.1108220109 (“Mitofusin 2 (Mfn2) links mitochondrial and endoplasmic reticulum function with insulin signaling and is essential for normal glucose homeostasis”; 643 citations; confirmed via PubMed PMID 22427360 + DOI lookup match; local PDF confirmed at )
• DOI note: Detmer & Chan cited as “Detmer 2008” in task brief — correct year is 2007 (PMID 17928812 confirmed; doi:10.1038/nrm2275; published November 2007 in Nat Rev Mol Cell Biol); corrected on page
• primary-source DOIs cited (7 footnotes, all confirmed in archive):
  • doi:10.1083/jcb.200211046 — Chen 2003, Mfn1/Mfn2 KO embryonic lethal + MEF fusion (J Cell Biol; 2,505 citations; OA bronze; PDF pending download)
  • doi:10.1038/ng1341 — Züchner 2004, MFN2 mutations cause CMT2A (Nat Genet; 1,537 citations; local PDF (stale local path))
  • doi:10.1038/nature07534 — de Brito & Scorrano 2008, MFN2 ER-mito tethering (Nature; 2,331 citations; local PDF )
  • doi:10.1093/hmg/ddq419 — Gegg 2010, MFN1/2 PINK1/Parkin-dependent ubiquitination (Hum Mol Genet; 889 citations; OA bronze; PDF pending download)
  • doi:10.1083/jcb.201007013 — Tanaka 2010, p97/proteasome degrades Parkin-ubiquitinated mitofusins (J Cell Biol; 1,323 citations; local PDF )
  • doi:10.1038/nrm2275 — Detmer & Chan 2007, mitochondrial dynamics review (Nat Rev Mol Cell Biol; 1,336 citations; not_oa; no local PDF)
  • doi:10.1073/pnas.1108220109 — Sebastián 2012, Mfn2 ER function + insulin signaling (PNAS; 643 citations; local PDF )
• implicit stubs created (new wikilinks to non-existent pages):
  • mitochondrial-dynamics — process page for mitochondrial network morphology; not yet seeded
  • drp1 — fission GTPase; planned R10d
  • opa1 — inner-membrane fusion GTPase; planned R10d
• existing pages correctly cross-linked: parkin (verified-partial), pink1 (verified), pink1-parkin-pathway (verified-partial), mitophagy (verified-partial), mitochondrial-dysfunction (verified), disabled-macroautophagy (verified), optn (drafted), ndp52 (drafted), miro1 (drafted this round)
• gaps surfaced:
  • needs-human-replication — MFN1/MFN2 decline in human aging (beyond T2D muscle); causal role not demonstrated
  • needs-replication — MFN1/MFN2 relative in vivo stoichiometric contributions; separation-of-function for MFN2 ER-tethering vs OMM fusion
  • no-mechanism — OMM bilayer merger conformational mechanism; no pre/post-fusion membrane-embedded structure
  • no-mechanism — age-associated MFN transcriptional downregulation mechanism
  • long-term-unknown — restoring MFN2 in aged muscle via NAD+/gene-therapy improves lifespan
• schema gaps:
  • PARALOG FAMILY PATTERN: Used protein-complex frontmatter (complex-subunits field) to encode a paralog family rather than a heterotrimer; MFN2 (aging/disease-relevant) set as primary uniprot: field; MFN1 listed in complex-subunits. This is a new application of an existing schema feature. Recommend documenting paralog-family-page convention in CLAUDE.md if additional families (e.g., MCL-1/BCL-xL) warrant combined pages.
• ROADMAP: mfn1 and mfn2 both marked [x] drafted 2026-05-04; resolved to molecules/proteins/mitofusins.md
• verification priority: HIGH — 16 combined inbound references; 4 local PDFs immediately actionable:
  • HIGHEST: Tanaka 2010 (local PDF) — verify K48 ubiquitin chain type and degradation-resistant MFN2 mutant blocking mitophagy claim
  • HIGH: de Brito & Scorrano 2008 (local PDF) — verify ER-targeted MFN2 rescue claim (the key separation-of-function evidence for tethering vs fusion)
  • HIGH: Züchner 2004 (local PDF) — verify CMT2A mutation domains and patient count
  • MEDIUM: Sebastián 2012 (local PDF) — verify MFN2 cKO muscle insulin resistance phenotype and T2D biopsy reduction
  • PENDING DOWNLOAD: Chen 2003 (OA bronze), Gegg 2010 (OA bronze) — trigger downloads for complete verification
  • NOT VERIFIABLE: Detmer & Chan 2007 (not_oa) — review claims must be checked against cited primary sources if challenged

[2026-05-05] verify | glut4

• verified: molecules/proteins/glut4.md
• source verified: Katz 1995 (10.1038/377151a0) — full PDF read; 5 pages
• sources unverifiable:
  • Sano 2003 (10.1074/jbc.C300063200) — hybrid OA, download failed (no candidate URLs); tagged no-fulltext-access
  • James 1989 (10.1038/338083a0) — not_oa; tagged no-fulltext-access
  • Klip 2014 (10.1152/ajpcell.00069.2014) — not_oa; tagged no-fulltext-access
• corrections from Katz 1995 PDF:
  • KO phenotype table: added quantitative heart:body ratio (2.3–2.5× age-matched controls); replaced “adipose dysmorphology” with “severely diminished adipose deposits (females: no dissectable ovarian fat pad)”; added sexually dimorphic glucose phenotype detail (females normal, males moderately abnormal — overt diabetes absent in both sexes); added fed hyperinsulinemia (~5–6× controls); added decreased longevity finding (absent from draft)
  • Compensatory mechanism text: removed incorrect attribution of GLUT1 compensation to skeletal muscle — Katz 1995 explicitly states GLUT1 and GLUT2 were NOT upregulated in skeletal muscle; GLUT1 upregulation (1.5-fold) was in heart only; GLUT2 upregulation (1.7-fold) was in liver only; hyperinsulinemia also identified as compensatory driver
  • Sano 2003 footnote: status updated from “pending (download queued)” to “failed (hybrid OA, no candidate URLs)”
• downstream pages to check: insr, akt, as160, type-2-diabetes — all cite GLUT4 translocation mechanism; compensatory mechanism correction is GLUT4-specific and unlikely to affect those pages directly

[2026-05-04] verify | age-1 (AGE-1 PI3K)

• verified: molecules/proteins/age-1.md
• sources verified:
  • doi:10.1093/genetics/118.1.75 (Friedman & Johnson 1988) — local PDF, full read (12 pages)
  • doi:10.1016/s1534-5807(04)00095-4 (Hertweck 2004) — local PDF, full read (12 pages)
• sources unverifiable:
  • doi:10.1038/382536a0 (Morris 1996) — not_oa; no-fulltext-access retained
• corrections made:
  1. DAF-16 dependence mis-citation removed — wiki cited [^friedman1988] for “daf-16 LoF suppresses age-1 longevity”; Friedman & Johnson 1988 predates DAF-16 identification by ~9 years and contains no DAF-16 mention. Claim now tagged needs-replication with suggested citations (Kenyon 1993, Ogg 1997, Paradis & Ruvkun 1998).
  2. Lifespan table 20°C row label: “Single mutant” → “fer-15 double mutant”; note added that single-mutant data at 20°C are not in this paper (requires Morris 1996, not_oa).
  3. “nine years later” → “eight years later” for Morris 1996 cloning timeline (Friedman 1988 published Jan 1988; Morris 1996 published Aug 1996 = ~8.5 years).
  4. Historical section expanded with exact strain genotype (MK546: age-1(hx546) fer-15(b26ts) II; unc-31(z1) IV), exact lifespan values (25.3 d vs 15.0 d mean; 46.2 d vs 22.0 d max at 25°C vs N2), 20°C figures (~40% mean / ~60% max), and fertility reduction magnitude (~75–80%, to 10–20% of wild-type).
  5. 25°C table row updated: comparator clarified as “vs wild-type N2” (not “fer-15 controls”); full strain genotype added; absolute lifespan values added.
  6. fer-15 background caveat expanded: allele correctly specified as fer-15(b26ts); explained tight linkage to age-1 on LG II (co-isolation, not deliberate crossing); added unc-31 co-allele context.
  7. Footnote [^friedman1988] expanded with cohort n range (6–49 per cohort), self-fertility reduction quantification, and explicit note that DAF-16 was not identified in this paper.
  8. ⚠️ auto-extraction banner removed; verified flag flipped true.
• Hertweck 2004 claims verified correct: sgk-1 LoF 63% (14.7→24.0 d), n=147, p<0.0001; akt-1 p=0.1642 NS; akt-2 p=0.3717 NS — no corrections needed.
• canonical IDs (UniProt Q94125, NCBI 174762) accepted per seeder confirmation; not independently re-queried against databases.

[2026-05-04] verify | mitofusins

Pages verified: 1 (partial scope — Züchner 2004 PDF inaccessible)

• molecules/proteins/mitofusins.md — verified: true (partial scope); corrections: 10

Sources checked:

• doi:10.1083/jcb.200211046 (Chen 2003) — downloaded (OA bronze, PMC); verified; Mfn1/Mfn2 KO lethality staging and fusion assay quantification confirmed
• doi:10.1038/ng1341 (Züchner 2004) — bronze OA; DOI lookup (stale local path) does not exist on this machine (stale path); qualitative claims cross-checked via PMID 15064763 abstract + citations in de Brito 2008; n= claim unverified; tagged no-fulltext-access on footnote
• doi:10.1038/nature07534 (de Brito 2008) — local PDF, verified; ER-mito tethering mechanism, Ca²⁺ transfer quantification, separation-of-function constructs confirmed
• doi:10.1093/hmg/ddq419 (Gegg 2010) — downloaded (OA bronze); verified; SH-SY5Y cells only (not HeLa or cortical neurons); PINK1 siRNA knockdown, not patient mutations
• doi:10.1083/jcb.201007013 (Tanaka 2010) — local PDF, verified; SH-SY5Y+HeLa+MEFs (not COS-7); p97 dominant-negative blocks mitophagy quantified; chain topology not specified in paper
• doi:10.1073/pnas.1108220109 (Sebastián 2012) — local PDF, verified; mef2C-Cre for skeletal muscle (not MLC1-Cre); liver cKO uses Alb-Cre; ER stress pathway confirmed
• doi:10.1038/nrm2275 (Detmer 2007) — not_oa; unverified; claims qualified as review-sourced

Corrections made:

1. Mfn1 KO lethality: “E8.5–E11.5” → “~E11.5–E12.5” (Chen 2003: normal to E10.5, 20% resorbed E11.5, 86% resorbed E12.5; E8.5 not stated for Mfn1)
2. Mfn2 KO lethality wording: updated to “~E10.5–E11.5; 29% resorbing at E10.5, 87% by E11.5” (Chen 2003 exact figures)
3. Sequence similarity: “~80% at the protein level” → “81% at the protein level” (Chen 2003 exact figure for mouse Mfn1/Mfn2; human paralogs noted as similarly close)
4. Sebastián 2012 Cre driver: “MLC1-Cre” → “mef2C-Cre (MEF2C promoter)” — in body table and footnote
5. Gegg 2010 footnote cell lines: “HeLa, SH-SY5Y, mouse cortical neurons” → “SH-SY5Y human dopaminergic neuroblastoma” only; HeLa and cortical neurons not used in this paper
6. Gegg 2010 body text: removed specific PINK1 mutants (W437X, Q456X, L347P) that do not appear in Gegg 2010 (paper used siRNA, not mutant proteins); added needs-replication for mutant characterization
7. K48-chain topology: removed “K48-linked chains on MFN1/MFN2 confirmed” from Tanaka 2010 attribution (paper does not directly characterize chain linkage type); replaced with note that UPS-dependent degradation is consistent with K48 but not directly typed; needs-replication added
8. “Fusion-dead MFN2 that resists ubiquitin-mediated degradation blocks mitophagy” (Tanaka 2010) → corrected to actual finding: p97^QQ dominant-negative blocks MFN degradation and mitochondrial elimination (~80% → <10% of cells lose all mitochondria, n≥50)
9. ER-mito tethering section: “phospholipid synthesis relay” tagged unsourced (not demonstrated in de Brito 2008); Ca²⁺ statistics added (peak [Ca²⁺]m reduced ~40%, n=5, P<0.001); ER-mito distance claim updated to Manders’ colocalization (~40% reduction) + electron tomography per de Brito 2008 methods
10. Züchner 2004 footnote: DOI lookup corrected to stale-path warning; “n=337 CMT2A patients + controls” and “6 novel mutations in 8 families” flagged as unverified (#gap/no-fulltext-access)

Unverifiable claims:

• Züchner 2004: n=, exact mutation count, exact number of families — PDF inaccessible (stale DOI lookup); re-download needed (bronze OA, should be retrievable)
• Detmer 2007 (not_oa): cardiac cKO cardiomyopathy phenotype, domain residue numbers, all review-sourced claims
• UniProt identity fields (Q8IWA4, O95140), HGNC/NCBI identifiers — not re-queried against live databases

Action required:

• Re-download Züchner 2004 (10.1038/ng1341) — stale DOI lookup; paper is bronze OA and was previously downloaded; path prefix (stale local path) needs to be corrected in DOI lookup
• File archive bug: returned (stale local path) for this DOI but that mount does not exist; should return the correct local path

Downstream pages to check (main agent):

• pink1-parkin-pathway — may inherit “K48-linked chains on MFN1/MFN2” from Tanaka 2010; verify attribution
• parkin — MFN1/MFN2 substrate description; chain topology claim; Sebastián 2012 Cre-driver attribution if cited
• mitophagy — MFN degradation checkpoint description; K48/K63 chain topology distinction if present
• downstream pages to check: sgk1 (Hertweck 2004 SGK-1 +63% claim — seeder confirms already verified), caenorhabditis-elegans (Apfeld 2004 AAK-2 parallel claim — seeder confirms already verified), daf-16 (DAF-16 dependence claim with mis-citation — needs corrected footnote if it cites friedman1988 for this), insulin-igf1 (age-1 pathway position claims)

[2026-05-04] ingest | alpha-synuclein (Round 10d)

• added: molecules/proteins/alpha-synuclein.md
• entity type: protein (SNCA; UniProt P37840; NCBI Gene 6622; HGNC 11138)
• inbound links satisfied: parkinsons-disease (verified-partial), neurodegeneration (verified-partial), chaperone-mediated-autophagy (verified-partial), loss-of-proteostasis (verified hallmark) — all 6 inbound references per task brief resolved
• DOIs confirmed in archive (all 100th citation percentile; none locally downloaded):
  • 10.1126/science.276.5321.2045 — Polymeropoulos 1997 — not_oa — 8,128 citations
  • 10.1038/42166 — Spillantini 1997 — bronze OA, download failed — 8,162 citations
  • 10.1126/science.1090278 — Singleton 2003 — not_oa — 4,334 citations
  • 10.1126/science.1101738 — Cuervo 2004 — not_oa — 1,966 citations
• gaps surfaced: native tetramer vs monomer controversy (#gap/contradictory-evidence); toxic species identity (#gap/contradictory-evidence); Braak staging exceptions (#gap/contradictory-evidence); human CMA causal chain (#gap/needs-human-replication); propagation mechanism (#gap/no-mechanism); immunotherapy failure mechanism (#gap/no-mechanism)
• implicit stubs created: braak-staging, ubiquitin-proteasome-system (if not already seeded), lrrk2 (noted as stub in parkinsons-disease.md cross-refs)
• all four primary PDFs closed-access; wiki-verifier cannot cross-check against local PDFs; page remains verified: false indefinitely unless OA copies become available

[2026-05-04] verify | round-10d

mitochondrial-biogenesis

Sources checked:

• doi:10.1016/s0092-8674(00)80611-x (Wu 1999, Cell) — local PDF verified end-to-end
• doi:10.1016/s0092-8674(00)81410-5 (Puigserver 1998, Cell) — local PDF verified end-to-end
• doi:10.1038/nature07813 (Canto 2009, Nature) — local PDF verified end-to-end
• doi:10.1016/j.cell.2013.05.039 (Lopez-Otin 2013) — background review citation confirmed

Corrections made:

1. “raises mitochondrial NAD+” → “raises the cellular NAD+/NADH ratio — driven by increased mitochondrial beta-oxidation” — Canto 2009 measures whole-cell NAD+ and NADH; compartment-specific mitochondrial NAD+ not separately quantified in the paper
2. Beta-adrenergic section rewritten: Puigserver 1998 shows isoproterenol/cAMP induction of PGC-1 mRNA phenomenologically and identifies three PKA consensus phosphorylation sites on PGC-1 protein; the PKA→CREB→PGC-1alpha transcriptional mechanism is a well-accepted inference, not directly demonstrated in Puigserver 1998 — now noted as such
3. [^canto2009] footnote updated: added animal model detail (C57BL/6J, 8-wk male, n=8-10/group); corrected “raising NAD+” to “raising cellular NAD+/NADH ratio (via increased beta-oxidation)”

Downstream pages to check (main agent):

• ampk — may describe “raises mitochondrial NAD+” with same imprecision; apply same correction if present
• sirt1 — Lys778 deacetylation site attribution; confirm it cites Rodgers 2005, not Canto 2009
• caloric-restriction — AMPK/SIRT1 biogenesis claims sourced to Canto 2009; apply NAD+/NADH ratio precision if needed

[2026-05-04] verify | round-10d

opa1

• page: molecules/proteins/opa1.md
• sources verified: Alexander 2000 (10.1038/79944, local PDF); Delettre 2000 (10.1038/79936, local PDF); Frezza 2006 (10.1016/j.cell.2006.06.025, local PDF); UniProt O60313 via REST API
• source unverifiable: Anand 2014 (10.1083/jcb.201308006, not_oa) — claims retained with existing no-fulltext-access markers
• corrections made (3):
  • ADOA prevalence “~1 in 35,000” → “1:12,000 to 1:50,000 (Alexander 2000); ~1 in 50,000 (Delettre 2000)” — neither paper states the wiki’s midpoint figure; Alexander 2000 gives the range explicitly; Delettre 2000 cites 1 in 50,000
  • Penetrance “~88%” → removed unsourced figure; replaced with “variable and incomplete penetrance (Delettre 2000 lod-score calculations use 75%)” — 88% appears in neither primary source; Delettre 2000 Table 1 header specifies penetrance = 75%
  • Cytochrome c cristae fraction “~85%” → refined to “~80–85% stored inside cristae” with source qualifier; Frezza 2006 actually states ~15–20% is IMS-free (not a stated 85% figure); the ~85% was a derived inference; IMS interactor description updated consistently
• confirmed correct: UniProt O60313 (OPA1_HUMAN, 960 aa, 111,631 Da); 8 splice isoforms (both Alexander and Delettre 2000 confirm); fusion independence from Mfn1 (Frezza 2006 Mfn1−/− MEF experiments confirmed); OPA1 knockdown augments and overexpression retards cytochrome c release (Frezza 2006 confirmed); transmembrane anchor topology (UniProt residues 97–113 confirmed, consistent with wiki’s ~89–120 approximate range)
• banner removed; verified: true

oxphos

• page: processes/oxphos.md
• sources verified: Trifunovic 2004 (10.1038/nature02517, local PDF); Singh 2022 ATLAS (10.1016/j.xcrm.2022.100633, local PDF); Murphy 2009 (10.1042/BJ20081386, downloaded PDF)
• source unverifiable: Mitchell 1961 (10.1038/191144a0, not_oa) — chemiosmotic mechanism claims are foundational consensus knowledge; retained without change
• corrections made (4):
  • Complex III mass “~500 kDa” → “~240 kDa” (Murphy 2009 p.6: “The monomer is ~240 kDa and comprises 11 polypeptides”)
  • ROS leak “~0.1–2% of electrons” qualified: added explicit caveat that this figure applies only to isolated mitochondria under mode 2 (high Δp, no ATP synthesis) and cannot be extrapolated to in vivo per Murphy 2009 — in vivo O₂•⁻ is “far, far lower”
  • PolG mouse median lifespan “~13–15 months” → “~48 weeks (~11 months)”; all homozygotes died before 61 weeks — Trifunovic 2004 Fig.2 caption and p.418 text
  • Urolithin A hamstring dose: was “improved at 1000 mg/day”; corrected to “significant at both 500 mg (p=0.027) and 1000 mg (p=0.029) vs placebo”; VO₂ peak and 6MWT significant only at 1000 mg/day (Singh 2022 Table 2)
• added to PolG phenotype: osteoporosis and reduced fertility (confirmed in Trifunovic 2004; previously missing from wiki text)
• footnote updated: Trifunovic 2004 survival curve n’s added (n=50 wt, n=38 mut/mut per Fig.2e caption)
• confirmed correct: CI ~45 subunits ~1 MDa; CIII 11 subunits (Murphy 2009); Singh 2022 n=88/79; 3-arm, 4-month, 40–64 yr; peak power ns; PolG D257A; C57BL/6 × 129Sv background; ROS elevation modest per Trifunovic 2004
• banner removed; verified: true (partial scope: Mitchell 1961 not_oa — full text cannot be verified)

[2026-05-04] verify | round-10d

alpha-synuclein

• page: molecules/proteins/alpha-synuclein.md
• sources: all four primary PDFs closed-access (not_oa or download failed):
  • 10.1126/science.276.5321.2045 — Polymeropoulos 1997 — not_oa
  • 10.1038/42166 — Spillantini 1997 — bronze OA, download failed (0 candidate URLs after filtering)
  • 10.1126/science.1090278 — Singleton 2003 — not_oa
  • 10.1126/science.1101738 — Cuervo 2004 — not_oa
• verification method: PubMed abstracts + MeSH efetch XML (PMID 15333840 for Cuervo 2004), Crossref bibliographic records for all four DOIs, UniProt REST API (P37840 JSON), HGNC REST API (confirmed HGNC:11138 = SNCA)
• corrections made (2):
  1. MEF2D attribution: body text and limitations section attributed MEF2D (neuronal survival TF blocked by A53T/A30P CMA blockade) to Cuervo 2004 [^cuervo2004] → corrected: MEF2D is from Yang et al. 2009 (Science, PMID 19119233, doi:10.1126/science.1166088); MEF2D is absent from Cuervo 2004 MeSH terms and abstract; new [^yang2009] footnote added; body text now correctly attributes general substrate blockade to Cuervo 2004 and MEF2D specifically to Yang 2009; Yang 2009 also establishes that WT (not only mutant) α-synuclein disrupts MEF2D–HSC70 binding — this nuance added
  2. Cuervo 2004 footnote cell model: “SH-SY5Y cells” → “PC12 cells” — confirmed by MeSH D016716 (PC12 Cells) present in Cuervo 2004 PubMed XML; SH-SY5Y is not a MeSH term for this paper; Wistar rats (MeSH D017208) also confirmed; mouse (MeSH D051379) also present; “isolated rat liver lysosomes” retained as consistent with Wistar rat + lysosomal membrane protein focus
  3. Cuervo 2004 footnote language: “dominant-negative receptor occupancy” → “uptake-blocker” — Cuervo 2004 abstract uses the term “uptake blockers”; “dominant-negative” is a reasonable mechanistic inference but is not the authors’ own language; updated in body text and extrapolation table accordingly
• confirmed correct (from PubMed abstract + Crossref): all four DOIs resolve correctly; journal/volume/page correct (Cuervo 2004: Science 305(5688):1292–1295; Singleton 2003: Science 302(5646):841; Spillantini 1997: Nature 388(6645):839–840; Polymeropoulos 1997: Science 276(5321):2045–2047); HGNC 11138 correct (not 11140, which is SNCB); UniProt P37840 confirmed 140 aa; Ser-129 and Tyr-125 phosphorylation confirmed via UniProt REST; PC12 cell model confirmed for Cuervo 2004
• unverifiable without PDF: Singleton 2003 onset age (~35 yr) and kindred clinical details; Spillantini 1997 Lewy neurite / DLB claims; Cuervo 2004 in-text kinetics and KFERQ motif sequence; any quantitative rates or fold-changes in any paper; NAC residue boundaries (61–95) from primary structural papers not assessed; Braak staging exception percentage (~6–8%) not source-traced
• banner removed; verified: true (partial scope — see verified-scope field in frontmatter)
• downstream pages to check (main agent): chaperone-mediated-autophagy (may repeat “dominant-negative” language and MEF2D attribution to Cuervo 2004 — apply same correction if present); cuervo-2004-alpha-syn-impairs-cma (study page frontmatter should list PC12 cells as model, not SH-SY5Y; update if seeded); parkinsons-disease (may cite MEF2D via Cuervo 2004 — check and correct); hsc70 (stub; may attribute MEF2D via Cuervo 2004 when seeded — note Yang 2009 as correct source)

jak-stat-pathway

• page: pathways/jak-stat-pathway.md
• sources verified against local PDF: Xu 2015 (10.1073/pnas.1515386112); James 2005 (10.1038/nature03546); Levy 2002 (10.1038/nrm909)
• sources unverifiable against PDF: Darnell 1994 (10.1126/science.8197455, closed-access; banner had already flagged this); Ridker 2017 (10.1056/NEJMoa1707914, PDF present but CANTOS claims focus-out-of-scope for this pass)
• corrections made (6):
  1. JAK inhibitor name: “JAK1 inhibitor 1” → “JAK inhibitor 1 (pan-JAK)” — Xu 2015 explicitly states it inhibits all four JAKs; the old name falsely implied JAK1 selectivity
  2. Physical function endpoints: expanded from “grip strength, walking speed” to full set confirmed in Fig. 8: grip strength, hanging endurance, ambulation, rearing, RotaRod coordination
  3. Mouse cohort n split: n=8 (CLAMS activity endpoints, Fig. 8A–C); n=9 (grip, hanging, RotaRod, Fig. 8D–F); ruxolitinib dose confirmed 60 mg/kg/day in food
  4. Mechanism qualifier: removed unsourced “NF-κB–dependent SASP transcription” attribution to Xu 2015 → corrected to “phospho-STAT3 (Tyr705) and downstream SASP transcription”; Xu 2015 demonstrates STAT3 Tyr705 reduction (Figs. 4F, 7C/D); NF-κB framing is broader SASP literature inference, not shown in this paper
  5. James 2005 V617F frequencies corrected (three wrong numbers):
     • PV: “>95%” → “40/45 (89%; abstract: ‘>80%’)” per James 2005 actual data
     • ET: “~50–60%” → “9/21 (~43%)”
     • PMF (idiopathic myelofibrosis): “~50–60%” → “3/7 (~43%)”
  6. James 2005 footnote: n=128 (fabricated) → actual cohorts (45 PV + 15 controls + 35 secondary erythrocytosis + 21 ET + 7 MF); p-value corrected to mouse transplant haematocrit experiment (p=0.003 control vs V617F, p=0.0002 WT vs V617F), not a clinical-cohort p-value
• confirmed correct: KEGG hsa04630; Reactome R-HSA-6785807; JAK domain structure (JH1 kinase, JH2 pseudokinase); JAK2 Y1007/1008 sites (confirmed Xu 2015 Fig. S2B and Levy 2002); STAT3 Tyr705, STAT1 Tyr701; SOCS3 on IL-6/STAT3; PIAS sumoylation; three JAKi tested in Xu 2015 (JAK inhibitor 1, momelotinib/CYT387, ruxolitinib/INCB18424); mouse age 24 months; treatment duration 10 weeks; young controls 6 months; V617F in JH2 pseudokinase domain; 0/15 controls carried V617F
• banner removed; verified: true (partial scope — Darnell 1994 closed-access; WikiPathways ID null retained with existing needs-canonical-id)
• downstream pages to check (main agent): senomorphics (already verified 2026-05-04; check for “JAK1 inhibitor 1” language — correct if present); cellular-senescence (may attribute NF-κB SASP mechanism to Xu 2015 — correct if so); xu-2015-jak-inhibition-sasp-frailty if seeded (check n values and drug name)

[2026-05-04] verify | round-10d

hsc70

• page: molecules/proteins/hsc70.md
• sources verified against local PDF:
  • 10.1083/jcb.99.2.723 (Schlossman 1984) — downloaded and read in full (10 pp); PDF verified 2026-05-04
  • 10.1007/s00018-004-4464-6 (Mayer & Bukau 2005) — downloaded and read in full (15 pp); PDF verified 2026-05-04
• sources trusted from chaperone-mediated-autophagy (not_oa, undownloadable):
  • 10.1126/science.2799391 (Chiang 1989) — not_oa
  • 10.1126/science.273.5274.501 (Cuervo 1996) — not_oa
  • 10.1126/science.1101738 (Cuervo 2004) — not_oa
• corrections made (3):
  1. Schlossman 1984 footnote model descriptor: “model: bovine brain clathrin-coated vesicles” → “model: calf brain cytosol (enzyme source); activity assayed using calf brain empty clathrin cages and coated vesicles” — the paper purifies the enzyme from calf brain cytosol (38-fold, >95% pure by SDS-PAGE densitometry); vesicles/cages are substrates, not the model system; added purification fold (38-fold), purity (>95%), and end-product identity (clathrin triskelions) to footnote; corrected “download pending” to “PDF verified 2026-05-04”
  2. Mayer & Bukau 2005 footnote: clarified that NEFs covered are Bag-1 through Bag-6 and HspBP1; HSPH1/HSP105 as a cytosolic NEF for Hsc70 is NOT in this 2005 review (that evidence was established post-2005 by Dragovic et al. 2006 and Raviol et al. 2006); updated “download pending (bronze OA)” to “PDF verified 2026-05-04”
  3. Chaperone cycle step 4 (HSPH1 NEF claim): added inline note that HSPH1/HSP105 NEF activity postdates Mayer & Bukau 2005; tagged unsourced with DOI hint (doi:10.1038/sj.emboj.7601049) for next pass
• confirmed correct against PDFs:
  • Schlossman 1984: 70,000-mol-wt uncoating ATPase; ATP-dependent; isolated as monomers+dimers both active; end product = clathrin triskelions; calf brain cytosol source — all match wiki body text
  • Mayer & Bukau 2005: two-state allosteric cycle (ATP-bound open/low-affinity, ADP-bound closed/high-affinity); J-protein stimulation of ATPase; BAG-family NEF nucleotide exchange; CHIP interaction via C-terminus of Hsc70 and Hsp90; HOP (STIP1) TPR interaction — all match wiki; ATPase domain ~45 kDa + SBD ~25 kDa consistent with domain architecture table
• unverifiable (not_oa sources): CMA steps 1/4/5 HSC70 roles; LAMP-2A bottleneck claim; α-synuclein CMA blockade mechanism; HSC70 level stability in aged tissues; all PTM sites (UniProt-sourced, not PDF-sourced)
• banner removed; verified: true (partial scope — see verified-scope in frontmatter)
• downstream pages to check (main agent): chaperone-mediated-autophagy (cross-links consistent; no corrections propagate from this page); alpha-synuclein (already verified; hsc70 HSC70 role description is consistent — no MEF2D/Yang 2009 misattribution found on this page)

[2026-05-05] verify | round-10d

ep300

• page: molecules/proteins/ep300.md
• sources verified against local PDF:
  • doi:10.1016/s0092-8674(00)81165-4 (Yao 1998, Cell 93:361-372) — already local; verified in full (12 pp)
  • doi:10.1074/jbc.M807135200 (Lee & Finkel 2009, JBC 284:6322-6328) — downloaded during pass (hybrid OA via PMC); verified in full (7 pp)
  • doi:10.1086/429130 (Roelfsema 2005, AJHG 76:572-580) — downloaded during pass (bronze OA via camoufox); verified in full (9 pp)
• sources trusted without re-verification: doi:10.1038/cdd.2014.215 (Pietrocola 2015) — per instruction, treated as verified from spermidine.md pass
• sources unverifiable (archive pending, not triggered): doi:10.1101/gad.8.8.869 (Eckner 1994) — cloning and domain architecture claims retained but not PDF-verified
• corrections made (6):
  1. Heterozygote phenotype (Yao 1998): “viable but display growth retardation and subtle developmental phenotypes” → background-dependent: in 129×BL6 mixed background heterozygotes recovered at normal Mendelian ratios with no overt phenotype; in 129/Sv inbred background ~55% fewer heterozygotes than expected survived to weaning (some died in utero); surviving heterozygotes lacked overt developmental defects in both backgrounds. Paper does not describe “growth retardation” in surviving heterozygotes.
  2. Compound heterozygote (Yao 1998): no correction to outcome (embryonic lethal); wording updated to be explicit (no viable double-heterozygotes at weaning, Table 2).
  3. RSTS2 EP300 mutation frequency (Roelfsema 2005): “~8% carry EP300 mutations” → paper found 3 EP300 mutations and 36 CBP mutations in 92 patients screened (ratio 1:12 EP300:CBP); “~8%” does not appear in this paper; corrected with gap note that later series define the clinical prevalence figure.
  4. RSTS2 phenotype description (Roelfsema 2005): “beaked nose, broad thumbs/toes” → replaced with features described for the three EP300 patients in this paper: arched eyebrows, long eye lashes, prominent nose with long hanging columella, pouting lower lip, broad thumbs, broad big toe.
  5. ATG acetylation substrate nomenclature (Lee 2009): “ATG5, ATG7, ATG12, LC3” → “ATG5, ATG7, ATG8 (LC3), ATG12” — paper assays Atg5, Atg7, Atg8, Atg12 in Fig 2; LC3-II/I conversion is a functional autophagy readout, not an additional direct acetylation substrate in this paper.
  6. Autophagy mechanism framing (Lee 2009): “blocking autophagosome initiation and elongation” → p300 suppresses autophagic flux; p300 knockdown increases LC3-II/I ratio and reduces p62 under fed and starved conditions; p300 overexpression inhibits starvation-induced autophagy; p300-Atg7 interaction reduced ~40% after 2h starvation (n=3 separate experiments; p<0.05). “Blocking initiation” language not used by Lee & Finkel.
• confirmed correct: UniProt Q09472; canonical isoform 2,414 aa; embryonic lethality window “days 9 and 11.5” (abstract, Yao 1998); n=92 in Roelfsema 2005; Journal/volume/page citations correct for all three verified sources; Pietrocola 2015 competitive inhibition and siRNA screen claims (trusted); C646 described in text correctly; Lee 2009 model is HeLa cells (confirmed).
• footnotes updated: Lee 2009 — PDF path recorded; n=3 experiments noted; model specified as HeLa. Roelfsema 2005 — PDF path recorded; CBP/EP300 mutation counts added. Yao 1998 — strain backgrounds specified.
• banner removed; verified: true (partial scope — Eckner 1994 not PDF-verified; canonical-database identity fields (UniProt Q09472, HGNC 3373, NCBI Gene 2033) not independently re-checked against live databases)
• downstream pages to check (main agent): autophagy (may cite “EP300 acetylates ATG5, ATG7, ATG12, LC3 blocking autophagosome formation” — correct ATG8/LC3 nomenclature and flux-suppression framing if present); spermidine (Pietrocola 2015 EP300 substrate list may use old LC3-separate nomenclature — check and align to ATG8/LC3 if needed)

deptor

• page: molecules/proteins/deptor.md
• sources verified against local PDF:
  • doi:10.1016/j.cell.2009.03.046 (Peterson 2009, Cell 137:873-886) — downloaded and read in full (14 pp); PDF verified 2026-05-04
  • doi:10.1016/j.molcel.2011.08.029 (Zhao, Xiong, Sun 2011, Mol Cell 44:304-316) — downloaded and read in full (13 pp); PDF verified 2026-05-04
  • doi:10.1016/j.molcel.2011.09.005 (Duan, Skaar, Kuchay, Pagano et al. 2011, Mol Cell 44:317-324) — downloaded and read in full (8 pp); PDF verified 2026-05-04
  • doi:10.18632/aging.203959 (Shi, Endicott, Miller 2022, Aging 14:2442-2457) — downloaded and read in full (16 pp); PDF verified 2026-05-04
• sources not independently verified: UniProt Q8TB45 identity fields (accession, HGNC 24784, NCBI Gene 64798) — not re-queried against live database
• corrections made (9):
  1. Frontmatter key-ptms: removed fabricated Ser235-phospho (appears in no primary source); added Ser293-phospho and Ser299-phospho (mTOR priming sites per Duan 2011 and Zhao 2011)
  2. Footnote labels corrected: [^gao2011] and [^zhao2011] were wrong author names — no “Gao” authored either paper; corrected to [^zhao2011_sun] (Zhao/Xiong/Sun, doi:…/08.029) and [^duan2011] (Duan/Pagano lab, doi:…/09.005); all in-body citations updated
  3. PDZ domain residues: “330-407” → “324-409 (approx.)” — Peterson 2009 Fig. 1I marks minimal mTOR-binding region as DEPTOR residues 324-409; “330-407” was unsourced
  4. Domain table PDZ function: “Mediates binding to mTOR FAT/kinase region” → “Mediates binding to mTOR FAT domain (residues 1483-2000 of mTOR)” — paper maps minimal mTOR-binding region to FAT domain only, not kinase domain
  5. Body PDZ description: removed “and kinase domain” from “FAT/kinase region” — same correction; paper specifies FAT only (mTOR aa 1483-2000)
  6. Regulation Step 1 mechanism: corrected false claim that “mTOR itself (direct)” phosphorylates degron serines Ser286/287/291 — mTOR phosphorylates priming sites Ser293/Ser299; CK1alpha then phosphorylates the degron; RSK1/S6K1 can also directly phosphorylate the degron; paragraph restructured to reflect two-stage mechanism (Zhao 2011 + Duan 2011)
  7. MM overexpression quantified and attribution broadened: added “28% of MMs (160/581 cases with mRNA ≥4-fold)” from Peterson 2009 text; added c-MAF/MAFB translocation subset alongside cyclin D — original wiki incorrectly described overexpression as driven only by IGH/cyclin-D translocations
  8. Miller 2022 tissue specificity and functional null result added: DEPTOR reduction is liver-specific (unchanged kidney; increased muscle); DEPTOR/PRAS40 RNAi knockdown in GHRKO fibroblasts did not alter mTORC1/2 substrate phosphorylation — indicating DEPTOR reduction is not functionally responsible for lower mTOR activity; TSC1/TSC2 elevation identified as dominant mechanism
  9. Miller 2022 footnote: “n=small groups” → “n=6 male + 6 female per genotype”; added authors, journal, and functional null result
• confirmed correct against PDFs: 409 aa (Peterson 2009 Fig. 1B); NCBI Gene 64798; DEP domains dispensable for mTOR binding; PDZ sufficient (Fig. 1H); dual mTORC1+mTORC2 inhibition; degron sequence SSGYFS (residues 286-291); all three Ser must be phosphorylated for beta-TrCP recognition; FBXW7 does NOT target DEPTOR; 8226/OCI-MY5 MM apoptosis upon DEPTOR knockdown; “bistable switch” corrected to “positive feedback loop” per Duan 2011 language
• banner removed; verified: true (partial scope — UniProt Q8TB45 identity fields not re-queried against live database)
• downstream pages to check (main agent):
  • mtor — may cite DEPTOR mTOR-binding region as “FAT/kinase domain”; correct to FAT domain only (mTOR residues 1483-2000)
  • raptor / rictor — DEPTOR interaction description may carry old “FAT/kinase” wording
  • Any page describing DEPTOR phosphorylation that says “mTOR directly phosphorylates Ser286/287/291” — incorrect; mTOR phosphorylates priming sites Ser293/Ser299; CK1alpha phosphorylates the degron

[2026-05-04] propagate | round-10d-corrections

R10d verification batch surfaced ~25 downstream propagation candidates. Applied the following high-confidence corrections; flagged conflicts/lower-priority items for next lint pass.

Applied

• hallmarks/mitochondrial-dysfunction.md — PolG mutator median lifespan corrected from “~60 weeks vs ~120 weeks” to “~48 weeks for homozygous mutators vs ~120 weeks; all homozygotes dead before 61 weeks per Trifunovic 2004 Fig. 2” (per oxphos.md verifier)
• molecules/compounds/spermidine.md — EP300 substrate list updated from “ATG5, ATG7, ATG12, and LC3” to “ATG5, ATG7, ATG8 (LC3), and ATG12” in two places; matches Lee & Finkel 2009 nomenclature per ep300.md verifier
• molecules/proteins/hsc70.md — Cuervo 2004 footnote cell model corrected from “human SK-N-SH neuroblastoma cells” to “PC12 cells (rat pheochromocytoma) and isolated Wistar rat liver lysosomes”; “competitively inhibiting” → “uptake-blockers” per Cuervo 2004’s own term (per alpha-synuclein.md verifier)
• processes/autophagy.md — spermidine row in CR-mimetics table updated to reflect Pietrocola 2015 EP300-selective refinement of the original Eisenberg 2009 “broad HAT” framing; verified-scope frontmatter updated
• molecules/proteins/cytochrome-c.md — cristae cytochrome c fraction “85%” → “~80–85%” with source qualifier (“Frezza 2006 states ~15–20% is IMS-free; the cristae-sequestered fraction is a derived inference”) per opa1.md verifier
• molecules/compounds/nmn.md — NAD+ decline figure “~30–50%” → “~30–90% tissue-dependent (per Yoshino 2018 review)” per nampt.md verifier

Flagged but not propagated (verifier conflict or low-priority)

• phenotypes/alzheimers-disease.md TREM2 R47H OR=4.5: alzheimers-disease.md verifier originally cited Guerreiro 2013 Table 2 discovery cohort OR=4.5 (CI 1.7–11.9); trem2.md verifier read end-to-end and found this figure does not appear in the paper, only the combined-series OR=5.05. Conflict between verifier passes — flagged for next lint pass to re-read PDF.
• studies/guerreiro-2013-trem2-ad, studies/jonsson-2013-trem2-ad, studies/hammond-2019-microglia-aging, studies/cuervo-2004-alpha-syn-impairs-cma, studies/xu-2015-jak-inhibition-sasp-frailty, studies/hutton-1998-mapt-ftdp17 — none yet seeded (R10f deferred); will inherit corrected wording when seeded
• mtor.md / raptor.md / rictor.md DEPTOR FAT-only binding (vs FAT/kinase): grepped these pages — none contain “FAT/kinase” wording; the corrected description lives only on deptor.md itself
• chronic-inflammation.md / tbk1.md “cGAS-STING in aged tissues” — checked; both pages have generic mentions that don’t reproduce Yang 2017’s overstated aged-tissue claim. No propagation needed.
• cellular-senescence.md / sasp.md Glück 2017 cell-line attribution — both pages already correctly attribute to MEFs + WI-38 (NOT IMR90). No propagation needed.