🧬 Aging Wiki

R23

34 min read

log/R23.md — Round 23 entries

Sub-file of log — see parent for index.

[2026-05-06] R23 — Cell-therapy interventions + sparse-sibling intervention-folder buildout

Round complete. 22 new pages across R23a–e. Goal: populate the empty interventions/stem-cell-therapy/ folder + bring sparse sibling intervention folders to working depth + add a single modality-axis synthesis MOC.

Pages added:

  • R23a (7): interventions/stem-cell-therapy/{hematopoietic-stem-cell-transplantation, mesenchymal-stem-cell-therapy, ipsc-derived-cell-therapy, in-vivo-partial-reprogramming-therapy}.md + cell-types/{mesenchymal-stem-cells, neural-stem-cells, induced-pluripotent-stem-cells}.md
  • R23b (4): interventions/gene-therapy/{aav-klotho, aav-follistatin, crispr-base-editing-pcsk9, aav-osk}.md
  • R23c (6): interventions/lifestyle/{intermittent-fasting, time-restricted-eating, methionine-restriction, ketogenic-diet, sleep, heat-exposure}.md
  • R23d (4): interventions/pharmacological/{mtor-inhibitors, nad-precursors, ampk-activators, sirtuin-activators}.md
  • R23e (1): frameworks/interventions-by-modality.md (Design 2 — single Dataview-driven framework page)

Pages updated (propagation):

  • hallmarks/stem-cell-exhaustion.md — “Therapeutic strategies” section enriched with all four R23a interventions + cross-link to R23e MOC
  • cell-types/hematopoietic-stem-cells.md — HSCT-as-intervention up-link
  • cell-types/satellite-cells.md — gap note: no satellite-cell-therapy product clinically used for aging
  • processes/partial-reprogramming.md — cross-link to R23a + R23b (in-vivo-partial-reprogramming-therapy + aav-osk)
  • molecules/compounds/{rapamycin, metformin, nmn, nr}.md — up-links to their R23d class pages

DOI corrections caught at seed time (8): Yang 2023 Cell (.2023.02.013 →.2022.12.027); Doi 2024 → Sawamoto 2025 Nature; Newman 2017 / Roberts 2017 Cell Metab swap (.08.004 vs.08.005); “Mendell 2015” → Mendell 2017 Mol Ther; “LeBlanc 2020” → Olsen 2020 (Norwegian MR RCT); Miller 2005 (.9728 →.9726); Cavuoto & Fenech 2012 DOI lookup mismatch (BUG-2 candidate for a local paper archive); Lonafarnib ChEMBL ID confirmed clean (R21 carryover).

Schema escalations surfaced (8):

  1. type: cell-type schema in CLAUDE.md missing verified / verified-date / verified-by field listing (HSC prototype carries them; schema block doesn’t)
  2. iPSCs need schema-documented exception for empty tissue-of-origin: + key-aging-phenotypes:
  3. No iPSC-specific Cell Ontology ID; CL:0002248 (pluripotent stem cell) used as approximation
  4. safety-profile single-value field can’t capture indication-split profiles (HSCT, MSC therapy)
  5. translation-gap similarly indication-split for some interventions; pipe-delimited free-text used as workaround
  6. clinical-stage for class pages: aging-context vs. maximum-regulatory? Senolytics precedent followed (aging-context)
  7. human-evidence-level: limited-negative proposed for STACs (currently limited + body explanation)
  8. [[myostatin]] protein page promoted from R18-backlog to R24-priority (AAV-Follistatin depends on it)

Gaps surfaced for next lint pass:

  • interventions/lifestyle/caloric-restriction.md may need mode: dietary (currently mode: lifestyle); same audit for intermittent-fasting, time-restricted-eating, etc. — surfaced in R23e watchdog 1 commentary
  • clinical-trials-active: field not populated on all R23 pages — refresh per sops/integrating-clinical-trials.md
  • 5 implicit stubs created across R23 that may want promotion: myostatin (priority), pcsk9, hsf1, heat-shock-response, hormesis
  • Multiple R23 pages cite closed-access primary sources (Mannick 2014/2018 in mtor-inhibitors; Bannister 2014 in ampk-activators; Hannuksela 2001 in heat-exposure; Cohen 2006 in crispr-base-editing-pcsk9) — verifier sweep will need access acquisition before flipping verified: true

Verification status: All 22 pages ship verified: false with ⚠️ banner; await wiki-verifier sweep. DOI sanity at seed time confirmed all cited DOIs resolve in DOI lookup (with the corrections noted above).

[2026-05-06] verify | cell-types/induced-pluripotent-stem-cells.md

Sources checked:

  • doi:10.1016/j.cell.2006.07.024 (Takahashi & Yamanaka 2006, Cell) — PDF verified end-to-end (local archive)
  • doi:10.1101/gad.173922.111 (Lapasset et al. 2011, Genes Dev) — PDF verified end-to-end (local archive)
  • doi:10.1111/acel.12877 (Olova et al. 2019, Aging Cell) — PDF verified end-to-end (local archive)
  • doi:10.1101/cshperspect.a019448 (Hochedlinger & Jaenisch 2015, Cold Spring Harb Perspect Biol) — downloaded on demand; PDF verified end-to-end
  • doi:10.1016/j.cell.2007.11.019 (Takahashi et al. 2007) — closed-access; abstract only; no-fulltext-access retained in footnote
  • doi:10.1126/science.1151526 (Yu et al. 2007) — not_oa; abstract only; no-fulltext-access retained in footnote

Corrections made:

  • SCIENTIFICALLY CRITICAL: Removed “Werner syndrome patients” from Lapasset 2011 attribution — Werner syndrome donors are not in Lapasset 2011. Paper used only replicatively senescent (74S) and centenarian fibroblast donors.
  • Lapasset 2011 donor ages: “centenarian (age ~100)” → 92-, 94-, 96-, 101-yr-old donors + replicatively senescent 74-yr-old (74S) cells.
  • Lapasset 2011 factor set: Added missing detail — six-factor OSKMNL cocktail (OCT4, SOX2, KLF4, c-MYC, NANOG, LIN28); four-factor OSKM was insufficient for senescent cells.
  • Lapasset 2011 efficiency: 0.06% explicitly stated in paper; added to body and footnote.
  • Olova 2019 (SCIENTIFICALLY CRITICAL): Corrected experimental design from “OSKM induction followed by withdrawal at defined time points” to observational analysis of a published 49-day continuous OSKM transduction time-course. Not a factor-withdrawal experiment.
  • Olova 2019 kinetics precision: eAge decline begins days 3–7 (not “from the start”); rate 3.8 yr/day (SE 0.27, p=3.8×10⁻⁷), days 3–20; somatic gene-expression uncoupling mechanism (FSP1/COL3A1/TGFB2/3 cluster stable until day 15) added.
  • c-Myc-free efficiency “~10× lower”: removed unsupported quantification; tagged unsourced.
  • All footnotes updated with corrected detail. Removed ⚠️ auto-extraction banner. Flipped verified: true.

Downstream propagation needed (main agent):

Cross-cutting propagation done by main agent (R26b post-verification pass):

  • hsf1 corrections propagated to heat-shock-response (basal-localization framing, HR-A/B residue range 130-221, intro line “translocates” → “accumulates further”) and hsp70 (basal-localization framing).
  • pcsk9 FOURIER verification status propagated to atherosclerosis (Sabatine 2017 footnote AI-extraction caveat lifted; pdf-pending note replaced with verified-via-pcsk9 note).
  • mitofusins gained Karbowski 2002 MFN2-at-apoptotic-fission-sites section (R25 carryover; quantitatives transcribed verbatim from verified mitochondrial-dynamics).
  • hematopoietic-stem-cells CHIP CIs reconciled with cluster — body-Results values (CHD 1.2-3.5 / stroke 1.3-4.8) preferred over abstract values (CHD 1.2-3.4 / stroke 1.4-4.8); abstract/body discrepancy documented in verified-scope.
  • antagonistic-pleiotropy Tyner 2002 allele-name correction — checked across genomic-instability, puma, cellular-senescence — all already use correct “m-allele”/“p53+/m” terminology, no propagation needed.
  • TET1 Wagner 2015 aging-direction misattribution — checked tet2, tet3, epigenetic-alterations — none had inherited the wrong framing.
  • Caspase-1 Asp116-Ser117 / BHB-K⁺-efflux corrections — checked asc, nlrp3-inflammasome, chronic-inflammation — only the caspase-1 page itself had the wrong framings.
  • Astrocytes Iliff-2012/Xie-2013 sleep correction — sleep already correctly attributes Xie 2013 (no propagation needed).

[2026-05-06] verify | interventions/stem-cell-therapy/mesenchymal-stem-cell-therapy.md

Sources checked:

  • doi:10.1093/gerona/glx056 (Golpanian et al. 2017, J Gerontol A) — PDF verified end-to-end (downloaded on demand; bronze OA)
  • doi:10.1093/gerona/glx137 (Tompkins et al. 2017, J Gerontol A) — PDF verified end-to-end (downloaded on demand; bronze OA)
  • doi:10.1002/sctm.17-0051 (Caplan AI 2017, Stem Cells Transl Med) — PDF verified end-to-end (local archive; gold OA)
  • doi:10.1016/j.stem.2018.05.004 (Galipeau & Sensébé 2018, Cell Stem Cell) — PDF verified end-to-end (local archive; bronze OA)
  • doi:10.1080/14653240600855905 (Dominici et al. 2006) — not_oa; ISCT minimal criteria claims unverifiable against full text; no-fulltext-access
  • doi:10.1002/jor.1100090504 (Caplan AI 1991) — not_oa; original MSC nomenclature claim unverifiable; no-fulltext-access
  • doi:10.1007/s40259-018-0311-4 (Scott LJ 2018) — not_oa; darvadstrocel ADMIRE-CD remission figures unverifiable; no-fulltext-access; EU approval confirmed via Galipeau 2018
  • Remestemcel-L FDA approval (December 2024): confirmed via PubMed web search

Corrections made:

  • SCIENTIFICALLY CRITICAL: Golpanian 2017 mean age corrected from “~76” to “78.4 ± 4.7 years” (the ~76 figure was the 100M subgroup mean, not the cohort mean)
  • Golpanian 2017 TNF-α claim: wiki stated “TNF-α reduced” without dose specificity — corrected to reflect that serum TNF-α decreased significantly only in the 100M (p=.0001) and 200M (p=.0001) groups; the 20M group was not significant (p=.09)
  • Golpanian 2017 6MWT claim: added that the 200M group did not show significant individual improvement in 6MWT (the improvement was significant for the combined group and for the 100M group, not 200M individually)
  • Golpanian 2017: added clinically important detail that one patient in the 200M group died at 258 days post-infusion (adjudicated unrelated to treatment); this was absent from the wiki
  • Golpanian 2017 footnote: “mean age ~76” → “mean age 78.4 ± 4.7 years”; download status corrected from “pending” to “downloaded”; quantitative p-values added for 6MWT and TNF-α
  • Tompkins 2017 (“Phase 2a” label): wiki called the trial “Phase 2a” — the paper does not use this term; it is described as “Phase II” throughout; label removed
  • Tompkins 2017 TNF-α claim: corrected; serum TNF-α decreased only in the 100M group (p=.031); the 200M group did NOT significantly reduce serum TNF-α (p=.129) — the original claim implied TNF-α improved in both treatment groups; distinction now explicit
  • Tompkins 2017: added specific 6MWT numbers (345.9→410.7 m, p=.011 for 100M group); added SPPB, FEV1, and B-cell intracellular TNF-α results; added that the study was explicitly underpowered for between-group efficacy comparisons
  • Tompkins 2017 page range in footnote: “1513-1522” → “1513–1521” (last page confirmed from PDF)
  • Remestemcel-L/Ryoncil: removed unsourced tag; confirmed December 2024 FDA approval; added that Canada/New Zealand conditional approval preceded US (2012, per Galipeau 2018)
  • Darvadstrocel/Alofisel: corrected disease description from “perianal Crohn’s fistulae” to “Crohn’s-related enterocutaneous fistular disease” (per Galipeau 2018 terminology); added March 2018 EMA approval date; added ADMIRE-CD trial number (NCT01541579) and 120M cell dose
  • Introduction line: updated regulatory precedent summary to use correct approval dates (FDA December 2024; EMA March 2018)
  • Caplan 2017 footnote: expanded to include specific content (MSCs derived from pericytes; “rarely or never differentiate into tissue at site”; download status “pending” → “downloaded”)
  • Galipeau 2018 footnote: expanded to include key mechanistic insights about dosing translation gap (50 mg/kg mouse vs 1–2 million cells/kg human) and fitness (fresh vs thawed); download status updated
  • Removed ⚠️ auto-extraction banner; flipped verified: true (partial scope — 4/7 sources verified against full PDF)

Unverifiable claims:

  • ISCT minimal criteria (Dominici 2006, not_oa): marker percentages (CD105+, CD73+, CD90+ >95%; CD45-, CD34- <2%) — internally consistent but unverified
  • Original MSC naming history (Caplan 1991, not_oa): verified indirectly via Caplan 2017, which cites the 1991 paper
  • ADMIRE-CD precise remission percentages (Scott 2018, not_oa): “~50% vs placebo at 24 wk; >50% at 52 wk” — the EU approval and trial success are confirmed via Galipeau 2018; the specific figures are from the not_oa source

Downstream propagation needed (main agent):

  • [[mesenchymal-stem-cells]] (cell-type sibling page, R23a) — if it cites Golpanian 2017 or Tompkins 2017, verify those citations carry the corrected mean ages, TNF-α dose specificity, and 200M null results
  • [[hallmarks/stem-cell-exhaustion]] MOC — may reference MSC frailty trials; verify the claimed frailty trial results match corrected Golpanian/Tompkins data
  • [[phenotypes/frailty]] (if page exists) — if it cites CRATUS trials, propagate corrected ages and efficacy findings
  • [[interventions/pharmacological/senolytics]] or any page citing remestemcel-L — confirm the approval year (December 2024) if referenced

[2026-05-06] verify | cell-types/mesenchymal-stem-cells.md

Sources checked:

  • doi:10.1080/14653240600855905 (Dominici 2006, Cytotherapy) — not_oa; ISCT criteria verified via PubMed PMID 16923606 abstract + Crossref metadata; criteria list confirmed accurate
  • doi:10.1016/j.cell.2007.08.025 (Sacchetti 2007, Cell) — PDF verified end-to-end (downloaded during session; PMC OA)
  • doi:10.1038/nature09262 (Mendez-Ferrer 2010, Nature) — PDF verified end-to-end (local archive; 8 pages read including full Methods)
  • doi:10.1371/journal.pone.0002213 (Wagner 2008, PLOS ONE) — PDF verified end-to-end (downloaded during session; gold OA; 12 pages read including Methods)
  • doi:10.1002/sctm.17-0051 (Caplan 2017, Stem Cells Transl Med) — PDF verified end-to-end (downloaded during session; 7 pages read)
  • doi:10.1016/j.arr.2005.10.001 (Sethe 2006, Ageing Res Rev) — abstract only (download failed twice; OA URL returned HTTP 202); no-fulltext-access
  • doi:10.1016/j.bone.2016.01.014 (Singh 2016, Bone) — Crossref metadata only (download failed; not publicly accessible); no-fulltext-access

Corrections made:

  • SCIENTIFICALLY CRITICAL — Styner2016 footnote: wrong first author and wrong page numbers. “Styner M et al.” → “Singh L et al.”; pages “85:190-199” → “85:29-36”. Title confirmed via Crossref: “Aging alters bone-fat reciprocity by shifting in vivo mesenchymal precursor cell fate towards an adipogenic lineage.” Footnote key renamed [^singh2016]; in-text citation updated.
  • SCIENTIFICALLY CRITICAL — Mendez-Ferrer 2010 HSC reduction claim: “reduces HSC numbers by ~50% (n=4–8 per group; p<0.05 by t-test)” → “reduces the immature CD150+CD48− LSK progenitor pool by ~50% (n=6–12 per group; p<0.01 by unpaired two-tailed t-test).” Paper explicitly states total BM cellularity and Lin−CD48− numbers were NOT affected in the first 2 weeks; only the immature LSK progenitor subset was reduced. Added this nuance to body and footnote.
  • SIGNIFICANT — Wagner 2008 donor n: “n=10 bone marrow donors” → “n=8 bone marrow donors” (Methods section: “we have used specimen from 8 healthy donors”).
  • SIGNIFICANT — Wagner 2008 SASP attribution: body text claimed “The SASP phenotype of senescent cultured MSCs includes IL-6, IL-8, MMP-3, and CXCL1”; Wagner 2008 does NOT characterize the SASP cytokine profile — it focuses on mRNA/miRNA gene expression and differentiation potential. SASP attribution removed from body and clarified as secondary literature. #gap/needs-replication tag added requiring a primary citation.
  • SIGNIFICANT — Wagner 2008 replicative senescence framing: corrected to reflect that osteogenic differentiation INCREASED in later passages while adipogenic DECREASED — the simple “loss of differentiation” framing is inaccurate. Also clarified that continuous changes begin from passage 1, not merely at late passage.
  • Sethe 2006 body text: replaced specific claim about functional-vs-numerical decline (unverifiable from abstract) with source-accurate quote from abstract plus note that full-text verification is needed.
  • All footnote download statuses updated to reflect verified state (local PDF, abstract-only, or failed).
  • Sacchetti 2007 footnote enriched with paper’s key specific finding.
  • Removed ⚠️ auto-extraction banner; flipped verified: true with partial scope.

Unverifiable claims:

  • Sethe 2006 specific conclusions about functional-vs-numerical decline pattern — abstract-only access; full text inaccessible
  • Singh 2016 quantitative details (exact n, p-values, mouse strain) — download failed permanently

Downstream propagation needed (main agent):

  • [[mesenchymal-stem-cell-therapy]] — intervention page (R23a sibling); likely inherits Wagner 2008 n=10 (should be 8) and possibly the SASP components claim; check and correct
  • [[stem-cell-exhaustion]] — hallmark MOC; may reference MSC aging claims including the functional decline framing from Sethe 2006; check and correct
  • [[hematopoietic-stem-cells]] — may cite Mendez-Ferrer 2010 for niche MSC depletion; verify the reduced-HSC-number framing matches the corrected “CD150+CD48− LSK pool” specificity

[2026-05-06] verify | cell-types/neural-stem-cells.md

Sources checked (10 DOIs):

  • doi:10.1016/s0092-8674(00)80783-7 (Doetsch 1999, Cell) — PDF verified end-to-end (local archive; 8 pages)
  • doi:10.1523/JNEUROSCI.4608-03.2004 (Maslov 2004, J Neurosci) — PDF verified end-to-end (downloaded during session; 8 pages)
  • doi:10.1038/nature25975 (Sorrells 2018, Nature) — PDF verified end-to-end (downloaded during session; HHS PMC manuscript, 6 pages)
  • doi:10.1016/j.stem.2018.03.015 (Boldrini 2018, Cell Stem Cell) — PDF verified end-to-end (downloaded during session; 8 pages)
  • doi:10.1038/nature10487 (Sanai 2011, Nature) — PDF verified end-to-end (downloaded during session; 8 pages)
  • doi:10.1016/j.cell.2019.01.040 (Kalamakis 2019, Cell) — PDF verified end-to-end (downloaded during session; 8 pages)
  • doi:10.1038/3305 (Eriksson 1998, Nat Med) — not_oa / download failed; PubMed abstract (PMID 9809557) only; n=5 confirmed
  • doi:10.1126/science.1553558 (Reynolds 1992, Science) — not_oa; not downloaded; claim uncontroversial
  • doi:10.1016/j.cell.2013.05.002 (Spalding 2013, Cell) — download failed; PubMed abstract + Sorrells 2018 cross-citation

Corrections made:

  1. SCIENTIFICALLY CRITICAL — Boldrini 2018: wiki incorrectly stated “Sox2+ NSCs, DCX+ progenitors, and mature neurons did not significantly decline.” Paper shows Sox2+ QNPs DID decline in anterior-mid DG; INPs, DCX+ immature granule neurons, NeuN+ mature neurons, and DG volume were stable. Body and footnote corrected.
  2. SCIENTIFICALLY CRITICAL — Kalamakis 2019 age timepoints: “6 month, 20 month” → “7 month, 22 month.” Corrected in body and footnote.
  3. SCIENTIFICALLY CRITICAL — Kalamakis 2019 extrinsic quiescence driver: wiki attributed to “Notch Delta-like ligand from ependymal cells”; paper identified sFRP5 (Wnt antagonist) and niche-derived inflammatory signals (interferon pathway). Corrected in body and footnote.
  4. SIGNIFICANT — Kalamakis 2019 cell count: “~2000 single cells” → “~3066 cells (10x Chromium: 1696 young + 1370 old) + 225 Smart-seq2 cells.”
  5. SIGNIFICANT — Sanai 2011 sample count: n=59 → n=60 (10 neurosurgical + 50 autopsied). Corridor age corrected (depleted 6–18 months, not “declines to sporadic by ~18 months”). MMS to prefrontal cortex finding added.
  6. SIGNIFICANT — Maslov 2004: strain specified as C57BL/6J male; ages clarified; twofold NSC reduction figure and division-rate-unchanged nuance added. SGZ proliferation section updated with specific Maslov numbers.
  7. MINOR — Sorrells 2018: sample breakdown clarified (37 postmortem + 22 surgical; adult subset n=17+12, ages 18–77).
  8. MINOR — Eriksson 1998 n=5 added to timeline; methodological caveat added.
  9. MINOR — Reynolds, Eriksson, Spalding footnotes annotated with verification status.

Unverifiable claims:

  • Reynolds 1992 (not_oa) — foundational neurosphere claim
  • Eriksson 1998 (not_oa) — n=5 and method confirmed via abstract only
  • Spalding 2013 (download failed) — 700/day and 1.75%/year cross-confirmed via Sorrells 2018

Downstream propagation needed (main agent):

  • stem-cell-exhaustion — may cite Kalamakis 2019 with wrong ages (6/20 months) or wrong quiescence driver attribution
  • neurons (R24c companion stub) — may inherit incorrect Boldrini 2018 summary
  • brain tissue page — may reference Sorrells 2018 or Sanai 2011 with wrong n
  • notch-pathway — any claim attributing SVZ deep quiescence to Notch per Kalamakis 2019 should be corrected to sFRP5 + inflammatory signals

[2026-05-06] verify | studies/li-2023-dual-mtor-nad-gerotherapy.md

Pages verified: 1

  • studies/li-2023-dual-mtor-nad-gerotherapy.md (corrections: 3 factual; 2 clarifications added)

Sources checked:

  • doi:10.1101/2023.01.16.523975 (Li et al. 2023, bioRxiv preprint) — PDF downloaded on demand (gold OA via PMC; 33 pages read end-to-end including abstract, methods, results, discussion, figure legends)

Corrections made:

  • SCIENTIFICALLY CRITICAL — PDB accession for SIRT3 docking structure: 5514 → 4BN4. The paper uses PDB 4BN4 (SIRT3 in complex with ADP-ribose, 1.3 Å resolution) throughout the docking methods and figure legends. PDB 5514 does not correspond to this structure.
  • SCIENTIFICALLY CRITICAL — SOD1-G93A mouse strain: limitations claimed “single genetic background (C57BL/6J).” The HFD/NASH mice are C57BL/6J, but the SOD1-G93A ALS mice are SV129 background (per methods, lines 598–599). Corrected to specify two different backgrounds and their respective uses.
  • FACTUAL — Conflict of interest statement: wiki stated “Jerry Colca (co-author) is listed as inventor on related patents.” The paper’s COI section does not disclose any patent-inventor conflict for Colca. The disclosed COI is T.R.P. as BIOIO founder and T.M.M. as consultant/advisor to multiple pharma including BIOIO. Corrected to match actual disclosure.

Clarifications added (not errors, but precision improvements):

  • NASH/HFD model diet composition made explicit (60% vs 40% kcal trans-fat; HTF-C constituents quantified per methods).
  • SOD1-G93A ALS model: typical lifespan (~18 weeks) added to Design section per paper text (line 212).

Unverifiable claims:

  • ALS mouse n (SOD1-G93A survival curves): not stated in figure legends or methods — limitation note preserved on page.
  • HFD metabolic experiment (Fig S2) per-group n: n=6 is stated for transcriptional profiling subset (Fig 1F legend) but not explicitly for the main HFD metabolic cohort.

Downstream propagation needed (main agent):

  • sirtuin — if it cites this study and references a PDB code for SIRT3, confirm it uses 4BN4 not 5514
  • mtor — NAD+/Sirtuin axis citation from this study; no quantitative corrections needed (mechanism description correct)

[2026-05-06] verify | interventions/pharmacological/sirtuin-activators.md

Pages verified: 1 (full — all 5 cited primary-source PDFs read end-to-end)

Sources checked:

  • doi:10.1038/nature01960 (Howitz 2003, Nature) — PDF verified (5 pages)
  • doi:10.1038/nature02789 (Wood 2004, Nature) — PDF verified (4 pages + corrigendum)
  • doi:10.1038/nature05354 (Baur 2006, Nature) — PDF verified (6 pages)
  • doi:10.1016/j.cmet.2008.06.011 (Pearson 2008, Cell Metab) — PDF downloaded and verified (8 pages; was marked “not downloaded” in banner)
  • doi:10.1074/jbc.M109.088682 (Pacholec 2010, JBC) — PDF verified (10 pages)
  • doi:10.1126/science.1231097 (Hubbard 2013, Science) — PDF verified (5 pages)

Corrections made:

  1. SCHEMA — human-evidence-level: limited → limited-negative (R26a update: STAC class has multiple Phase 2/3 null trials; confirmed appropriate).
  2. SCIENTIFICALLY IMPORTANT — Baur 2006 n-per-group: footnote said n=~20/group; source states n=60 (SD) and n=55 (HC, HCLR, HCR) at start. Corrected in footnote and body.
  3. SCIENTIFICALLY IMPORTANT — Pearson 2008 n-per-group: footnote said n=~20–30/group; source states n=60 (SD), 55 (SDLR), 54 (SDR) at start. Corrected in footnote and body.
  4. SCIENTIFICALLY IMPORTANT — Pearson 2008 transcriptional overlap: footnote claimed “144 genes shared with CR signature” — this number does not appear in Pearson 2008. Pearson 2008 reports pathway-level overlap percentages (82%/76%/96%/64% across liver/muscle/adipose/heart vs EOD feeding). “144 of 153 pathways” is from Baur 2006 (pathway analysis, not gene count). Corrected in footnote; body text updated to report Pearson 2008’s actual metric.
  5. SIGNIFICANT — Pacholec 2010 off-target claim: body and footnote said “14–38 unrelated kinases and receptors” for “all four STACs” — misleadingly collapses per-compound data. Source specifies: SRT1720 >50% inhibited 38 targets, SRT2183 14, SRT1460 20, resveratrol 7 (panel of >100 targets at 10 µM). Corrected in body and footnote.
  6. SIGNIFICANT — Wood 2004 C. elegans lifespan: footnote claimed ~15% extension. Source shows ~10% averaged across food conditions (up to ~14% on dead E. coli at 100 µM). Corrected in footnote.
  7. MINOR — Baur 2006 body text: added hazard ratios (HCR vs HC: 0.69, p=0.020; HCR vs SD: 1.03, p=0.88) and specified strain as C57BL/6NIA.
  8. MINOR — Pearson 2008 body text: added tissue-specific pathway-overlap percentages; clarified dose in food (400 mg/kg food ≈ 22 mg/kg/day); clarified health benefits (bone, cataract, vascular, motor) achieved without lifespan extension.
  9. MINOR — Pacholec 2010 footnote: added ob/ob mouse 100 mg/kg lethality detail (3/8 mice died).
  10. MINOR — Removed ⚠️ auto-extraction banner; flipped verified: true.

Unverifiable claims (remained in place with gap tags):

  • Mercken NHP resveratrol study — DOI not confirmed; unsourced tag preserved
  • SRT2104 GSK trial publication status — unsourced tag preserved

Downstream propagation needed (main agent):

  • resveratrol compound page (when seeded) — will inherit n-per-group, lifespan %, and off-target data from this class page; ensure compound page uses corrected figures
  • sirtuin pathway page — may summarize STAC mechanism; check for “14–38 off-target” or “15% C. elegans” claims
  • nad-precursors sister page — not affected by these corrections but referenced as sibling class
  • deregulated-nutrient-sensing hallmark page — may cite Baur 2006 or Pearson 2008 with old n values

[2026-05-06] verify | interventions/pharmacological/mtor-inhibitors.md

Pages verified: 1

  • interventions/pharmacological/mtor-inhibitors.md (corrections: 7 — see below)

Sources checked:

  • doi:10.1038/nature08221 (Harrison 2009, Nature) — local PDF, read end-to-end (5 pages)
  • doi:10.1016/j.cmet.2009.11.010 (Bjedov 2010, Cell Metabolism) — local PDF, read end-to-end (12 pages)
  • doi:10.18632/aging.206235 (Moel/PEARL 2025, Aging) — local PDF, read end-to-end (29 pages including methods and supplementary)
  • doi:10.1111/acel.12405 (Arriola Apelo 2016, Aging Cell) — local PDF, read end-to-end (11 pages)
  • doi:10.1126/scitranslmed.3009892 (Mannick 2014, Sci Transl Med) — not_oa; PubMed abstract + Crossref metadata only no-fulltext-access
  • doi:10.1126/scitranslmed.aaq1564 (Mannick 2018, Sci Transl Med) — not_oa; Crossref abstract only no-fulltext-access
  • doi:10.1016/S2666-7568(21)00062-3 (Mannick 2021, Lancet Healthy Longevity) — not_oa; PubMed PMID:33977284 full abstract used no-fulltext-access
  • doi:10.1038/nature11861 (Johnson 2013 review, Nature) — not_oa; not verified

Corrections made:

  1. SCIENTIFICALLY CRITICAL — Mannick 2018 trial identity and n were wrong. Wiki attributed n=652 and “RTB101 ± everolimus” to doi:10.1126/scitranslmed.aaq1564. The actual Mannick 2018 Sci Transl Med paper is a Phase 2a study (n=264) using BEZ235 (catalytic mTOR inhibitor, the compound that later received the code RTB101) + RAD001/everolimus combination vs placebo. The n=652 Phase 2b and n=1,024 Phase 3/PROTECTOR1 data are from a separate publication: Mannick 2021, Lancet Healthy Longevity (PMID:33977284). Body text, footnote, and clinical summary table corrected.
  2. SCIENTIFICALLY CRITICAL — PROTECTOR1 population eligibility corrected. Wiki said “elderly with ≥2 risk factors for respiratory illness” — the Phase 3 PROTECTOR1 trial excluded COPD and current smokers (different from Phase 2b). Corrected to “adults ≥65 years without COPD and not current smokers.” Outcome details added (OR 1.07, p=0.65).
  3. SIGNIFICANT — Mannick 2018 primary endpoint clarified. Wiki said primary endpoint was “respiratory illness rate” and that RTB101+everolimus significantly reduced it. The actual Phase 2a (n=264) primary endpoint was infection rate over 1-year follow-up; the BEZ235+RAD001 combination showed p=0.001. The narrative about “RTB101 alone did not reach significance” was from the Phase 2b pooled analysis (2021 paper), not the 2018 paper. These have been disentangled.
  4. SIGNIFICANT — Phase 2b (Mannick 2021) finding clarified. The Phase 2b trial showed RTB101 10 mg/day alone significantly reduced lab-confirmed RTIs in the pooled parts 1+2 analysis (19% vs 28%; OR 0.601, p=0.02). The implication section previously stated “mTORC1 inhibition alone may be insufficient” based on misattribution — corrected to note the Phase 3 failure is more plausibly explained by endpoint differences (symptomatic vs lab-confirmed) and population differences.
  5. Bjedov 2010 Drosophila section: enriched with mechanism detail from paper (TORC1 specificity confirmed; 4E-BP and autophagy ATG5 both required; CR confound excluded; extension beyond DR maximum demonstrated).
  6. PROTECTOR1 primary endpoint description corrected: was “no significant reduction in symptomatic respiratory illness confirmed by PCR” — Phase 3 endpoint was clinically symptomatic respiratory illness not required to be PCR-confirmed (different from Phase 2b). Corrected.
  7. RTB101 compound table row: updated to reflect BEZ235=RTB101 identity and Phase 2a/2b/Phase 3 progression.

Pages unverifiable (not_oa):

  • Mannick 2014 (doi:10.1126/scitranslmed.3009892) — abstract confirms n≈218 and ~20% vaccine titer improvement; detailed arm-level data not independently verifiable without full text
  • Mannick 2018 (doi:10.1126/scitranslmed.aaq1564) — abstract confirms n=264 and BEZ235+RAD001 design; p=0.001 from PubMed abstract
  • Mannick 2021 (doi:10.1016/S2666-7568(21)00062-3) — PubMed full abstract used; n, OR, and p-values confirmed

Downstream pages needing review (for main agent propagation):

  • rapamycin (molecules/compounds/rapamycin.md) — likely carries Mannick 2018 n=652/RTB101 framing inherited from this class page; should be corrected to match
  • everolimus (molecules/compounds/everolimus.md, if it exists) — may cite Mannick 2014/2018 data
  • mtor (pathways/mtor.md) — may cite Mannick trial data for human clinical evidence; check for n=652 attribution
  • deregulated-nutrient-sensing (hallmarks/) — may reference ITP lifespan data or Mannick human evidence
  • hyperfunction-theory (hypotheses/) — may cite Harrison 2009 and Mannick trials as supporting evidence

[2026-05-06] verify | interventions/lifestyle/sleep.md

Pages verified: 1 (partial — Xie 2013 PDF unavailable; Carroll 2015 verified via publisher abstract only)

Sources checked:

  • doi:10.1126/scitranslmed.3003748 (Iliff et al. 2012, Sci Transl Med) — PDF verified end-to-end (local archive; glymphatic system architecture)
  • doi:10.1016/j.neuron.2017.02.004 (Mander et al. 2017, Neuron) — PDF verified end-to-end (downloaded on demand; SWS aging review)
  • doi:10.1093/sleep/33.5.585 (Cappuccio et al. 2010, Sleep) — PDF verified end-to-end (downloaded on demand; sleep-mortality meta-analysis)
  • doi:10.1161/JAHA.117.005947 (Yin et al. 2017, JAHA) — PDF verified end-to-end (downloaded on demand; dose-response mortality meta-analysis)
  • doi:10.5665/sleep.4398 (Carroll et al. 2015, Sleep) — PDF download failed (bronze OA, URL mismatch); verified via publisher abstract and Crossref metadata only
  • doi:10.1126/science.1241224 (Xie et al. 2013, Science) — PDF download failed (no OA copy in archive); attribution confirmed correct via cross-reference with citing literature

Corrections made:

  1. SCIENTIFICALLY CRITICAL — Carroll 2015 age-comparison direction inverted: wiki stated TLR4/NF-kB activation was “stronger in older vs younger adults” after sleep deprivation. The actual finding is the opposite: younger adults showed increased TLR4-stimulated monocyte inflammation after partial sleep deprivation; older adults showed reduced baseline TLR4-stimulated inflammation that was NOT further activated by sleep loss. Age × sleep-loss interaction p<0.05. Corrected in body text and footnote.
  2. Carroll 2015 n and age: “n=29 older adults (mean ~61y) + younger comparison group” → “n=49 older adults (ages 60–84) + n=21 younger adults (ages 25–39); total n=70.” Also corrected “total sleep deprivation” → “partial sleep deprivation (03:00–07:00 window).”
  3. Cappuccio 2010 OR → RR: wiki used “OR” (odds ratio) throughout; paper reports relative risk (RR). Short sleep RR 1.12 (95% CI 1.06–1.18); long sleep RR 1.30 (95% CI 1.22–1.38). Corrected in mortality table, body text, and footnote.
  4. Cappuccio 2010 n and cohort count: “~1.3 million across 16 prospective cohort studies” → “n=1,382,999; 16 studies providing 27 independent cohort samples; 112,566 deaths.” Corrected in footnote.
  5. Mander 2017 SWS amplitude reduction range: “50–75% reduction in slow-wave activity amplitude” → “~75–80% reduction in frontal SWA amplitude (Figure 1B; maximal in prefrontal region during first NREM cycle).” The 50% lower bound was not supported by stated paper figures.
  6. Yin 2017 n and study count: “~74 prospective cohort studies” → “67 articles (141 independent reports); n=3,582,016 total participants.” Corrected in footnote.
  7. Mortality table restructured: source column added to distinguish Yin 2017 dose-response from Cappuccio 2010 categorical; “OR”/“HR” changed to “RR” throughout.
  8. Iliff 2012 footnote expanded with exact strain (C57BL/6 8–12 wk), Aqp4-null solute clearance reduction (~70%), AQP4-null Aβ clearance (~55% reduction); marked “verified against full PDF.”
  9. Mander 2017 footnote expanded with SWA amplitude finding and sleep-need debate conclusion.
  10. Xie 2013 footnote: “pending download” → “failed (no OA copy)”; no-fulltext-access tag added.

Pages unverifiable:

  • doi:10.1126/science.1241224 (Xie 2013) — no OA copy; PDF download failed twice; tagged no-fulltext-access in footnote
  • doi:10.5665/sleep.4398 (Carroll 2015) — bronze OA but URL mismatch in archive; verified via abstract only; quantitative values approximate

Downstream propagation needed (for main agent):

  • astrocytes — confirmed R26b propagation need is resolved: sleep.md correctly attributes 60% interstitial space expansion to Xie 2013, not Iliff 2012
  • alzheimers-disease — verify glymphatic Aβ clearance claim attributes mechanistic architecture to Iliff 2012 and sleep-dependence to Xie 2013 separately, not conflated
  • chronic-inflammation hallmark — if it cites Carroll 2015 for sleep-deprivation → TLR4/NF-kB claim, the directionality correction (effect primarily in younger adults, blunted in older adults) must be propagated
  • loss-of-proteostasis — if it cites Xie 2013 for glymphatic clearance, verify attribution is to Xie 2013 (not Iliff 2012)

[2026-05-06] verify | interventions/lifestyle/intermittent-fasting.md

Pages verified: 1 (partial — de Cabo & Mattson 2019 NEJM and Goodrick 1990 closed-access; Trepanowski 2017 and Sutton 2018 download failed, verified via PubMed abstract + PMC full text; Liu 2022 NEJM full PDF read end-to-end)

Sources checked:

  • doi:10.1056/NEJMoa2114833 (Liu D et al. 2022, NEJM) — local PDF downloaded and read end-to-end
  • doi:10.1016/j.cmet.2018.04.010 (Sutton EF et al. 2018, Cell Metab) — PMC full text (PMC5990470) read; download failed (no PDF captured)
  • doi:10.1001/jamainternmed.2017.0936 (Trepanowski JF et al. 2017, JAMA Intern Med) — PubMed abstract + Crossref metadata; download failed (HTTP 403)
  • doi:10.1056/NEJMra1905136 (de Cabo R, Mattson MP 2019, NEJM) — not_oa; PubMed + Crossref metadata only; no-fulltext-access
  • doi:10.1146/annurev-nutr-071816-064634 — confirmed this DOI resolves to Patterson & Sears 2017 (NOT Mattson); corrected footnote
  • doi:10.1016/0047-6374(90)90107-q (Goodrick CL et al. 1990, Mech Ageing Dev) — not_oa; Crossref metadata only; no-fulltext-access
  • doi:10.1002/oby.22065 (Anton SD, Mattson MP et al. 2018, Obesity) — PubMed abstract; used as replacement for wrong Mattson DOI

Corrections made:

  1. Liu 2022 weight loss numbers reversed: wiki said “~6.3 kg TRE vs ~8.0 kg CR”; actual paper reports TRE −8.0 kg (95% CI −9.6 to −6.4) vs CR −6.3 kg (95% CI −7.8 to −4.7); P=0.11. SCIENTIFICALLY CRITICAL — the arms were swapped.
  2. Liu 2022 caloric targets: “1200–1500 kcal” stated for both sexes; corrected to sex-stratified: 1500–1800 kcal/day men, 1200–1500 kcal/day women.
  3. Liu 2022 n: n=118 (completers only) → n=139 randomized (118 completed 12-month follow-up).
  4. [^mattson2017annu] DOI attribution wrong: doi:10.1146/annurev-nutr-071816-064634 resolves to Patterson & Sears 2017, NOT Mattson. SCIENTIFICALLY CRITICAL — replaced entire footnote with Anton et al. 2018 (Obesity, doi:10.1002/oby.22065), the actual Mattson-co-authored metabolic-switch review; footnote label changed to [
  5. Sutton 2018 eating window: “5-hr eating window” → “6-hr eating window (dinner before 3pm)” per PMC full text.
  6. Sutton 2018 n: expanded to “n=8 completers (12 enrolled; 4 withdrew)” with elibility criteria (men only, age 35–70, prediabetes defined as elevated HbA1c AND impaired OGTT); added specific BP and insulin sensitivity numbers from PMC.
  7. Sutton 2018 weight: added explicit confirmation that body weight was maintained between arms (p=0.12), strengthening the “calorie-independent” claim.
  8. de Cabo 2019 citation count removed: “1,439 citations” removed from body (dynamic value; Crossref shows 1,311 at time of verification — discrepancy between counting systems).
  9. Mitchell 2019 (doi:10.1016/j.cmet.2018.08.011) added: Cell Metab paper showing fasting-specific survival benefits independent of caloric restriction in C57BL/6 mice; added to Rodent IF lifespan section and footnotes.

Pages unverifiable (closed-access):

  • doi:10.1056/NEJMra1905136 (de Cabo & Mattson 2019) — not_oa; tagged no-fulltext-access
  • doi:10.1016/0047-6374(90)90107-q (Goodrick 1990) — not_oa; quantitative lifespan magnitudes unverified

Downstream propagation needed (for main agent):

  • caloric-restriction (interventions/lifestyle/caloric-restriction.md) — may cite Mitchell 2019 or reference the IF/CR confound; if not, consider adding Mitchell 2019 to the “meal-frequency vs caloric restriction” discussion
  • time-restricted-eating (interventions/lifestyle/time-restricted-eating.md) — inherits Liu 2022 and Sutton 2018 citations; check for reversed weight-loss numbers and 5-hr vs 6-hr eating window errors
  • deregulated-nutrient-sensing (hallmarks/) — may reference IF human evidence; check for de Cabo 2019 citations with overstated conclusions
  • ketogenic-diet (interventions/lifestyle/ketogenic-diet.md or molecules/compounds/) — shares BHB/ketosis mechanism claims sourced from what was [^mattson2017annu]; those claims now link to Anton 2018; check if ketogenic-diet page cites the same wrong DOI
  • autophagy (processes/autophagy.md) — [^martinez-lopez2017] autophagy induction threshold remains unsourced; if autophagy page carries this 12–16 hr threshold claim, it needs primary citation

[2026-05-06] verify | interventions/lifestyle/ketogenic-diet.md

Pages verified: 1 (partial — Shimazu 2013 PDF download failed)

Sources checked:

  • doi:10.1016/j.cmet.2017.08.004 (Newman et al. 2017, Cell Metabolism) — PDF downloaded on demand and verified end-to-end
  • doi:10.1016/j.cmet.2017.08.005 (Roberts et al. 2017, Cell Metabolism) — PDF downloaded on demand and verified end-to-end
  • doi:10.2196/diabetes.6981 (McKenzie et al. 2017, JMIR Diabetes) — PDF downloaded on demand and verified end-to-end
  • doi:10.1038/nm.3804 (Youm et al. 2015, Nature Medicine) — local archive; verified end-to-end
  • doi:10.1126/science.1227166 (Shimazu et al. 2013, Science) — green OA; DOI lookup failed twice; no-fulltext-access in footnote; cross-checked via Roberts 2017 citations only

Corrections made:

  1. SCIENTIFICALLY CRITICAL — Newman 2017 lifespan framing: wiki claimed “~13% improvement in median lifespan” for Newman 2017. Actual finding is reduced MIDLIFE MORTALITY (final log-rank p=0.78 — overall lifespan NOT significantly extended; maximum lifespan NOT increased). The ~13.6% median-lifespan figure belongs to Roberts 2017, not Newman. Section heading, body, and footnote all corrected.
  2. SCIENTIFICALLY CRITICAL — Newman 2017 sample sizes added: N=61 Control, 81 Cyclic KD, 36 Cyclic HF, 37 KD continuous, 26 HF continuous.
  3. SCIENTIFICALLY CRITICAL — Roberts 2017 lifespan: median 1003 vs 886 days (+13.6%, p<0.05) confirmed; maximum lifespan p=0.16 (NOT significant). Wiki said “median AND maximum extended” — maximum extension is not statistically significant.
  4. Roberts 2017: n=43-44/group added; LCD intermediate result (943 days) documented.
  5. SCIENTIFICALLY CRITICAL — McKenzie 2017 HbA1c: “6.3%” to “6.6%” (Table 1 all-subjects mean). Reduction is 1.0 pp (not 1.3 pp).
  6. SCIENTIFICALLY CRITICAL — McKenzie 2017 weight loss: “~12%” to “7.2% (SD 3.7%; -8 kg)”. The original was ~1.7x too high.
  7. McKenzie 2017 medication claim: “56% eliminated” to “56.8% had >=1 medication reduced OR eliminated.”
  8. BHB-NLRP3 mechanism (R26b propagation): added Section 3c with K+-efflux / ASC-oligomerization inhibition framing per Youm 2015; AMPK/ROS/autophagy/GPR109a/HDAC all ruled out as mediators; MCC950 ATPase mechanism noted as distinct.
  9. New Youm 2015 footnote added; nlrp3-inhibition added to mechanisms frontmatter.

Pages unverifiable: Shimazu 2013 — no-fulltext-access in footnote; retry on next lint pass.

Downstream propagation needed (main agent):

  • caloric-restriction — verify KD comparison table does not misframe Newman 2017 as median-lifespan extension
  • intermittent-fasting — same KD comparison table check
  • nlrp3-inflammasome (if seeded) — verify BHB mechanism uses K+-efflux/ASC framing, not ATPase
  • chronic-inflammation hallmark MOC — verify KD-NLRP3 framing post-R26b
  • beta-hydroxybutyrate (stub, pending seed) — when seeded, BHB-NLRP3 must use K+-efflux / ASC oligomerization per Youm 2015

[2026-05-06] R26c — R23 atomic-page verification sweep (DONE; 22/22 pages)

All 22 R23 atomic pages verified. Each page now ships verified: true with verified-scope line documenting PDF-verified vs partial/abstract-only sources. Auto-extraction banners removed across the entire R23-output set.

Wiki-wide state after R26c close: 0 pages remain verified: false (excluding meta files: ROADMAP.md, CLAUDE.md, log.md, sops/,.claude/agents/). The wiki is fully verified for the first time.

Pages verified (sub-batch / page → critical-correction one-liner):

R26c batch 1 — stem-cell-therapy + cell-types (7):

  • mesenchymal-stem-cells — [^styner2016] wrong first author (Styner→Singh) and wrong pages; Mendez-Ferrer 2010 depletion claim narrowed (immature CD150+CD48− LSK pool only, not all HSCs); Wagner 2008 SASP attribution removed (paper has no SASP profiling).
  • neural-stem-cells — Boldrini 2018 mischaracterized (Sox2+ QNPs DID decline selectively in anterior-mid DG; wiki said no decline); Kalamakis 2019 timepoints (6/20 mo → 7/22 mo); Kalamakis deep-quiescence driver (Notch DLL → sFRP5 + IFN signals); Sanai 2011 corridor depletion timeline.
  • induced-pluripotent-stem-cells — Lapasset 2011 Werner syndrome claim removed (paper has only senescent + centenarian donors, no Werner); Olova 2019 mischaracterized (observational time-course of continuous OSKM, NOT factor-withdrawal); 6-factor OSKMNL methodology specified (4-factor OSKM fails on senescent cells).
  • HSCT — Jazbec 2022 strain (C57BL/6 → BALB/c, matters for Th1/Th2 skew); Gibson 2022 DNMT3A-CH direction reversed (improves OS HR=0.78 via GVL, NOT worsens; risk only with calcineurin not PTCy); Kollman 2016 cohort breakdown.
  • mesenchymal-stem-cell-therapy — Golpanian 2017 mean age 76→78.4 (the 76 was the 100M subgroup mean); Tompkins 2017 TNF-α dose-stratified; Remestemcel-L confirmed Dec 2024 FDA approval.
  • ipsc-derived-cell-therapy — Doi 2025 Parkinson’s iPSC trial: bilateral injections (not unilateral); ¹⁸F-DOPA PET (not DAT-SPECT); efficacy denominator 4/6 not 4/7; Mandai 2017 (3 CNV deletions, no SNV oncogenic variants); Lapasset 2011 epigenetic-clock language removed (Lapasset measures telomere/transcriptome, not Horvath clock — Horvath published 2013).
  • in-vivo-partial-reprogramming-therapy — Yang 2023 Aging MEK inhibitor fabrication (paper uses GSK-3α/β + TGF-β + valproic acid, NO MEK inhibitor); Lapasset 2011 OSKMNL caveat added; Yang 2023 Aging primary readout corrected (transcriptomic clocks, not DNAm).

R26c batch 2 — gene-therapy (4):

  • aav-klotho — Kurosu 2005 strain C57BL/6 → C3H (4× backcross); Kurosu circulating Klotho ~2-4× → ~2×; fabricated p53/p21 claim removed; per-line lifespan breakdown added (EFmKL46 male +20.0%, EFmKL48 male +30.8%, females ~19%); Davidsohn 2019 AAV serotype AAV9 → AAV8; Castner 2023 half-life “minutes-hours” → 29.5h NHP / ~7min rodent.
  • aav-osk — Lu 2020 vision-arm ages 12 mo → 3 and 11 mo; Lu 2020 systemic safety 20 mo → 21 mo; Tet-On vs Tet-Off architecture clarified per arm; Yang 2023 marmoset claim removed (fabricated; not in paper); Yang 2023 ICE timeline added (4-6 mo treatment, ~10 mo post-treatment, 14-16 mo at assessment, 57% DNAm reversal); Ocampo 2016 strain corrected.
  • aav-follistatin — Lee 2001 follistatin effect “comparable to KO” → “met or exceeded KO” (paper attributes to GDF-11/activins also blocked); Mendell 2017 6MWT magnitudes ~15-50m → +58 to +153m for 4 responders; Mendell biopsy endpoint corrected (no myostatin reduction in paper); registered primary endpoint clarified (safety, 6MWT post-hoc).
  • crispr-base-editing-pcsk9 — Chadwick 2017 delivery (hydrodynamic → adenoviral); Chadwick 2017 PCSK9 reduction ~90% → ~56% (90% belongs to Musunuru 2021 NHP); Lee 2022 dose-stratified results; pvad095 author attribution.

R26c batch 2 — pharmacological class pages (4):

  • mtor-inhibitors — Mannick 2018 trial identity completely wrong (n=652 belongs to Mannick 2021 Lancet Healthy Longevity PROTECTOR1 Phase 3, NOT the n=264 BEZ235+RAD001 Phase 2a paper); PROTECTOR1 reframed (Phase 2b RTB101-only DID work, OR 0.601 p=0.02; Phase 3 failure attributable to switch to symptomatic-illness endpoint and lower-risk population); separate [^mannick2021protector] footnote added.
  • ampk-activators — Cabreiro 2013 dose breakdown (100 mM near-null, NOT in 18-36% range); metformin AMPK demoted (primary mechanism is mitochondrial Complex I inhibition + fructose-1,6-bisphosphatase via AMP per Foretz 2014 LKB1/AMPKα-null studies; AMPK is consequence not driver); Martin-Montalvo χ² values added; UKPDS n framing corrected (1704 = total overweight pool; primary metformin vs conventional was 342 vs 411).
  • nad-precursors — Martens 2018 ~2.7-fold belongs to Trammell 2016 n=1 pilot, not Martens cohort (correct Martens: ~60% over placebo p=0.048); Yoshino 2021 muscle steady-state NAD+ DID NOT change (only turnover metabolites; physical function and mitochondrial respiratory capacity null); Mills 2016 delivery drinking water not SC; Massudi 2012 tissue (pelvic skin only); Grozio 2019 Slc12a8 does NOT transport NR.
  • sirtuin-activators — human-evidence-level: limited-negative confirmed appropriate; Baur 2006 + Pearson 2008 n-per-group corrected (60/55 not ~20-30); Pearson 2008 transcriptional 144-genes-shared-with-CR misattribution (belongs to Baur 2006); Pacholec 2010 off-target counts disambiguated (SRT1720 38 / SRT2183 14 / SRT1460 20 / resveratrol 7); Wood 2004 C. elegans ~15% → ~10% (averaged).

R26c batch 2 — lifestyle (6):

  • sleep — Carroll 2015 age-comparison direction inverted (TLR4 activation actually stronger in YOUNGER, blunted in older); Cappuccio 2010 OR → RR throughout; cohort count and n corrected (1,382,999 across 27 independent cohort samples); Mander 2017 SWA reduction range corrected (75-80%, not 50-75%).
  • intermittent-fasting — Liu 2022 arms reversed (TRE actually lost MORE weight: TRE −8.0kg vs CR −6.3kg, P=0.11); wrong Mattson DOI (resolves to Patterson & Sears 2017, replaced with Anton 2018 Obesity); Liu 2022 sex-stratified caloric targets; Mitchell 2019 added for fasting-vs-CR confound discussion.
  • methionine-restriction — Miller 2005 strain (C57BL/6×DBA/2)F1 → (BALB/c×C57BL/6)F1 (CB6F1); Olsen 2020 duration 8 weeks → 7 days; Plummer & Johnson 2022 author attribution wrong (“Ables et al.” → Plummer JD/Johnson JE); GCN2 mechanism corrected (Gen2-/- alone insufficient — PERK provides redundant ISR activation per Fang 2022); Tang 2022 model (HEK293T → Drosophila SAMTOR crystal + functional mutagenesis).
  • ketogenic-diet — Newman 2017 13% lifespan figure belongs to Roberts 2017 (Newman shows midlife mortality reduction only, p=0.78 final log-rank); McKenzie 2017 HbA1c 6.3%→6.6% and weight loss 12%→7.2% (was ~1.7× too high); Roberts 2017 maximum lifespan p=0.16 NOT significant; BHB-NLRP3 mechanism corrected per R26b (K⁺-efflux/ASC oligomerization, not ATPase).
  • heat-exposure — HSF1 localization framing per R26b (basal predominantly nuclear, not cytoplasm→nucleus translocation); BMC Medicine 2018 sex-specificity narrowed (women: 0/73 events, association not estimable); Laukkanen 2017 dementia paper p-values added; KIHD baseline CHD ~24% prevalent at baseline (not “no CVD”).
  • time-restricted-eating — Sutton 2018 cortisol claim fabricated (paper showed unchanged Δ=−0.1±1.3 μg/dl p=0.95); Liu 2022 “ad libitum” framing wrong (both arms followed identical prescribed CR; spontaneous-400-kcal claim deleted); Sutton 2018 insulin endpoint (HOMA-IR → 3-hour incremental AUC ratio); Sutton 2018 BP separated systolic/diastolic; Hatori 2012 strain + n details.

R26c batch 2 — studies (1):

  • li-2023-dual-mtor-nad-gerotherapy — PDB 5514 → 4BN4 (single critical accession ID); SOD1-G93A strain SV129 (NASH/HFD experiments are C57BL/6J — two-background design); COI disclosure replaced (Jerry Colca patent claim absent from preprint).

R26c sweep statistics:

  • 22 pages verified
  • ~110 source-text-level corrections
  • ~25 scientifically critical (residue/strain/cohort/PDB-ID drift, fabricated experiments, inverted directions, swapped first-authors, wrong DOIs)
  • 4 fabricated experimental claims removed (Yang 2023 marmoset; Sutton 2018 cortisol; Kurosu 2005 p53/p21; Yang 2023 Aging MEK inhibitor)
  • 4 reversed-direction errors caught (Boldrini 2018 Sox2+ NSC decline; Gibson 2022 DNMT3A-CH outcome direction; Carroll 2015 TLR4 age-direction; Liu 2022 TRE/CR weight-loss arms)

Cross-cutting propagation done by main agent (R26c post-verification pass):

  • klotho protein page — Kurosu 2005 strain (C57BL/6→C3H), fold-change (~2-4×→~2×), p53/p21 fabrication removed, per-line lifespan breakdown added; footnote propagation note documenting the “Dubal 2014 methods” misattribution origin.
  • All other propagation needs spot-checked across atomic pages — most claims were contained on the source page; minor inheritances (sleep page Iliff/Xie attribution; intermittent-fasting “ad libitum” framing; nlrp3-inflammasome BHB ATPase; clonal-hematopoiesis donor-CHIP framing; partial-reprogramming MEK/Werner; mitofusins Karbowski 2002) all checked via grep and confirmed clean or fixed.

Graph View

Backlinks