R6

log/R6.md — Round 6 entries

Sub-file of log — see parent for index.

[2026-05-04] ingest | round-6e-followup

bcl2l13

• added: molecules/proteins/bcl2l13.md
• entity type: protein (485 aa, isoform 2 OMM-targeted; BCL-2 family; mitophagy receptor; functional Atg32 ortholog)
• canonical IDs: UniProt Q9BXK5 (confirmed via UniProt REST API text), NCBI Gene 23786 (confirmed via NCBI Gene page), HGNC 15769 (confirmed from task brief + NCBI Gene)
• DOIs cited: 10.1074/jbc.m010520200 (Kataoka 2001 JBC — original cloning; archive pending), 10.1038/ncomms8527 (Murakawa 2015 Nat Commun — Atg32 ortholog; CRITICAL; 491 citations; gold OA, archive pending), 10.1080/15548627.2015.1084459 (Otsu 2015 Autophagy — commentary; 95 citations; archive pending), 10.1016/j.celrep.2024.115001 (Murakawa 2024 Cell Reports — AMPKα2 Ser272 cardiac axis; 11 citations; gold OA, archive pending), 10.1080/15548627.2025.2465408 (Murakawa 2025 Autophagy — cardiac phospho-BCL2L13; 4 citations; closed-access not_oa), 10.3389/fcell.2022.1065702 (Kataoka 2022 Frontiers review; gold OA, archive pending)
• DOI correction (task brief): brief cited DOI 10.1016/s0960-9822(01)00373-6 for Kataoka 2001 (Curr Biol). This DOI resolves in DOI lookup to “A natural variability in the proline-rich motif of Nef modulates HIV-1 replication in primary T cells” (Curr Biol 11:920-924, 2001) — confirmed BUG-2-type mismatch. Correct Kataoka 2001 DOI: 10.1074/jbc.m010520200 (JBC 276:19548-19554), confirmed via PubMed PMID 11262395 efetch.
• CRITICAL kinase correction (task brief): brief stated “ULK1 phosphorylates BCL2L13 Ser272 → enhances LIR-LC3 binding.” Primary literature (Murakawa 2024 Cell Reports + Murakawa 2025 Autophagy) establishes the kinase as AMPKα2 (PRKAA2), not ULK1. Page uses AMPKα2 throughout with a correction note in the auto-extraction banner.
• implicit stubs created: kataoka-2001-bcl-rambo-apoptosis, murakawa-2015-bcl2l13-atg32-ortholog, otsu-2015-bcl2l13-atg32-commentary, murakawa-2024-ampk-bcl2l13-cardioprotection, murakawa-2025-bcl2l13-ampk-cardiac, kataoka-2022-bcl-rambo-review
• gaps surfaced: exact LIR core residue positions (WXXI) in human BCL2L13 not confirmed from primary sequence; GenAge ID null; Ensembl ID null; BCL2L13 aged-tissue expression trajectory not established; BH domain binding affinities to BCL-2/BCL-xL not quantified; apoptosis mechanism via C-terminal extension unresolved; cardiac aging human data lacking; no ULK1 vs AMPKα2 competitive data
• ROADMAP: bcl2l13 marked [x] drafted R6e 2026-05-04

ndp52

• added: molecules/proteins/ndp52.md
• entity type: protein (446 aa; selective autophagy cargo receptor; mitophagy + xenophagy)
• canonical IDs: UniProt Q13137 (CALCOCO2_HUMAN), NCBI Gene 10241, HGNC 29912 — not live-re-verified; recommend lint-pass re-check
• DOIs cited (archive-confirmed): 10.1038/ni.1800 (Thurston 2009 Nat Immunol — NDP52 + TBK1 + bacteria; local PDF at ); 10.1038/nature14893 (Lazarou 2015 Nature — NDP52+OPTN primary receptors; not_oa, no local PDF); 10.1016/j.molcel.2012.08.024 (von Muhlinen 2012 Mol Cell — LC3C/NDP52; pending download); 10.1038/ncb2589 (Tumbarello 2012 NCB — myosin VI + receptors; pending download, green OA)
• BUG-2 mismatch logged: DOI 10.1016/j.molcel.2012.04.014 (cited in task brief) resolves in archive to XIAP/LUBAC paper; correct DOI is 10.1016/j.molcel.2012.08.024 (confirmed Crossref)
• DOI not found: 10.1083/jcb.201503102 (Tumbarello 2015 JCB from task brief) — unresolvable in archive; information not cited
• implicit stubs created: tbk1, xenophagy, lc3c, thurston-2009-ndp52-xenophagy, lazarou-2015-pink1-receptors-mitophagy, vonmuhlinen-2012-ndp52-lc3c, tumbarello-2012-myosin-autophagy-receptors
• gaps surfaced: NDP52 aged-tissue expression unsourced; Parkin-independent PINK1 route in-vivo mechanism unclear; LC3C-NDP52 complex PDB ID absent; galectin-8/NDP52 binding interface uncrystallized; no GenAge entry; Lazarou 2015 not_oa (no PDF verification possible)
• ROADMAP: ndp52 marked [x] drafted R6e 2026-05-04, verified: false

atg13

• added: molecules/proteins/atg13.md
• entity type: protein (517 aa; ULK1 complex scaffold; mTORC1 regulatory target)
• canonical IDs: UniProt O75143 (confirmed via REST API), NCBI Gene 9776, HGNC 29091
• DOIs cited: 10.1091/mbc.e08-12-1248 (Hosokawa 2009, local PDF confirmed), 10.1074/jbc.m900573200 (Ganley 2009, local PDF confirmed), 10.4161/auto.5.5.8249 (Mercer 2009 ATG101 discovery, archive pending), 10.4161/auto.5.7.9296 (Hosokawa 2009b ATG101 Autophagy, archive pending), 10.1074/jbc.M115.689646 (Puente 2016 ATG13 phospho-sites, archive pending), 10.1080/15548627.2015.1091144 (Suzuki 2015 open/closed HORMA, archive pending), 10.1038/nature04723 (Komatsu 2006, local PDF available — cited for neurodegeneration analogy only)
• DOI scrutiny: user brief suggested doi:10.1016/j.molcel.2011.06.025 for “Joo 2011 ATG13 phospho” — DOI lookup confirms this resolves to “Distinct Interactions Select and Maintain a Specific Cell Fate” (138 citations, cell biology paper unrelated to ATG13). DOI NOT used. Correct ATG13 phospho mapping sourced to Puente 2016 (10.1074/jbc.M115.689646, confirmed via PubMed PMID 26801615).
• implicit stubs created: fip200, atg101 (already seeded in R6e parallel), hosokawa-2009-ulk1-atg13-fip200-mtor, ganley-2009-ulk1-atg13-fip200-essential, mercer-2009-atg101-discovery, hosokawa-2009-atg101, puente-2016-atg13-phosphorylation
• gaps surfaced: no ATG13 KO mouse data; no ATG13 Mendelian disease; human HORMA structure inferred from yeast/mouse; Ser-224 kinase identity needs replication; ATG13 aging protein-level data unsourced; GenAge ID absent
• schema decisions: no schema gaps encountered; standard protein frontmatter used; complex-subunits: field not used (ATG13 is a subunit, not the complex page — the complex page, if created, would be ulk1-complex.md)
• ROADMAP: atg13 marked [x] (drafted, verified: false) under Round 6 Core machinery

optn

• added: molecules/proteins/optn.md
• entity type: protein (577 aa; selective autophagy cargo receptor; ALS gene)
• canonical IDs: UniProt Q96CV9 (confirmed via REST API), NCBI Gene 10133, HGNC 17142
• DOIs cited: 10.1126/science.1066901 (Rezaie 2002 Science — POAG characterization; not_oa), 10.1038/nature08971 (Maruyama 2010 Nature — ALS12; local PDF confirmed), 10.1126/science.1205405 (Wild 2011 Science — TBK1/Ser177 xenophagy; not_oa), 10.1073/pnas.1405752111 (Wong 2014 PNAS — mitophagy receptor + E478G; archive pending), 10.1038/nature14893 (Lazarou 2015 Nature — primary receptor establishment; not_oa)
• DOI note: user brief cited 10.1074/jbc.275.27.21036 as Schwamborn 2000 JBC (original FIP-2 cloning). This DOI is absent from the archive index and returned zero PubMed hits across multiple search strategies. Not cited on this page. Verifier should attempt publisher resolution; tagged as gap on page.
• DOI note: user brief cited 10.1038/ncomms3537 for “Wong 2014 Nat Commun OPTN mitophagy” — DOI lookup confirms this is a Condensin-I/chromosome biology paper (78 citations). Wrong DOI. Correct Wong 2014 mitophagy DOI is 10.1073/pnas.1405752111 (PNAS, 775 citations; confirmed via DOI lookup + PubMed efetch PMID 25294927).
• implicit stubs created: tbk1, xenophagy, rezaie-2002-optn-glaucoma, maruyama-2010-optn-als, wild-2011-optn-xenophagy-tbk1, wong-2014-optn-mitophagy, lazarou-2015-pink1-mitophagy-receptors
• gaps surfaced: no GenAge entry (no lifespan experiment in model organism); Schwamborn 2000 DOI unconfirmable; age-related OPTN expression data unsourced; NF-κB/inflammaging link not dissected in aged tissues; E478G dominant-negative vs LoF debate unresolved
• schema decisions: no schema gaps; standard protein frontmatter used
• ROADMAP: optn marked [x] (drafted, verified: false) under Round 6e selective autophagy receptors

atg101

• added: molecules/proteins/atg101.md
• entity type: protein (218 aa; metazoan-specific; no yeast ortholog)
• canonical IDs: UniProt Q9BSB4, NCBI Gene 60673, HGNC 25758 (verify on next lint pass)
• DOIs cited: 10.4161/auto.5.5.8249 (Mercer 2009 — discovery; archive pending), 10.4161/auto.5.7.9296 (Hosokawa 2009 Autophagy — parallel discovery; archive pending), 10.1038/nsmb.3036 (Suzuki 2015 NSMB — HORMA crystal structure; closed-access, no local PDF), 10.1016/j.str.2015.07.011 (Qi 2015 Structure — human ATG13–ATG101 HORMA heterodimer; archive pending)
• DOI correction: user brief cited 10.4161/auto.5.5.8504 for Mercer 2009 — archive resolves that DOI as a rapamycin/mTORC1 paper. Correct Mercer 2009 ATG101 DOI is 10.4161/auto.5.5.8249 (confirmed via PubMed PMID 19287211 efetch).
• DOI correction: user brief cited 10.1038/nsmb.3162 for Suzuki 2015 HORMA structure — archive resolves that DOI as an m6A epitranscriptomics paper. Correct Suzuki 2015 DOI is 10.1038/nsmb.3036 (confirmed via PubMed PMID 26299944 efetch).
• implicit stubs created: atg13 (R6e parallel, in-progress), fip200, gabarapl1, gabarapl2, lc3c, mercer-2009-atg101-discovery, hosokawa-2009-atg101, suzuki-2015-atg101-atg13-horma, qi-2015-human-atg13-atg101-horma
• gaps surfaced: no germline Atg101 KO mouse lifespan data; ATG101 PTM landscape uncharacterized; no aging-specific expression data; WF-finger functional mammalian KI data lacking; Ensembl ID absent (null in frontmatter)
• ROADMAP: atg101 marked drafted (unverified)

[2026-05-04] verify | fundc1

Pages verified: 1 (partial scope — 5 of 9 cited sources verified against primary PDFs; 2 closed-access, 1 corrupted download, 1 download failure)

• molecules/proteins/fundc1.md — corrections made; verified: true (partial scope)

Sources checked (5 verified, 4 unverifiable):

• 10.1002/embr.201438501 Wu 2014 EMBO Rep — VERIFIED
• 10.4161/auto.29568 Wu 2014 Autophagy — VERIFIED
• 10.1080/15548627.2016.1238552 Kuang 2016 Autophagy — VERIFIED
• 10.15252/embj.201593102 Wu 2016 EMBO J — VERIFIED
• 10.1080/15548627.2019.1580095 Lampert 2019 Autophagy — VERIFIED
• 10.1038/ncb2422 Liu 2012 Nat Cell Biol — not_oa; tagged no-fulltext-access
• 10.1016/j.molcel.2014.02.034 Chen 2014 Mol Cell — not_oa; tagged no-fulltext-access
• 10.1080/15548627.2016.1193656 Wu 2016 Autophagy — corrupted download (BUG-4 filed)
• 10.1161/CIRCULATIONAHA.117.030235 Wu 2017 Circulation — download failed; tagged no-fulltext-access
• DOI 10.1126/science.1257529 (excluded by seeder) — confirmed non-existent; exclusion correct

Key corrections to molecules/proteins/fundc1.md:

1. LIR motif sequence CORRECTED: “Tyr18-Asp19-Asp20-Leu21” → “Tyr18-Glu19-Val20-Leu21” (Y18EVL21)
2. Kuang 2016 “~10-fold Ser17 enhancement” REMOVED — paper does not report this; attribution was wrong
3. Kuang 2016 ITC Kd values ADDED: 0.40 µM (unphospho), 1.72 µM (pTyr18, ~4-fold weaker), 0.60 µM (pSer13, minor effect)
4. FUNDC1 LIR noncanonical conformation ADDED: Val20 inserts into HP1 alongside Tyr18
5. ULK1/Ser17 starvation specificity ADDED: Ser17 phosphorylation induced by hypoxia/FCCP but NOT nutrient starvation
6. DRP1 interaction domain specified: cytosolic loop AA 96–138 (Wu 2016 EMBO J)
7. Lampert 2019 mechanism corrected: siRNA knockdown (not KO); PRKN-independent noted; mtDNA mutation finding added
8. “tail-anchored” descriptor removed from intro (incorrect — 3 TM segments, not a tail anchor)
9. All 9 footnotes rewritten with accurate model, quantitative, and access-status details
10. BUG-4 filed in a local paper archive/FEATURE_REQUESTS.md for Wu 2016 Autophagy corrupted download

Downstream pages to check (main agent):

• processes/mitophagy.md — may cite LIR sequence or the erroneous ~10-fold claim
• studies/liu-2012-fundc1-mitophagy-receptor.md — LIR sequence may need correction if it exists

[2026-05-04] verify | spermidine

Pages verified: 1 (molecules/compounds/spermidine.md) — partial scope; see verified-scope in frontmatter

Sources checked:

• 10.1038/nm.4222 (Eisenberg 2016) — local PDF verified end-to-end
• 10.1038/cdd.2014.215 (Pietrocola 2015) — local PDF verified end-to-end
• 10.1093/ajcn/nqy102 (Kiechl 2018) — abstract only (PDF failed); partial
• 10.1038/ncb1975 (Eisenberg 2009) — abstract only (PDF failed); partial; no-fulltext-access
• 10.1126/science.aan2788 (Madeo 2018) — not_oa; permanently unverifiable; no-fulltext-access

Corrections: (1) Eisenberg 2016 strain C57BL/6→C57BL/6J; (2) added exact n=40/41 and n=91/86, p=0.018/0.012; (3) cardioprotection endpoints specified (LV EDP, stiffness β, LV mass/TL, VVC); (4) IL-6 removed from inflammation claim (only TNF-α confirmed); (5) mechanism reframed HAT-broad→EP300-specific per Pietrocola 2015 conclusions; (6) histone site H3K14ac→H3K56 per Pietrocola Fig.6; (7) Pietrocola cell models: removed erroneous C57BL/6 mice; (8) Kiechl n ~800→829, HR clarified to per-1-SD=0.74(0.66–0.83,p<0.001), unverified tertile HR estimate removed, SAPHIR validation added; (9) frontmatter mechanisms: hat-broad-inhibition→ep300-inhibitor.

Downstream propagation needed (for main agent):

• processes/autophagy.md — HAT-broad mechanism framing may need alignment with EP300-specific correction
• Any study pages for Eisenberg 2016/Pietrocola 2015 if they exist

[2026-05-04] verify | lipophagy

Pages verified: 1 (partial — Settembre 2013 PDF unavailable; LIPA/CLEAR claim tagged no-fulltext-access)

• processes/lipophagy.md — corrections made; verified: true (partial scope)

Sources checked:

• 10.1038/nature07976 (Singh 2009, Nature) — local PDF. VERIFIED end-to-end.
• 10.1016/j.cmet.2011.04.004 (Singh & Cuervo 2011, Cell Metab) — downloaded and verified (review paper).
• 10.1038/ncb3166 (Kaushik & Cuervo 2015, Nat Cell Biol) — downloaded and verified.
• 10.1080/15548627.2015.1124226 (Kaushik & Cuervo 2016, Autophagy) — downloaded and verified.
• 10.1038/ncb2718 (Settembre 2013, Nat Cell Biol) — PARTIAL. PDF download failed (status: failed despite green OA at PMC3699877). Key quantitative TFEB/HFD claims partially verified via PMC full-text fetch; LIPA/CLEAR binding claim NOT independently confirmed.

Corrections made to processes/lipophagy.md:

1. Leupeptin mis-attribution: Discovery section said “(Atg5-knockout) or pharmacologically (3-methyladenine, leupeptin)” treating leupeptin as an autophagy inhibitor. Corrected: leupeptin is a lysosomal protease inhibitor used to trap cargo, not an autophagy inhibitor. 3MA is the pharmacological autophagy inhibitor; siAtg5 and Atg5-KO MEFs are the genetic tools.
2. In-vitro model specification: Added that the in-vitro model is the rat hepatocyte line RALA255-10G (not generic “hepatocytes”). Atg5-KO used MEFs, not hepatocytes.
3. Atg7-KO age: Added “examined at four months of age” per the paper (previously unspecified).
4. Term correction: “named the process ‘lipophagy’” → “coined the term ‘macrolipophagy’” (the paper’s precise coinage).
5. PLIN2 CMA-mutant description: Changed “Hsc70/LAMP2A-binding site mutant” → precise notation (414SLKVQ→414SLKAA CMA-targeting-motif mutant preventing Hsc70 binding) per Kaushik 2015.
6. AMPK/PLIN2 phosphorylation causal order CORRECTED (critical): Wiki stated “AMPK phosphorylation of PLIN2 at Ser492/Ser494 is required for PLIN2 recognition by Hsc70.” Both the causal order and the site specification are wrong. Per Kaushik 2016: Hsc70 binds first, THEN AMPK phosphorylates PLIN2; the specific residues (Ser492/Ser494) are NOT stated in either Kaushik 2015 or 2016 — tagged unsourced.
7. Metabolic regulation table: PLIN2 Ser492/Ser494 removed from AMPK row; replaced with “(CMA priming; specific residues not identified in primary sources) unsourced”.
8. Settembre 2013 TFEB section: Added actual quantitative results from PMC fetch (40% kcal fat, 10–12 wks, n=10/group, p<0.001 hepatic lipid, p<0.0001 serum TAG); added no-fulltext-access and verification caveat on LIPA/CLEAR claim; ChIP-qPCR evidence for autoregulatory loop explicitly noted.
9. Pharmacological table: “Compound C” removed from TFEB activators list — Compound C is an AMPK inhibitor (per Kaushik 2016), not a TFEB activator. Corrected with appropriate agents.
10. All footnotes updated: removed “pending download” status from all four downloaded sources; added accurate model descriptions, specific quantitative findings, and precise cell/animal model details.
11. Banner removed; verified flag flipped.

Pages unverifiable (download failed):

• 10.1038/ncb2718 (Settembre 2013) — PDF download failed (download_status: failed). Tagged no-fulltext-access on LIPA/CLEAR claim and quantitative LD data. Partially checked via PMC full-text fetch.

Downstream pages that may need updates (for main agent — not propagated here):

• molecules/proteins/plin2.md (if it exists) — check if Ser492/Ser494 phosphorylation sites are listed; correct/remove if so (unsourced in primary literature)
• molecules/proteins/plin3.md (if it exists) — check PLIN3 CMA-substrate claim; consistent with Kaushik 2015
• processes/chaperone-mediated-autophagy.md (if it exists) — should reflect that CMA is upstream of both cytosolic lipolysis AND macrolipophagy per Kaushik 2015
• molecules/proteins/tfeb.md (if it exists) — LIPA/CLEAR claim needs PDF verification of Settembre 2013
• studies/singh-2009-autophagy-lipid-metabolism.md — if exists, update with precise in-vitro model (RALA255-10G), KO age (4 months), and term “macrolipophagy”

---

[2026-05-04] verify | bnip3

Pages verified: 1 (partial — Bruick 2000 not_oa; HIF-1α claims unverified against full text)

• molecules/proteins/bnip3.md — 9 corrections; verified: true with scope

Sources checked (6 of 7):

• 10.1084/jem.186.12.1975 (Chen 1997, JEM) — downloaded and verified end-to-end.
• 10.1074/jbc.m111.322933 (Hanna 2012, JBC) — downloaded and verified end-to-end.
• 10.4161/auto.6.7.13005 (Quinsay 2010, Autophagy) — downloaded and verified end-to-end.
• 10.1073/pnas.0708818104 (Schweers 2007, PNAS) — downloaded and verified end-to-end.
• 10.1038/nature07006 (Sandoval 2008, Nature) — local PDF, verified end-to-end.
• 10.1038/cdd.2010.146 (Rikka 2011, Cell Death Differ) — downloaded and verified end-to-end.
• 10.1073/pnas.97.16.9082 (Bruick 2000, PNAS) — not_oa. Tagged no-fulltext-access on footnote.

Corrections made to molecules/proteins/bnip3.md:

1. LIR motif sequence CORRECTED (critical): W18-V-V-L21 → W18-V-E-L21 (WVEL). Hanna 2012 Fig 2A shows hBnip3 “6 - APGMQEESLQGSWVELHFS - 24” with WVEL bolded; W18A mutation abrogates LC3 binding.
2. BNIP3/GABARAP interaction CORRECTED: BNIP3 binds LC3 but NOT GABARAP per Hanna 2012 (NIX shows opposite preference). GABARAPL1/2 removed from interactors list.
3. Chen 1997 model system CORRECTED: “COS-7/transfection” → “Rat-1 fibroblasts + MCF-7 breast carcinoma cells”.
4. Hanna 2012 model system CORRECTED: “cardiomyocytes + HEK293” → “HeLa cells + adult rat hearts”.
5. Quinsay 2010 model system CORRECTED: “neonatal rat cardiomyocytes” → “adult rat cardiac myocytes (Sprague Dawley, 200–250g males)”. Ca²⁺/ROS independence added.
6. Schweers 2007 quantitative claims CORRECTED and EXPANDED: 40% from Fig. 2G (all erythrocytes, MitoTracker Red); 19% vs 3% from Fig. 2B (reticulocytes, Thiazole Orange). BAX/BAK/BCL-XL/BIM/PUMA independence noted.
7. Rikka 2011 model system CORRECTED: “cardiomyocytes + Bnip3 transgenic mice” → “HL-1 myocytes + Bax/Bak DKO MEFs, adenoviral overexpression”. 59% ± 2 respiration reduction (p<0.01) added.
8. Boyd 1994 DOI RESOLVED: Correct DOI confirmed as 10.1016/0092-8674(94)90202-x (Cell 79:341–351). Wrong DOI 10.1038/371375a0 is Nossal 1994.
9. Sandoval 2008 footnote EXPANDED: Quantitative data added (28%/12% Mito+CD71+ at 3/6 wks; n=3); C57BL/6 background confirmed; ΔΨm mechanism focus noted.

Downstream pages that may need updates (for main agent — not propagated here):

• studies/chen-1997-nip3-proapoptotic.md — model Rat-1/MCF-7 not COS-7
• studies/hanna-2012-bnip3-lc3-mitophagy.md — model HeLa+rat hearts; LIR is WVEL; BNIP3 does not bind GABARAP
• studies/quinsay-2010-bnip3-mitophagy-mptp.md — adult (not neonatal) rat cardiomyocytes; Ca²⁺/ROS independence
• studies/schweers-2007-nix-erythrocyte-mitophagy.md — both quantitative measurements (Fig 2B and Fig 2G)
• studies/rikka-2011-bnip3-bioenergetics-turnover.md — HL-1 + Bax/Bak DKO MEFs (not transgenic mice); 59% respiration reduction
• processes/mitophagy.md — if uses WVVL for BNIP3 LIR, correct to WVEL; update GABARAP statement
• molecules/proteins/nix.md or bnip3l.md — NIX prefers GABARAP over LC3 (opposite of BNIP3); erythrocyte quantitative data

---

[2026-05-04] verify | atg5

Pages verified: 1 (partial — Mizushima 1998 not_oa; domain architecture source unverifiable)

• molecules/proteins/atg5.md — corrections made; verified: true with scope

Sources checked:

• 10.1038/nature03029 (Kuma 2004, Nature) — local PDF. VERIFIED end-to-end.
• 10.1038/nature04724 (Hara 2006, Nature) — local PDF. VERIFIED end-to-end. CONFIRMED: this is an Atg5-flox × nestin-Cre paper, NOT Atg7.
• 10.1038/ncomms3300 (Pyo 2013, Nat Commun) — local PDF. VERIFIED end-to-end.
• 10.1083/jcb.152.4.657 (Mizushima 2001, J Cell Biol) — downloaded and verified.
• 10.1038/26506 (Mizushima 1998, Nature) — not_oa. Tagged no-fulltext-access.

Corrections made to molecules/proteins/atg5.md:

1. Hara 2006 footnote model: “neuron-specific Atg7-null mice” → “neuron-specific Atg5-null mice (Atg5^flox/flox; nestin-Cre)” — critical misattribution corrected; Hara 2006 IS the ATG5 paper.
2. Hara 2006 study page link: hara-2006-neuronal-atg7-ko-neurodegeneration → hara-2006-neuronal-atg5-ko-neurodegeneration.
3. Conditional KO body text: removed erroneous description of Hara 2006 as “Nestin-Cre × Atg7-flox”; replaced with accurate phenotype description from PDF (Purkinje cell loss, axonal swelling, ubiquitin-positive NeuN+ inclusions, progressive motor deficits from 3 weeks).
4. KO neonatal lethality: “universally die within 24 hours” → 51/52 Atg5−/− died within 1 day; exact caesarean survival times added (WT/het 20.6±3.2 h vs KO 12.4±1.3 h, P<0.01); one survivor lived to day 9.
5. KO biochemistry: expanded to include specific amino acid types (essential AAs, BCAAs), AMPK activation, ECG ST-segment elevation, and milk-rescue result — all from Kuma 2004.
6. Mendelian ratios added for Atg5 KO: +/+:+/−:−/− = 150:371:147 (from Kuma 2004 Fig. 2 text).
7. Domain structure section: Mizushima 2001 citation removed as support for crystal-structure-derived domain architecture — that paper does not describe crystal structure; clarified that UblA/HBR/UblB are from later structural work.
8. K130R functional data: correctly attributed to Mizushima 2001 (K130R mutant cannot be conjugated, fails elongation).
9. ATG12–ATG5 age-decline claim: Mizushima 2001 misattribution corrected; flagged unsourced — Mizushima 2001 studies ES cells, not aged tissues. Primary source needed.
10. Pyo 2013 “no increase in cancer incidence” claim removed — paper does not report formal cancer incidence as an endpoint; replaced with accurate necropsy description.
11. Pyo 2013 footnote: added max lifespan data (WT 781±22d vs Tg 900±34d), leanness quantification (−12% body weight after 12 months), confirmed log-rank test.
12. Mizushima 2001 footnote: completely rewritten to accurately describe what the paper shows (phagophore localization, K130R, LC3 dependence on Atg12-Apg5, protein degradation quantification); PDF path added.
13. Kuma 2004 footnote: “universally die within 24 h” → precise statistics; amino acid/energy phenotype detail added.
14. ⚠️ banner removed; verified flag flipped.

Downstream pages to check:

• studies/hara-2006-neuronal-atg7-ko-neurodegeneration — study page may need to be renamed/corrected to reflect ATG5 (not ATG7) genotype; likely created from same misattribution
• molecules/proteins/atg7.md — may contain erroneous Hara 2006 attribution to ATG7; check footnotes for hara-2006 wikilink
• molecules/proteins/lc3.md and molecules/proteins/ulk1.md — may cite Hara 2006 under ATG7; check per concurrent verifier report that these pages were already corrected

[2026-05-04] verify | pink1-parkin-pathway

Pages verified: 1 (partial — Lazarou 2015 and Koyano 2014 unverifiable; not_oa)

• pathways/pink1-parkin-pathway.md — corrections made; verified: true (partial scope)

Sources checked:

• 10.1083/jcb.200809125 (Narendra 2008, J Cell Biol) — local PDF. VERIFIED. Parkin recruitment topology, ATG5-dependency confirmed. Cell models: HEK293, HeLa (YFP-Parkin), Mfn1−/−Mfn2−/− MEFs, rat cortical neurons.
• 10.15252/embj.2020104705 (Onishi 2021, EMBO J) — local PDF. VERIFIED. PINK1 domain map (Fig 3A), Box 1 Parkin activation mechanism, cargo receptor diagram (Fig 3C), autophosphorylation at Ser228 + Ser402 confirmed.
• 10.15252/embj.201592237 (Sauvé/Wauer 2015, EMBO J) — local PDF. VERIFIED. pUb binding site confirmed: RING1 His302/Arg305 (60-fold affinity loss with H302A). Ubl Ser65 phosphorylation releases Ubl from RING1. Competitive antagonism between pUb and Ubl for RING1 confirmed by ITC + NMR. Domain map confirmed from Fig 1A: Ubl 1–76, linker 77–140, RING0 141–225, RING1 226–327, IBR 327–378, REP 378–410, RING2 410–465.
• 10.1073/pnas.0602493103 (Yang et al. 2006, PNAS) — local PDF. VERIFIED. Author attribution was wrong (wiki said “Clark et al.” — paper is Yang, Gehrke, Imai, Huang, Ouyang, Wang, Yang, Beal, Vogel, Lu 2006). Model is dPink1 RNAi (not KO). hParkin OE rescues muscle + PPL1 dopaminergic phenotypes; DJ-1 OE does not. dParkin protein reduced in dPink1 RNAi animals.
• 10.1038/nature14893 (Lazarou 2015, Nature) — UNVERIFIABLE. not_oa. NDP52/OPTN primary receptor claims and PINK1-only path claims retain no-fulltext-access tag in footnote.
• 10.1038/nature13392 (Koyano 2014, Nature) — UNVERIFIABLE. not_oa. pUb Ser65 claim retained; footnote has no-fulltext-access.
• KEGG hsa04137 — confirmed via API. Pathway name: “Mitophagy - animal - Homo sapiens.”
• Reactome R-HSA-5205685 — confirmed via API. Display name: “PINK1-PRKN Mediated Mitophagy Pathway.”

Corrections made to pathways/pink1-parkin-pathway.md:

1. PARL step: Added contradictory-evidence noting Narendra 2010 finding that PARL is dispensable in mammalian cells (PARL KD/KO does not reduce PINK1 cleavage); propagated from pink1.md verify pass.
2. “Clark et al. 2006” → “Yang et al. 2006”: Author attribution corrected throughout; footnote key renamed from [^clark2006] to [^yang2006]; footnote rewritten with accurate description (RNAi knockdown model, not KO; hParkin rescues; DJ-1 does not; PPL1 cluster specificity noted).
3. pUb binding site description: Corrected Step 4 to explicitly state pUb binds RING1 at His302/Arg305 (not generically “RING1”); mechanism of Ubl-RING1 competition via allostery added; both events shown to be synergistic, not redundant (Wauer 2015).
4. Wauer 2015 footnote: Removed “pending download — partially unverified” status; updated with accurate full author list (Sauvé, Lilov, Seirafi et al.), method details (Delta86-130 crystal at 2.54 Å, ITC, NMR, SAXS), His302/Arg305 affinity data (60-fold reduction with H302A).
5. Overview section PARL description: “cleaved by the PARL protease” → “cleaved by a rhomboid protease (canonically attributed to PARL, though disputed — see Step 2)” to match corrected body text.
6. Banner removed; verified flag flipped.

Pages unverifiable (closed-access):

• 10.1038/nature14893 (Lazarou 2015) — not_oa; no-fulltext-access in footnote
• 10.1038/nature13392 (Koyano 2014) — not_oa; no-fulltext-access in footnote

Downstream pages that may need updates (for main agent — not propagated here):

• processes/mitophagy.md — check if it carries “Clark 2006” attribution for dPink1 epistasis; correct to Yang 2006 if so
• processes/mitophagy.md — check if PARL contradictory-evidence has been added (flagged by pink1.md verifier; should now be in sync)
• molecules/proteins/pink1.md — already updated with PARL contradiction per batch-1 verify pass; no further action needed
• molecules/proteins/parkin.md — already updated with RING1 His302/Arg305 pUb binding site per batch-1 verify pass; no further action needed

---

[2026-05-04] verify | p62

Pages verified: 1 (partial — multiple sources unverifiable)

• molecules/proteins/p62.md — corrections made; verified: true with scope

Sources checked:

• 10.1038/ncb2021 (Komatsu 2010, Nat Cell Biol) — local PDF. VERIFIED end-to-end.
• 10.1083/jcb.200507002 (Bjørkøy 2005, J Cell Biol) — downloaded and verified.
• 10.1016/j.celrep.2019.06.070 (Aparicio 2019, Cell Rep) — downloaded and verified.
• 10.1016/j.cell.2007.10.035 (Komatsu 2007, Cell) — CORRUPTED PDF (downloaded a Cell Press media kit). Claims retained with warning; BUG-3 filed in FEATURE_REQUESTS.md.
• 10.1074/jbc.M702824200 (Pankiv 2007, JBC) — download failed. Tagged no-fulltext-access.
• 10.4161/auto.2.2.2405 (Bjørkøy/Lamark 2006 review, Autophagy) — download failed. Tagged no-fulltext-access.
• 10.1042/ebc20170035 (Lamark 2017, Essays Biochem) — not_oa. Tagged no-fulltext-access.

Corrections made to molecules/proteins/p62.md:

1. KIR domain residues: “~348–352” → “346–359” (Komatsu 2010 crystal structure confirmed; V348–L355 electron density; residues 345–359 sufficient and essential by deletion mapping)
2. Ser349/mTORC1 attribution removed: Komatsu 2010 shows D349A structurally impairs KEAP1 binding but does NOT attribute Ser349 phosphorylation to mTORC1 — that kinase attribution is from Ichimura 2013 (Mol Cell 56:793–805). Gap tagged.
3. Bjørkøy 2005 cell lines in footnote: “HEK293, MCF7” → “HeLa, NIH3T3, SHSY-5Y” (actual paper cell lines)
4. Bjørkøy 2005 body text: removed unsupported “PB1-deficient p62 cannot rescue autophagy flux in p62-null cells” claim; replaced with accurate PB1+UBA body formation description
5. Aparicio 2019 sex specificity: females only (males: n.s. per Figure S1E). Added.
6. Aparicio 2019 mechanism: corrected “proteostasis and reduced oxidative damage” → mitochondrial fission/mitophagy as the required mechanism (Drp1K38A and parkin RNAi abolish longevity benefit)
7. Komatsu 2010 footnote: added Kd (ITC), Cre driver, stoichiometry, Ser349/mTORC1 clarification
8. Aparicio 2019 footnote: added sex, n, mechanistic detail
9. Komatsu 2007, Pankiv 2007, Bjørkøy 2006 review, Lamark 2017 footnotes: flagged with access blockers
10. ⚠️ banner removed; verified flag flipped

BUG-3 filed in (corrupted Komatsu 2007 Cell PDF)

Downstream pages for main agent to check:

• processes/autophagy.md — KIR residue and Ser349/mTORC1 claim
• processes/mitophagy.md — Komatsu 2007 Cell double-KO claims
• hallmarks/disabled-macroautophagy.md — p62 flux readout caveat
• molecules/proteins/lc3.md — Pankiv 2007 LIR claims (unverified source)

---

[2026-05-04] verify | pink1

Pages verified: 1 (partial — Valente 2004 unverifiable)

• molecules/proteins/pink1.md

Sources checked:

• 10.1126/science.1096284 (Valente 2004) — UNVERIFIABLE. DOI lookup failed (figshare OA via figshare returned HTTP 202, no PDF; paper is Science paywalled). Tagged no-fulltext-access.
• 10.1083/jcb.200809125 (Narendra 2008) — local PDF. VERIFIED. Footnote enriched with quantitative data and cell-line details.
• 10.1371/journal.pbio.1000298 (Narendra 2010) — downloaded and verified. VERIFIED. Sufficiency claim confirmed. Key nuance added: PARL KD/KO did not reduce PINK1 cleavage in mammalian cells (contradicts canonical PARL model).
• 10.1083/jcb.201402104 (Kane 2014) — downloaded and verified. VERIFIED. Ub Ser65 phosphorylation confirmed by MS (TLSDYNIQKEpSTLHLVLR, ETD + HCD). UbS65A inhibits Parkin translocation >50%.
• 10.1042/BJ20140334 (Kazlauskaite 2014) — downloaded and verified. VERIFIED. ~14-fold phospho-Ub enrichment by SILAC-MS. Both Ub Ser65 and Parkin UBL Ser65 phosphorylations confirmed required for full Parkin activation.
• 10.15252/embj.2020104705 (Onishi 2021) — local PDF. VERIFIED. Domain map from Fig 3A is primary reference for residue corrections.

Corrections made to molecules/proteins/pink1.md:

1. TM domain start: 94 → 85 (Onishi 2021 Fig 3A; 94 matched neither source)
2. Kinase domain end: 509 → 511 (Onishi 2021 Fig 3A; Narendra 2010 gave 509, minor discrepancy)
3. Domain table: removed unsourced “Insertion loop 1 ~290–315” row; added correct i1/i2/i3 insert description; CTR named per Onishi 2021
4. Parkin recruitment: “REI domain” → “Ubl domain / RING0/UPD” (REI not a term in primary sources)
5. PARL cleavage: added contradictory-evidence; Narendra 2010 finding vs. canonical PARL model documented in body and Limitations
6. Valente 2004 claims: tagged no-fulltext-access throughout
7. Footnotes: all four downloaded papers updated from “pending” to “local PDF available” with enriched quantitative detail
8. Verified flag flipped true (partial scope); banner removed

Pages unverifiable: 10.1126/science.1096284 (Valente 2004) — no-fulltext-access

Downstream pages for main agent to check:

• processes/mitophagy.md — check PARL claim; add contradiction note if needed
• pathways/pink1-parkin-pathway.md — use corrected domain residues when created

---

[2026-05-04] verify | parkin

Pages verified: 1

• molecules/proteins/parkin.md — corrections made (see below); verified: true (partial scope — canonical-DB IDs not live-verified)

Sources checked:

• 10.1083/jcb.200809125 (Narendra 2008, J Cell Biol) — local PDF. VERIFIED. Model systems confirmed: HEK293 cells (endogenous Parkin), HeLa cells (YFP-Parkin), Mfn1−/−Mfn2−/− MEFs, rat cortical neurons. Foundational recruitment study; does not characterise PINK1 mechanism or pUb (those were 2014).
• 10.1038/ncb2012 (Geisler 2010, Nat Cell Biol) — local PDF. VERIFIED. Key findings confirmed: PINK1 kinase activity and MTS required for Parkin translocation; Parkin forms K27- and K63-linked polyUb chains; VDAC1 is a Parkin substrate ubiquitinated via K27-linked chains predominantly; VDAC1 siRNA abrogates both Parkin translocation and mitochondrial clearance; p62 essential for clearance not translocation. Model: HeLa + SH-SY5Y. HK2 is NOT identified in this paper.
• 10.1083/jcb.201007013 (Tanaka 2010, J Cell Biol) — downloaded and verified. Model: SH-SY5Y (endogenous Parkin) + HeLa (YFP-Parkin) + MEFs. MFN1/2 proteasomal degradation via p97/VCP confirmed; Drp1 fission required for Parkin mitophagy; p97 and proteasome activity both required.
• 10.1038/nature12043 (Sarraf 2013, Nature) — downloaded and verified (author manuscript). QdiGLY/SILAC proteomics in HCT116^PARKIN and HeLa^PARKIN. Hundreds of ubiquitylation sites; 36 MOM proteins among 60 Class 1/2 targets. Substrates include MFN1/2, RHOT1/2, VDAC1/2/3, HK1/HK2, TOMM20, TOMM70.
• 10.1073/pnas.1216197110 (Rana 2013, PNAS) — downloaded and verified. Ubiquitous Parkin OE: ~28% increase in both mean AND maximum lifespan in female flies (log-rank, n > 200). Neuronal OE: also extends lifespan. Reduces protein aggregates; reduces dMfn; increases citrate synthase + complex I/II activity. No physiological tradeoffs.
• 10.15252/embj.201592237 (Wauer/Sauvé 2015, EMBO J) — downloaded and verified. Delta86-130 crystal structure at 2.54 Å. pUb binds RING1 at His302/Arg305 (not Ubl-RING0 interface as wiki stated). Ubl phosphorylation releases Ubl from RING1 and increases UbcH7 affinity. Both pUb and pUbl required for maximal activity.
• 10.15252/embj.2020104705 (Onishi 2021, EMBO J) — local PDF. Already verified on mitophagy.md; review context confirmed.

Not_oa (unverifiable):

• 10.1038/33416 (Kitada 1998, Nature) — not_oa. Kitada 1998 claims (PRKN mutations in AR-JP pedigrees) retained unverified.
• 10.1126/science.1237908 (Trempe 2013, Science) — not_oa. Crystal structure claims and autoinhibition contact details retained with attribution, but cannot confirm exact residue contacts against full text.
• 10.1038/nature13392 (Koyano 2014, Nature) — not_oa. Feed-forward pUb amplification loop claims retained; binding site description corrected using Wauer 2015 (open access) which resolved the molecular detail.

Corrections made to molecules/proteins/parkin.md:

1. pUb binding site: “binds the Ubl-RING0 interface” → “binds a site on RING1 formed by His302 and Arg305” (Wauer 2015). Previous description was mechanistically incorrect.
2. Domain boundaries table: RING0 “79–148” → “~141–225”; linker “149–237” removed and replaced with correctly-bounded “disordered linker 77–140”; RING1 “238–293” → “~226–326”; IBR “314–377” → “~327–378”; REP “~383–410” → “~378–410”; RING2 “411–449” → “~410–465” (Wauer 2015 Fig 1A).
3. HK2 substrate attribution: “HK2” in table cited to Geisler 2010 → removed from Geisler; changed to “HK1, HK2” cited to Sarraf 2013 with correct description of 14 diGLY sites.
4. Sarraf 2013 substrate count: “>100 OMM-resident ubiquitinated substrates” → “hundreds of ubiquitylation sites in dozens of proteins; ≥36 confirmed MOM proteins among 60 Class 1/2 targets” (precise per paper).
5. Lede substrate count: “ubiquitinates >100 OMM proteins” → “ubiquitinates dozens of OMM-resident proteins (≥36 confirmed MOM proteins).”
6. Rana 2013 lifespan: “extends median lifespan” → “extends both mean and maximum lifespan (~28% increase in female flies).”
7. Rana 2013 mechanism: added citrate synthase activity and respiratory complex I/II increases (from paper); removed unsupported locomotor claim.
8. VDAC1 table row: added “K27-linked” chain type specificity (Geisler 2010).
9. All footnotes updated to accurate descriptions with correct model systems, methods, and authors; “archive: pending download” status updated to “local PDF available” for Wauer 2015, Sarraf 2013, Rana 2013, Tanaka 2010.

Pages unverifiable (closed-access):

• 10.1038/33416 (Kitada 1998) — not_oa; retained
• 10.1126/science.1237908 (Trempe 2013) — not_oa; retained
• 10.1038/nature13392 (Koyano 2014) — not_oa; retained; pUb binding site corrected using Wauer 2015

Downstream pages that may need updates (for main agent — not propagated here):

• processes/mitophagy.md — if it carries the “pUb binds Ubl-RING0 interface” mechanistic error, correct to RING1 His302/Arg305 (Wauer 2015)
• pathways/pink1-parkin-pathway.md — same mechanistic correction for pUb binding site may be needed
• Any entity page citing “Geisler 2010” as source for HK2 — HK2 is a Sarraf 2013 finding

[2026-05-04] verify | lc3

Pages verified: 1

• molecules/proteins/lc3.md — corrections made (see below)

Sources checked:

• 10.1038/35044114 (Ichimura 2000, Nature) — local PDF. VERIFIED. Paper establishes E1 (Apg7) and E2 (Apg3) for yeast Apg8-PE conjugation; does NOT demonstrate the E3-like role of ATG12–ATG5–ATG16L1 for mammalian LC3-PE (that came in 2007–2008 papers by Hanada/Fujita).
• 10.1093/emboj/19.21.5720 (Kabeya 2000, EMBO J) — downloaded via PMC. VERIFIED. Observed MW: LC3-I ~18 kDa, LC3-II ~16 kDa (rat LC3 in brain and HeLa cells). Cleavage at Gly120 confirmed by LC3G120A mutant analysis. LC3-II present on both inner and outer autophagosomal membrane leaflets confirmed by protease protection and immunogold EM.
• 10.1038/nature04724 (Hara 2006, Nature) — local PDF. VERIFIED. Critical error corrected: Hara 2006 used Atg5 (nestin-Cre), not Atg7 (CaMKII-Cre) as the wiki stated.
• 10.1038/nature04723 (Komatsu 2006, Nature) — local PDF. VERIFIED. Atg7 / nestin-Cre correct. All Atg7flox/flox; nestin-Cre mice dead within 28 weeks, n=26 mutant / 41 control, P<0.01.
• 10.1080/15548627.2020.1797280 (Klionsky 2021, Autophagy) — downloaded. VERIFIED. Guidelines recommend comparing LC3-II to housekeeping proteins (not to LC3-I); LC3-II/LC3-I ratio is discouraged. LC3 subfamily per Klionsky 2021: LC3A (two splice variants), LC3B, LC3B2, LC3C (four members, not three).
• 10.1074/jbc.M702824200 (Pankiv 2007) — download failed (no OA URL). UNVERIFIABLE. p62 LIR/WSTL claims retained unverified.
• 10.4161/auto.4600 (Mizushima & Yoshimori 2007) — download failed (no OA URL). UNVERIFIABLE. LC3-I/LC3-II migration and flux-interpretation claims retained unverified (consistent with Klionsky 2021 which covers same ground).

Corrections made to molecules/proteins/lc3.md:

1. MW apparent values: “LC3-I ~16 kDa, LC3-II ~14 kDa” → “LC3-I ~16–18 kDa, LC3-II ~14–16 kDa” (Kabeya 2000 rat data; Klionsky 2021 range)
2. E3 attribution: removed [^ichimura2000] from ATG12–ATG5–ATG16L1 E3 step; added note that Ichimura 2000 only established E1/E2 in yeast; E3 mammalian evidence is from later work (Hanada 2007, Fujita 2008); tagged unsourced for E3 citation
3. Hara 2006 footnote: “Atg7 neural conditional KO mice (CaMKII-Cre)” → “Atg5 neural conditional KO mice (nestin-Cre)”; study page slug updated from hara-2006-atg7-neural-neurodegeneration to hara-2006-atg5-neural-neurodegeneration
4. Body text (two locations + extrapolation table): corrected “Atg7” to “Atg5” for Hara 2006 model; clarified distinct models for Hara vs Komatsu
5. LC3A paralog table note: “Three splice isoforms (LC3A-I, -II, -III)” → “Two splice variants”; clarified that -I/-II/-III notation refers to processing state, not splice variant count
6. Klionsky 2021 footnote: updated description to accurately reflect the paper’s recommendation (HKP normalization, not “absolute LC3-II preferred over ratio”)
7. LC3-II/LC3-I ratio body text: updated to reflect Klionsky 2021’s actual recommendation (normalize to HKP, with caveats)

Pages unverifiable (no OA PDF):

• 10.1074/jbc.M702824200 (Pankiv 2007) — tagged in footnote as pending download; p62 LIR/WSTL claim unverified
• 10.4161/auto.4600 (Mizushima & Yoshimori 2007) — tagged in footnote; migration behavior and flux interpretation unverified against PDF

Downstream pages that may need updates (for main agent — not propagated here):

• studies/hara-2006-atg7-neural-neurodegeneration — if this study page exists, it needs renaming to hara-2006-atg5-neural-neurodegeneration and complete correction of the described model (consistent with correction already noted from atg7.md verification on 2026-05-04)
• molecules/proteins/atg5.md — may cite Hara 2006 with wrong gene attribution; check
• Any other entity page with [^hara2006] footnote citing “Atg7/CaMKII-Cre” — systematic grep needed

---

[2026-05-04] verify | atg7

Pages verified: 1

• molecules/proteins/atg7.md — corrections made (see below)

Sources checked:

• 10.1038/nature04724 (Hara 2006) — local PDF. VERIFIED. Critical error corrected: Hara 2006 is an Atg5 paper (Atg5^flox/flox; nestin-Cre), not Atg7. The wiki incorrectly attributed Nestin-Cre × Atg7-flox to this paper.
• 10.1038/nature04723 (Komatsu 2006) — local PDF. VERIFIED. Critical error corrected: wiki said “CamKII-α-Cre × Atg7-flox”; actual design is Nestin-Cre × Atg7^flox/flox. Added quantitative data: n=26 mutant/41 control, dead within 28 weeks, P<0.01 log-rank.
• 10.1038/ncomms3300 (Pyo 2013) — local PDF. VERIFIED. Lifespan claim 17.2% correct; χ²=17.32, P<0.001; n=70 WT/65 Tg (line 25). “By analogy” framing accurate.
• 10.1091/mbc.10.5.1367 (Tanida 1999) — downloaded and verified. Yeast Apg7p/Cvt2p; active-site Cys507 (yeast) confirmed; human Cys572 distinction maintained correctly on wiki page.
• 10.1083/jcb.200412022 (Komatsu 2005) — downloaded and verified. Atg7^-/- neonatal lethality within 1 day confirmed (n=19 neonates in caesarean study); Cre drivers were Zp3-Cre (germline) + Mx1-Cre (liver-specific). Footnote updated with n and driver details.
• 10.1016/j.cmet.2009.10.008 (Masiero 2009) — downloaded and verified. Cre driver corrected: wiki said “HSA-Cre”; actual driver is MLC1f-Cre (myosin light chain 1 fast promoter). Added quantitative CSA data: ~40% reduction, n=5/group, p<0.001.
• 10.1016/j.cell.2011.10.026 (Mizushima 2011) — downloaded. Review paper; ATG7 decline claim consistent with review text.
• 10.1038/nsmb.2067 (Noda 2011) — not_oa; cannot verify. Tagged no-fulltext-access in footnote.

Corrections made to molecules/proteins/atg7.md:

1. Hara 2006 attribution: “Nestin-Cre × Atg7-flox” → corrected to Atg5^flox/flox; nestin-Cre (ATG5 paper, not ATG7)
2. Komatsu 2006 Cre driver: “CamKII-α-Cre” → corrected to “Nestin-Cre”; added n=26/41, 28-week survival endpoint
3. Section heading: “CamKII-Cre; Nestin-Cre” → “Nestin-Cre”; body rewritten to clearly distinguish ATG5 (Hara) vs ATG7 (Komatsu) papers
4. Masiero 2009 Cre driver: “HSA-Cre” → “MLC1f-Cre (myosin light chain 1 fast promoter)”; added CSA quantification (~40%, n=5, p<0.001)
5. All footnotes: updated Cre drivers, added quantitative n’s, updated local PDF paths (all 4 pending papers now downloaded)
6. Pyo 2013 footnote: added n=70 WT/65 Tg, χ²=17.32, P<0.001, log-rank details
7. Noda 2011 footnote: added no-fulltext-access tag
8. Verified flag flipped to true; banner removed

Pages unverifiable (closed-access):

• 10.1038/nsmb.2067 (Noda 2011) — tagged no-fulltext-access; FAP motif claim remains unverified against PDF

Downstream pages that may need updates (for main agent — not propagated here):

• molecules/proteins/atg5.md — log entry describes Hara 2006 as “neuron-specific Atg7 KO”; needs correction to “Atg5 conditional KO”
• Study page slug studies/hara-2006-neuronal-atg7-ko-neurodegeneration — title/slug is wrong; paper is about Atg5, not Atg7; rename to studies/hara-2006-neuronal-atg5-ko-neurodegeneration
• studies/komatsu-2006-neuronal-atg7-ko-p62 — if this page exists, verify Cre driver is documented as nestin-Cre, not CamKII-Cre

---

[2026-05-04] ingest | round-6d-spermidine-atg-conjugation

atg12

• added: molecules/proteins/atg12.md
• entity: protein (ATG12 / autophagy related 12 / APG12); UniProt O94817; NCBI Gene 9140; HGNC 588
• canonical sources pulled: UniProt O94817 (140 aa, Gly140 C-terminal conjugation site, disordered N-terminal region, Ubl fold residues 55-140, interactors ATG5/ATG7/ATG10/ATG16L1/ATG3/TECPR1); accessed 2026-05-04
• HGNC conflict caught and resolved: briefing said hgnc: 589 for ATG12, but HGNC:589 = ATG5, HGNC:588 = ATG12. Corrected in frontmatter. Escalated in schema-gaps section of return summary.
• GenAge: ATG12 not listed (no genage-id)
• primary sources cited (footnotes):
  • 10.1038/26506 (Mizushima 1998, Nature — yeast ATG12-ATG5 conjugation system; 1644 citations; closed OA; not downloaded)
  • 10.1074/jbc.273.51.33889 (Mizushima 1998, J Biol Chem — human ATG12 conjugation system; Gly140-Lys130 isopeptide; 489 citations; hybrid OA; pending)
  • 10.1091/mbc.e07-12-1257 (Fujita 2008, Mol Biol Cell — ATG16L complex specifies LC3 lipidation site; 1050 citations; green OA; pending)
  • 10.1074/jbc.C700195200 (Hanada 2007, J Biol Chem — ATG12-ATG5 E3-like activity; 1081 citations; hybrid OA; pending)
  • 10.1016/j.molcel.2011.10.014 (Rubinstein 2011, Mol Cell — free ATG12 binds Bcl-2/Mcl-1, promotes apoptosis; 295 citations; bronze OA; pending)
  • 10.1038/nature03029 (Kuma 2004, Nature — Atg5 KO neonatal lethality; used as mechanistic analogy for Atg12 KO; 2808 citations; local PDF)
  • 10.1371/journal.pone.0023367 (Kang 2011, PLoS ONE — ATG12 depletion induces premature senescence in human fibroblasts; 247 citations; gold OA; pending)
  • 10.1038/ncomms3300 (Pyo 2013, Nat Commun — ATG5 OE lifespan extension, requires ATG12-ATG5 conjugate; 689 citations; local PDF)
• BUG-2 caught: briefing DOI 10.1016/j.molcel.2011.05.030 for Rubinstein 2011 resolves to an unrelated transcription paper (DOI lookup confirmed). Correct DOI 10.1016/j.molcel.2011.10.014 confirmed via PMID 22152474. Noted in page banner and footnote.
• implicit stubs created (new wikilinks to non-existent pages): atg10 (already noted on atg5), atg16l1 (already noted on atg5), spermidine (already noted on atg5)
• gaps surfaced:
  • Atg12 germline KO neonatal lethality: widely cited claim but no dedicated primary Atg12 KO paper found in citation search — needs-replication; claim rests on mechanistic equivalence with Atg5-null (Kuma 2004)
  • Rubinstein 2011 free-ATG12 pro-apoptotic moonlighting: single-study in-vitro; in-vivo relevance not demonstrated — needs-replication
  • ATG12-ATG5 conjugate decline with aging in human tissues: small cross-sectional studies only — needs-replication
  • Free ATG12 accumulation with age → apoptotic shift hypothesis: speculative, no primary data — no-mechanism unsourced
  • ATG12 OE or augmentation in human aging context: no interventional data — needs-human-replication
• ROADMAP.md updated: atg12 marked [x] drafted 2026-05-04

atg5

• added: molecules/proteins/atg5.md
• entity: protein (ATG5 / autophagy related 5 / APG5); UniProt Q9H1Y0; NCBI Gene 9474; HGNC 589
• canonical sources pulled: UniProt Q9H1Y0 (275 aa, K130 conjugation site, E122 critical residue, UblA/HBR/UblB domain structure, SCAR25 disease association, interactors ATG12/ATG7/ATG10/ATG16L1); accessed 2026-05-04
• primary sources cited (footnotes):
  • 10.1038/26506 (Mizushima 1998, Nature — ATG12-ATG5 isopeptide conjugation system discovery; 1644 citations; closed OA; not downloaded)
  • 10.1083/jcb.152.4.657 (Mizushima 2001, J Cell Biol — ATG5 phagophore localization + ATG16L1 association; 1327 citations; bronze OA; pending download)
  • 10.1038/nature03029 (Kuma 2004, Nature — Atg5 KO neonatal lethality; 2808 citations; local PDF available)
  • 10.1038/nature04724 (Hara 2006, Nature — neuron-specific Atg7 KO neurodegeneration; 3796 citations; local PDF available — used as mechanistic analogy for neuronal autophagy dependency)
  • 10.1038/ncomms3300 (Pyo 2013, Nat Commun — Atg5 OE extends lifespan ~17.2%, n=65+70, χ²=17.32 p<0.001; 689 citations; local PDF available)
• implicit stubs created (new wikilinks to non-existent pages): atg10, atg12, atg16l1, spermidine
• gaps surfaced:
  • Atg5 OE lifespan extension mouse-only — needs-human-replication
  • Age-associated ATG12-ATG5 decline in human tissues — small studies only; needs-replication
  • Cardiomyocyte- and muscle-specific Atg5 conditional KO primary citations unconfirmed — excluded; unsourced
  • Mechanism of ATG5/ATG12-ATG5 age-decline regulation — unknown; no-mechanism
  • Spermidine mechanism in mammalian aging models — needs-replication
• ROADMAP.md updated: atg5 marked [x] drafted 2026-05-04

[2026-05-04] ingest | round-6b-pink1-parkin-p62

p62

• added: molecules/proteins/p62.md
• entity: protein (SQSTM1 / Sequestosome-1 / p62); UniProt Q13501; NCBI Gene 8878; HGNC 11280
• canonical sources pulled: UniProt Q13501 (440 aa, PB1 3-102, ZZ 123-173, LIR 336-341, KIR 348-352, UBA 389-434, key PTMs Ser403/Ser349 phospho, Lys420 acetyl, Cys289 palmitoyl; disease: PDB3/FTDALS3/NADGP/DMRV); accessed 2026-05-04
• primary sources cited (footnotes):
  • 10.1083/jcb.200507002 (Bjorkoy 2005, J Cell Biol — p62 aggregates + LIR-dependent selective autophagy; 3,207 citations; pending download)
  • 10.1074/jbc.M702824200 (Pankiv 2007, J Biol Chem — LIR W338-x-x-L341 + Ser403 phosphorylation; 4,465 citations; pending download)
  • 10.1016/j.cell.2007.10.035 (Komatsu 2007, Cell — Sqstm1-/- x Atg7fl/fl shows p62 required for ubiquitin inclusions; 2,094 citations; pending download)
  • 10.1038/ncb2021 (Komatsu 2010, Nat Cell Biol — KIR/KEAP1/Nrf2 axis; 2,347 citations; LOCAL PDF available)
  • 10.4161/auto.2.2.2405 (Bjorkoy-Lamark 2006, Autophagy review; 322 citations; pending download)
  • 10.1042/ebc20170035 (Lamark 2017, Essays Biochem review; 710 citations; not_oa)
  • 10.1016/j.celrep.2019.06.070 (Aparicio 2019, Cell Rep — Ref(2)P OE Drosophila lifespan; 134 citations; gold OA; pending download)
• Park 1995 DOI excluded — BUG-2 archive mismatch (10.1074/jbc.270.27.16243 resolves to Photosystem I paper); flagged needs-canonical-id on page
• implicit stubs created (new wikilinks to non-existent pages): aggrephagy, xenophagy, pexophagy
• gaps surfaced:
  • Park 1995 original sequestosome DOI unverified (BUG-2) — needs-canonical-id
  • p62 as flux readout: transcription confound requires mRNA controls — needs-replication
  • Age-dependent p62 decline in human tissues — sparse; needs-human-replication
  • p62 OE longevity in mammals — untested; needs-human-replication
  • mTORC1-RAGULATOR-p62 interaction mechanism — debated; no-mechanism
  • Pathogenic vs protective p62 inclusions in ALS/FTD/PD — contested; contradictory-evidence
• ROADMAP.md updated: p62 marked [x] drafted 2026-05-04

pink1

• added: molecules/proteins/pink1.md
• entity: protein (PINK1 / PTEN-induced kinase 1 / PARK6); UniProt Q9BXM7; NCBI Gene 65018; HGNC 14581
• canonical sources pulled: UniProt Q9BXM7 (581 aa, domain structure, key PTMs Ser228/Ser402 autophosphorylation, PARL cleavage at ~Ala103, PARK6 disease association); all 6 DOIs confirmed via
• primary sources cited (footnotes):
  • 10.1126/science.1096284 (Valente 2004, Science — PINK1 mutations cause PARK6 autosomal recessive PD; 3415 citations; green OA; pending download)
  • 10.1083/jcb.200809125 (Narendra 2008, J Cell Biol — Parkin recruits selectively to depolarized mitochondria; 3859 citations; local PDF available)
  • 10.1371/journal.pbio.1000298 (Narendra 2010, PLoS Biol — PINK1 sufficient for Parkin recruitment; selectively stabilized on impaired OMM; 2821 citations; gold OA; pending download)
  • 10.1083/jcb.201402104 (Kane 2014, J Cell Biol — PINK1 phosphorylates ubiquitin Ser65; 1245 citations; bronze OA; pending download)
  • 10.1042/bj20140334 (Kazlauskaite 2014, Biochem J — parallel Ser65-Ub discovery + Parkin UBL Ser65 phosphorylation; 798 citations; hybrid OA; pending download)
  • 10.15252/embj.2020104705 (Onishi 2021, EMBO J — comprehensive mitophagy review; 1238 citations; local PDF available — already cited on mitophagy.md)
• implicit stubs created (new wikilinks to non-existent pages): pink1-parkin-pathway, parkin, optn, ndp52, parkinsons-disease
• gaps surfaced:
  • PINK1 protein age-decline trajectory in human tissues (vs. rodent) — sparse; tagged needs-human-replication
  • Causality of PINK1 decline vs. consequence in aged muscle/brain — tagged no-mechanism
  • PINK1 activators as clinical-stage candidates — none confirmed; tagged needs-replication
  • GenAge entry ID for PINK1 unconfirmed — tagged needs-canonical-id
  • PINK1 exact PARL cleavage residue range — minor uncertainty; tagged needs-replication
  • PINK1/NLRP3 inflammasome link in aged human tissue — tagged unsourced
• ROADMAP.md updated: pink1 marked [x] drafted 2026-05-04

parkin

• added: molecules/proteins/parkin.md
• entity: protein (PRKN / PARK2 / Parkin); UniProt O60260; NCBI Gene 5071; HGNC 8607
• canonical sources pulled: UniProt O60260 (465 aa, O60260 accession, Cys431 active site, 6-domain RBR architecture, AR-JP PARK2 disease association) — UniProt API unavailable during seeding; identity fields not live-verified; flagged needs-canonical-id for lint pass
• primary sources cited (footnotes):
  • 10.1038/33416 (Kitada 1998, Nature — PRKN cloning + AR-JP identification; 5,156 citations; not_oa)
  • 10.1083/jcb.200809125 (Narendra 2008, J Cell Biol — Parkin selective recruitment to depolarized mitochondria; 3,859 citations; local PDF available)
  • 10.1038/ncb2012 (Geisler 2010, Nat Cell Biol — VDAC1 + p62 dependency in Parkin-mediated mitophagy; 2,729 citations; local PDF available)
  • 10.1083/jcb.201007013 (Tanaka 2010, J Cell Biol — Parkin ubiquitinates MFN1/2 → K48 proteasomal degradation; 1,323 citations; bronze OA; pending download)
  • 10.1038/nature12043 (Sarraf 2013, Nature — Parkin-dependent ubiquitylome landscape >100 OMM substrates; 1,000 citations; green OA; pending download)
  • 10.1126/science.1237908 (Trempe 2013, Science — full-length Parkin crystal structure + auto-inhibition mechanism; 519 citations; not_oa)
  • 10.1038/nature13392 (Koyano 2014, Nature — phospho-Ub Ser65 activates Parkin; feed-forward loop; 1,435 citations; not_oa)
  • 10.15252/embj.201592237 (Wauer 2015, EMBO J — two-hit activation mechanism: phospho-Ub + Parkin UbL Ser65 phosphorylation; 211 citations; hybrid OA; pending download)
  • 10.1073/pnas.1216197110 (Rana 2013, PNAS — Parkin OE extends Drosophila lifespan; 328 citations; bronze OA; pending download)
  • 10.15252/embj.2020104705 (Onishi 2021 EMBO J review — background aging context; local PDF available)
• implicit stubs created (new wikilinks to non-existent pages): mfn1, mfn2, miro1, miro2, vdac1, optn, ndp52, pink1-parkin-pathway (also in pink1.md), parkinsons-disease (also in pink1.md)
• gaps surfaced:
  • UniProt identity fields not live-verified (API unavailable) — needs-canonical-id
  • Age-related Parkin protein decline in human post-mitotic tissues — small studies only; needs-replication
  • Parkin OE lifespan extension in mice — untested; needs-human-replication
  • MIRO1/2 ubiquitination mechanism — cell-line data only; in-vivo aged-tissue confirmation lacking; needs-replication
  • Sporadic-PD S-nitrosylation of Parkin in aged human brain — quantitative in-vivo data lacking; no-mechanism
  • Parkin small-molecule activators — preclinical only; no clinical-stage candidates; long-term-unknown
• schema note: Riley 2013 DOI (10.1074/jbc.M112.420661) not found in archive — not cited; Trempe 2013 used instead for Parkin structure
• ROADMAP.md updated: parkin marked [x] drafted 2026-05-04

pink1-parkin-pathway

• added: pathways/pink1-parkin-pathway.md
• entity: pathway (PINK1–Parkin mitophagy pathway); KEGG hsa04137 (Mitophagy - animal); Reactome R-HSA-5205685 (PINK1-PRKN Mediated Mitophagy); WikiPathways null
• canonical sources pulled: KEGG hsa04137 (confirmed via REST API — key nodes PINK1, PRKN, MFN1/2, OPA1, TBK1, ULK1, LC3, BECN1, SQSTM1/p62); Reactome R-HSA-5205685 (confirmed via ContentService API — 6 contained reactions spanning PINK1 recruitment through autophagosome coupling)
• primary sources cited (footnotes):
  • 10.1083/jcb.200809125 (Narendra 2008, J Cell Biol — foundational Parkin selective recruitment; 3859 citations; local PDF available)
  • 10.15252/embj.2020104705 (Onishi 2021, EMBO J — comprehensive mitophagy review; 1238 citations; local PDF available — cross-cited from mitophagy.md verified)
  • 10.1038/nature14893 (Lazarou 2015, Nature — NDP52/OPTN primary receptors; PINK1-only Parkin-independent path; 2620 citations; not_oa — unverified no-fulltext-access)
  • 10.15252/embj.201592237 (Wauer 2015, EMBO J — Parkin Ubl/ubiquitin activation switch; 211 citations; pending download — partially unverified)
  • 10.1038/nature13392 (Koyano 2014, Nature — ubiquitin Ser65 phosphorylation by PINK1; 1435 citations; not_oa — unverified no-fulltext-access)
  • 10.1073/pnas.0602493103 (Clark 2006, PNAS — Drosophila pink1/parkin epistasis; 771 citations; local PDF available)
  • 10.1016/j.xcrm.2022.100633 (Singh 2022, Cell Rep Med — ATLAS trial urolithin A; 182 citations; local PDF available — cross-cited from mitophagy.md verified)
• implicit stubs created (new wikilinks to non-existent pages): ndp52, optn, tbk1, mfn1, mfn2, miro1, vdac1, oxidative-stress, parkinsons-disease, inflammaging, mitosens
• gaps surfaced:
  • PINK1-only (Parkin-independent) path quantitative contribution in aged neurons — no-mechanism
  • PINK1/Parkin flux measurements in post-mitotic human tissues — technically difficult; needs-human-replication
  • Pharmacological PINK1/Parkin activation — no human interventional trial data; long-term-unknown
  • SENS MitoSENS mapping to PINK1/Parkin (organelle clearance vs. mtDNA deletion accumulation) — conceptual gap; flagged for review
• ROADMAP.md updated: pink1-parkin-pathway marked [x] drafted 2026-05-04; Tier A queue entry updated to ✓ seeded R6b

[2026-05-04] ingest | round-6a-core-autophagy

atg7

• added: molecules/proteins/atg7.md
• entity: protein (ATG7); UniProt O95352; NCBI Gene 10533; HGNC 16935
• canonical sources pulled: UniProt O95352 (function, domains, Cys572 active site, SCAR31 disease variants, TRIM32/Lys45 ubiquitination); all 7 DOIs confirmed via
• primary sources cited (footnotes):
  • 10.1091/mbc.10.5.1367 (Tanida 1999, Mol Biol Cell — Apg7p yeast E1-like enzyme; 398 citations; green OA/PMC; pending download)
  • 10.1083/jcb.200412022 (Komatsu 2005, J Cell Biol — Atg7 KO neonatal lethal; 2301 citations; bronze OA; pending download)
  • 10.1038/nature04724 (Hara 2006, Nature — neuronal Atg7 KO neurodegeneration + ubiquitin inclusions; 3796 citations; local PDF available)
  • 10.1038/nature04723 (Komatsu 2006, Nature — parallel neuronal Atg7 KO + p62 accumulation; 3439 citations; local PDF available)
  • 10.1016/j.cmet.2009.10.008 (Masiero 2009, Cell Metabolism — muscle-specific Atg7 KO + sarcopenia-like phenotype; 1242 citations; bronze OA; pending download)
  • 10.1038/ncomms3300 (Pyo 2013, Nat Commun — Atg5 OE extends lifespan ~17.2%; 689 citations; hybrid OA; local PDF available)
  • 10.1016/j.cell.2011.10.026 (Mizushima & Komatsu 2011, Cell — autophagy review; 6154 citations; pending download)
  • doi:10.1038/nsmb.2067 (Noda 2011 — ATG7 NTD/FAP motif structural characterization; bare DOI footnote)
• implicit stubs created (new wikilinks to non-existent pages): atg10, atg3, atg12, atg5, lc3 (Tier B/C stub already queued), bnip3 (Tier C stub), fundc1 (Tier C), sarcopenia (phenotype page)
• gaps surfaced:
  • Cardiac-specific Atg7 KO aging phenotype commonly cited in reviews but primary reference not independently confirmed — excluded; flagged in Limitations section
  • ATG7 protein age-decline in human tissues — small studies only; tagged needs-replication
  • ATG7 OE lifespan extension in mammals — untested; tagged needs-replication + needs-human-replication
  • TRIM32/Lys45 regulatory mechanism in aging — tagged no-mechanism

beclin-1

• added: molecules/proteins/beclin-1.md
• entity: protein (BECN1); UniProt Q14457; NCBI Gene 8678; HGNC 1034
• canonical sources pulled: UniProt Q14457 (function, domains, PTMs, interactors); all 6 DOIs confirmed via
• primary sources cited (footnotes):
  • 10.1038/45257 (Liang 1999, Nature — original cloning + autophagy + tumor suppressor; 3270 citations; not_oa)
  • 10.1016/j.cell.2005.07.002 (Pattingre 2005, Cell — BCL-2 inhibits Beclin-1 autophagy; 3443 citations; pending download)
  • 10.1016/j.molcel.2008.06.001 (Wei 2008, Mol Cell — JNK1/BCL-2 phosphorylation; 1289 citations; pending)
  • 10.1038/s41586-018-0162-7 (Fernández 2018, Nature — F121A lifespan ~10%; 621 citations; green OA/PMC)
  • 10.1073/pnas.2436255100 (Yue 2003, PNAS — Becn1−/− embryonic lethal E7.5; haploinsufficiency; 2070 citations; green OA/PMC)
  • 10.4161/auto.5.5.8625 (Zalckvar 2009, Autophagy — DAPK1 Thr119 phosphorylation; 240 citations; not_oa)
• implicit stubs created (new wikilinks to non-existent pages): vps34, atg14l, uvrag, dapk1, ubiquitin-proteasome-system, disabled-macroautophagy (may already exist)
• gaps surfaced:
  • GenAge entry ID for BECN1 unconfirmed — tagged needs-canonical-id
  • F121A lifespan single-study — tagged needs-replication
  • AMPK direct Beclin-1 phosphorylation sites (Ser90/Ser93) — tagged unsourced
  • Beclin-1 C-fragment pro-apoptotic in-vivo function — tagged no-mechanism
  • Human Beclin-1 age-decline quantitative data sparse — tagged needs-human-replication

[2026-05-04] ingest | round-6a-core-autophagy

tfeb

• added: molecules/proteins/tfeb.md
• entity: TFEB (Transcription Factor EB) — protein
• canonical IDs: UniProt P19484, NCBI Gene 7942, HGNC 11753
• DOIs cited (all confirmed via ):
  • 10.1126/science.1174447 (Sardiello 2009, Science, 2434 citations) — CLEAR network discovery; not_oa
  • 10.1126/science.1204592 (Settembre 2011, Science, 3131 citations) — TFEB master autophagy/lysosomal regulator; OA-pending
  • 10.1038/emboj.2012.32 (Settembre 2012, EMBO J, 1865 citations) — mTORC1 Ser211 phosphorylation / 14-3-3 sequestration; OA-pending
  • 10.1126/scisignal.2002790 (Roczniak-Ferguson 2012, Sci Signaling, 1313 citations) — independent Ser211/14-3-3 confirmation; not_oa
  • 10.1038/ncb3114 (Medina 2015, Nat Cell Biol, 1335 citations) — calcineurin / Ca²+ arm; not_oa
  • 10.1038/ncomms3267 (Lapierre 2013, Nat Commun, 524 citations) — HLH-30 / C. elegans longevity; OA-pending
• gaps surfaced:
  • unsourced — “~500 CLEAR genes” resolved to Sardiello 2009 (471 bioinformatic candidates; 96 confirmed core); Palmieri 2011 (PMID 21358634) flagged for ChIP-seq follow-up
  • needs-human-replication — TFEB overexpression aging rescue is mouse-only
  • no-mechanism — direct AMPK-TFEB phosphorylation unresolved
  • dose-response-unclear — rapamycin vs torin1 TFEB activation window not characterized in vivo
  • needs-canonical-id — GenAge entry for TFEB not confirmed
• implicit stubs created: mtor-complex-1, 14-3-3-proteins, calcineurin, rag-gtpases, tfe3, mitf, tfec, trpml1
• propagation note: autophagy.md #gap/unsourced on CLEAR “~500 genes” is resolved on this page; main agent should update autophagy.md to remove the gap tag and cross-link to tfeb
• verified: false — all 6 DOIs confirmed present in archive (4 not_oa; 2 OA-pending); no local PDFs available; wiki-verifier should download OA copies and cross-check quantitative claims

lc3

• added: molecules/proteins/lc3.md (type: protein, verified: false)
• entity: MAP1LC3B / LC3B — canonical autophagosome membrane marker; ATG8 family member; 125 aa pro-form / 120 aa mature
• canonical DB IDs confirmed:
  • UniProt Q9GZQ8 (MAP1LC3B_HUMAN, Swiss-Prot reviewed) — confirmed via.txt REST endpoint
  • NCBI Gene 81631 — from briefing; not independently live-checked (#gap/needs-canonical-id if discrepant)
  • HGNC 13352 — from briefing; not independently live-checked
  • Paralog accessions confirmed: LC3A Q9H492, LC3C Q9BXW4 (UniProt REST response)
• DOIs cited (all confirmed via except Mizushima 2010):
  • 10.1093/emboj/19.21.5720 (Kabeya 2000, EMBO J, 6,474 citations) — archive pending download; bronze OA
  • 10.1038/35044114 (Ichimura 2000, Nature, 2,016 citations) — local PDF available
  • 10.1074/jbc.m702824200 (Pankiv 2007, J Biol Chem, 4,465 citations) — archive pending download; hybrid OA
  • 10.1038/nature04724 (Hara 2006, Nature, 3,796 citations) — local PDF available
  • 10.1038/nature04723 (Komatsu 2006, Nature, 3,439 citations) — local PDF available
  • 10.4161/auto.4600 (Mizushima & Yoshimori 2007, Autophagy, 2,575 citations) — archive pending download; hybrid OA
  • 10.1080/15548627.2020.1797280 (Klionsky et al. 2021, Autophagy, 2,434 citations) — archive pending download; bronze OA
  • NOTE: Mizushima 2010 (10.4161/auto.6.1.10498) — DOI returns 404 from Crossref and not found in archive; Mizushima & Yoshimori 2007 used instead as the LC3 blotting-guidelines citation
• implicit stubs created (new wikilinks to non-existent pages):
  • sqstm1, atg4b, atg3, bnip3, fundc1, ndp52, optn, nbr1
• gaps surfaced:
  • needs-human-replication — Atg7-KO neurodegeneration is mouse-only; human counterpart genetic evidence lacking
  • unsourced — systematic multi-tissue LC3-II flux profiling with age in humans; multi-tissue ATG5/ATG7/Beclin-1 decline citation
  • unsourced — GABARAP subfamily aging-specific profiling vs LC3 subfamily
  • needs-replication — paralog compensation in LC3B single-KO models
• schema decisions: paralog sub-table format adopted (consistent with complex-subunits pattern in tsc1-tsc2.md); no novel schema invented; not escalating
• ROADMAP updated: Proteins section, [[lc3]] entry added after [[beclin-1]], marked (drafted, verified: false)
• suggested verification priority: Hara 2006 + Komatsu 2006 have local PDFs — verify neurodegeneration phenotype descriptions first; Kabeya 2000 + Ichimura 2000 carry the core mechanistic claims (LC3-I/II conversion; ubiquitin-like cascade) and should be downloaded (OA available for Kabeya via bronze OA URL) before biochemistry claims are signed off

[2026-05-04] ingest | round-6c-variants-receptors

lipophagy

• added: processes/lipophagy.md
• entity: process — selective autophagic degradation of lipid droplets
• canonical DB IDs: none applicable (process page; no UniProt/PubChem ID)
• DOI correction note: user briefing cited doi:10.1038/ncb3066 for Kaushik 2015 — this is a BUG-2 mismatch (resolves to SNAP-29 O-GlcNAc paper); correct DOI is 10.1038/ncb3166 (confirmed via PubMed PMID 25961502 + DOI lookup). Briefing also cited 10.1016/j.cmet.2011.01.018 for Singh & Cuervo 2011 review — mismatch (resolves to Paneni endothelial insulin paper); correct DOI is 10.1016/j.cmet.2011.04.004. Briefing cited 10.1038/ncomms3886 for Martinez-Lopez 2013 — mismatch (resolves to ribosome profiling paper); true Martinez-Lopez 2013 autophagy/adipose paper is PMID 23907538, doi:10.1038/embor.2013.111 (Myf5+ progenitors / brown fat) — deemed insufficiently focused on lipophagy for inclusion; omitted.
• primary sources cited (footnotes) — all confirmed via :
  • 10.1038/nature07976 (Singh 2009, Nature — lipophagy discovery; 3,816 citations; LOCAL PDF available)
  • 10.1016/j.cmet.2011.04.004 (Singh & Cuervo 2011, Cell Metab — review; 778 citations; pending download; bronze OA)
  • 10.1038/ncb3166 (Kaushik & Cuervo 2015, Nat Cell Biol — CMA degrades PLIN2/3; 688 citations; pending download; green OA — PMC4449813)
  • 10.1080/15548627.2015.1124226 (Kaushik & Cuervo 2016, Autophagy — AMPK-PLIN2-CMA; 244 citations; pending download; bronze OA)
  • 10.1038/ncb2718 (Settembre 2013, Nat Cell Biol — TFEB drives lipophagy; 971 citations; pending download; green OA — PMC3699877)
• implicit stubs created (new wikilinks to non-existent pages): lipa, plin2, plin3, rab7, nafld, ulk1 (may already exist — check)
• gaps surfaced:
  • no-mechanism — lipid droplet “eat-me” signal / cargo receptor for lipophagy unknown; FAF2/Rab7/PLIN dynamics implicated but no single receptor identified
  • needs-human-replication — controlled lipophagy-flux studies in human liver/hepatocytes sparse; most data from rodent liver
  • needs-replication — Rab7 requirement in mammalian lipophagy; FAF2/ubiquitin-LD link
  • needs-replication — microlipophagy characterization in mammals (mostly yeast evidence)
  • dose-response-unclear — TFEB activators as lipophagy inducers; tissue-specific dosing in aging
  • needs-replication — ectopic lipid accumulation via impaired lipophagy in non-hepatic aging tissues
• schema decisions: no novel frontmatter fields required; standard process schema used; selective-variants: [] left empty (lipophagy has no named sub-variants); no escalation needed
• ROADMAP updated: Tier B queue — lipophagy marked ✓ seeded R6c 2026-05-04; Round 6 Variants bullet updated to [x]
• suggested verification priority: Singh 2009 local PDF available — verify LD quantification, Atg7-KO phenotype numbers, and lipase activity data first; Kaushik 2015 (PMC4449813, OA) should be downloaded for PLIN2/3 CMA claims; Settembre 2013 (PMC3699877, OA) should be downloaded for TFEB overexpression + liver steatosis data

[2026-05-04] ingest | round-6c-variants-receptors

chaperone-mediated-autophagy

• added: processes/chaperone-mediated-autophagy.md (type: process, verified: false)
• inbound context: 3 inbound references per ROADMAP; referenced as [[chaperone-mediated-autophagy]] in processes/autophagy.md (selective-variants) and other pages seeded this round
• canonical DB IDs pulled: none — CMA has no KEGG pathway ID, Reactome pathway ID, or canonical small-molecule DB entry as a process; process-type schema does not require canonical DB IDs (#gap/needs-canonical-id not applicable here — by schema design)
• DOI corrections applied at seeding (4 of 7 task-supplied DOIs were wrong — BUG-2 class):
  • Chiang 1989: task brief supplied 10.1126/science.2538923 (resolves to “A Multiubiquitin Chain Is Confined to Specific Lysine” — wrong paper); corrected to 10.1126/science.2799391 (“A Role for a 70-Kilodalton Heat Shock Protein in Lysosomal Degradation”) per PMID 2799391 and DOI lookup title match
  • Cuervo & Dice 2000: task brief supplied 10.1126/science.288.5475.2199 (not found in archive); actual journal is J Biol Chem; corrected DOI 10.1074/jbc.M002102200 confirmed via PMID 10806201 and DOI lookup (“Age-related Decline in Chaperone-mediated Autophagy”)
  • Bandyopadhyay 2008: task brief supplied 10.1128/MCB.00115-08 (not found in archive); corrected to 10.1128/MCB.02070-07 (“The Chaperone-Mediated Autophagy Receptor Organizes in Dynamic Protein Complexes at the Lysosomal Membrane”) per PMID 18644871 and DOI lookup title match
  • Schneider 2014: task brief supplied 10.1016/j.cmet.2014.07.011 (resolves in archive to “Mitochondrial Complex I Activity Suppresses Inflammation and Enhances Bone Resorption” — wrong paper; BUG-2); corrected to 10.1016/j.cmet.2014.06.009 (“Deficient Chaperone-Mediated Autophagy in Liver Leads to Metabolic Dysregulation”) confirmed via Crossref query and DOI lookup
• primary sources cited (footnotes) — archive/PubMed status:
  • PMID 2103896 / Dice 1990, Semin Cell Biol — no DOI in PubMed (pre-DOI era); cited as PMID only; not in archive; needs-canonical-id
  • 10.1126/science.2799391 / Chiang 1989, Science — archive: not_oa; 959 citations
  • 10.1126/science.273.5274.501 / Cuervo 1996, Science — archive: not_oa; 937 citations
  • 10.1074/jbc.M002102200 / Cuervo & Dice 2000, J Biol Chem — archive: pending download (hybrid OA); 636 citations
  • 10.1128/MCB.02070-07 / Bandyopadhyay 2008, Mol Cell Biol — archive: pending download (green OA); 529 citations
  • 10.1016/j.cmet.2014.06.009 / Schneider 2014, Cell Metab — archive: pending download (bronze OA); 321 citations
  • 10.1038/s41580-018-0001-6 / Kaushik & Cuervo 2018, Nat Rev Mol Cell Biol — archive: download failed (green OA PMC6399518; should be re-attempted); 1,278 citations
  • 10.1080/15548627.2021.1935007 / Bourdenx 2021, Autophagy — not checked in archive; included for neurodegeneration section
• implicit stubs created (new wikilinks to non-existent pages): hsc70, lamp-2a, hsp90 (these three not confirmed as existing protein pages), alpha-synuclein, hip, hop, chip
• gaps surfaced:
  • No human in vivo CMA flux measurement in aging cohorts — needs-human-replication
  • KFERQ motif prevalence “~30%” is a computational prediction — needs-replication
  • CMA activators QX77/CA77.1 and AR.7 not yet tested in aged animal models — long-term-unknown dose-response-unclear
  • CMA–macroautophagy compensation threshold in aged tissue undefined — no-mechanism
  • Dice 1990 has no DOI in PubMed (pre-DOI era) — needs-canonical-id
  • Alpha-synuclein LAMP-2A dominant-negative blockade established in cells/mouse; human PD-brain confirmation correlative — needs-human-replication
• schema decisions: no novel frontmatter fields required; standard process schema used; selective-variants: [] left empty (CMA has no sub-variants analogous to mitophagy/lipophagy); no escalation needed
• ROADMAP updated: Processes section — chaperone-mediated-autophagy added as [x] drafted 2026-05-04
• suggested verification priority: Schneider 2014 (bronze OA, should download) — verify LAMP-2A overexpression quantitative rescue numbers and metabolic phenotype data first; Cuervo 2000 (hybrid OA, should download) — verify the “>30% decline” figure and LAMP-2A identification as cause; Bandyopadhyay 2008 (green OA, should download) — verify LAMP-2A complex size (~700 kDa) and lys-hsc70/hsp90 roles; Kaushik 2018 review (re-attempt PMC6399518 download) — verify pharmacological CMA activators section

fundc1

• added: molecules/proteins/fundc1.md
• entity: protein (FUNDC1 / FUN14 domain containing 1); UniProt Q8IVP5; NCBI Gene 139341; HGNC 28746
• inbound context: 2 inbound references per ROADMAP; referenced as [[fundc1]] in molecules/proteins/atg7.md and the mitophagy stack
• canonical sources pulled: UniProt Q8IVP5 (155 aa; N-terminal cytosolic tail + 3 TM segments + short IMS C-terminus; LIR motif Tyr18-x-x-Leu21; PTMs Tyr18/Ser13/Ser17; interactors MAP1LC3A/B, GABARAP, DNM1L, GPX4, USP19, PGAM5; accessed 2026-05-04 via rest.uniprot.org); NCBI Gene 139341 and HGNC 28746 confirmed from briefing (not independently live-re-verified against NCBI/HGNC)
• DOI corrections applied at seeding: briefing supplied 10.15252/embr.201438501 for Wu 2014 EMBO Rep — correct DOI is 10.1002/embr.201438501 (confirmed via DOI lookup; PMID 24671035; title match); briefing supplied 10.1126/science.1257529 for Chen 2014 Src paper — not found in archive; Src-related mechanistic content recovered from Liu 2012 (10.1038/ncb2422) and Chen 2014 Mol Cell (10.1016/j.molcel.2014.02.034); Science DOI excluded; briefing supplied 10.1126/scisignal.aaq1188 for Wu 2017 — not found in archive; correct DOI confirmed as 10.1161/CIRCULATIONAHA.117.030235 (Circulation; PMID 28942427; DOI lookup title match)
• primary sources cited (footnotes) — archive status:
  • 10.1038/ncb2422 (Liu 2012, Nat Cell Biol — FUNDC1 discovery; LIR Tyr18; 1,567 citations; not_oa — no local PDF)
  • 10.1016/j.molcel.2014.02.034 (Chen 2014, Mol Cell — PGAM5/CK2/Ser13 loop; 542 citations; not_oa — no local PDF)
  • 10.1002/embr.201438501 (Wu 2014, EMBO Rep — ULK1 Ser17; 551 citations; bronze OA — pending download)
  • 10.4161/auto.29568 (Wu 2014, Autophagy — BCL2L1/PGAM5; 177 citations; bronze OA — pending download)
  • 10.1080/15548627.2016.1238552 (Kuang 2016, Autophagy — LIR structural phospho-switch; 139 citations; bronze OA — pending download)
  • 10.15252/embj.201593102 (Wu 2016, EMBO J — FUNDC1 at MAM + DRP1 fission; 335 citations; green OA PMC4864280 — pending download)
  • 10.1080/15548627.2016.1193656 (Wu 2016, Autophagy — FUNDC1 as MAM protein; 181 citations; bronze OA — pending download)
  • 10.1161/CIRCULATIONAHA.117.030235 (Wu 2017, Circulation — cardiac KO + human heart failure; 301 citations; green OA PMC5716911 — pending download)
  • 10.1080/15548627.2019.1580095 (Lampert 2019, Autophagy — NIX+FUNDC1 cardiac progenitor remodeling; 299 citations; bronze OA — pending download)
• implicit stubs created (new wikilinks to non-existent pages): drp1, pgam5, bnip3l, ip3r2
• gaps surfaced:
  • FUNDC1 mitophagy flux in aged human tissues (heart/liver/muscle) not measured — needs-human-replication
  • Human heart failure FUNDC1 data: small cohort; causal direction unclear — needs-replication
  • Src inactivation mechanism at OMM under hypoxia — not mechanistically resolved — no-mechanism
  • FUNDC1–GPX4 ferroptosis interaction — direct vs. indirect unclear — no-mechanism
  • MAM-DRP1 fission function: cell-line data primarily; aged-tissue in vivo lacking — needs-replication
  • FUNDC1/BNIP3/NIX pathway redundancy under physiological hypoxia — undefined — needs-replication
  • GenAge entry for FUNDC1 not confirmed — needs-canonical-id
  • No clinical-stage pharmacological FUNDC1 activator — long-term-unknown
• schema decisions: no novel frontmatter fields required; standard protein schema; genage-id: null with gap tag; briefing-supplied Science DOI (10.1126/science.1257529) excluded due to archive non-confirmation; not escalating schema
• ROADMAP updated: Selective autophagy receptors line — [[fundc1]] marked [x] drafted 2026-05-04, verified: false
• suggested verification priority: Liu 2012 (not_oa — most critical for LIR discovery claims; recommend download attempt); Wu 2017 Circulation (green OA PMC5716911 — download and verify human heart failure cohort size and FUNDC1 protein quantification); Chen 2014 Mol Cell (not_oa — key for PGAM5/CK2/Ser13 mechanistic claims); Kuang 2016 (bronze OA — download for ~10-fold affinity enhancement figure)

[2026-05-04] ingest | round-6d-spermidine-atg-conjugation

spermidine

• added: molecules/compounds/spermidine.md
• entity: compound — natural polyamine; autophagy inducer; CR-mimetic; PubChem CID 1102; ChEMBL CHEMBL1374440
• canonical sources pulled: PubChem CID 1102 confirmed (MW 145.25, InChIKey ATHGHQPFGPMSJY-UHFFFAOYSA-N, logP -1.0, C7H19N3, IUPAC N’-(3-aminopropyl)butane-1,4-diamine); accessed 2026-05-04
• DOI corrections — 2 of 5 task-supplied DOIs were wrong (BUG-2 class):
  • Kiechl 2018: brief supplied 10.3945/ajcn.117.165886 — archive matched this to “Can changes in the plasma lipidome help explain cardiovascular benefits of Mediterranean diet?” (~1 citation); correct DOI 10.1093/ajcn/nqy102 confirmed via PubMed PMID 29955838 (“Higher spermidine intake is linked to lower mortality”; ~225 citations)
  • Pietrocola 2015: brief supplied 10.1016/j.cmet.2015.05.012 — archive matched to “A Periodic Diet that Mimics Fasting Promotes Multi-System Regeneration” (fasting-mimicking diet paper; 836 citations — different paper); correct DOI 10.1038/cdd.2014.215 confirmed via PubMed PMID 25526088 (“Spermidine induces autophagy by inhibiting the acetyltransferase EP300”; Cell Death and Differentiation, NOT Cell Metabolism)
• model-organism lifespan attribution (critical distinction):
  • Eisenberg 2009 (10.1038/ncb1975, Nat Cell Biol): yeast chronological lifespan + C. elegans lifespan + Drosophila lifespan extension confirmed; mouse data = oxidative stress reduction ONLY — no mouse lifespan data in this paper
  • Eisenberg 2016 (10.1038/nm.4222, Nat Med): mouse lifespan extension + cardioprotection (diastolic function, hypertrophy, arterial stiffness) — this is the correct citation for mouse longevity claims
• primary sources cited (footnotes) — archive status:
  • 10.1038/ncb1975 (Eisenberg 2009, Nat Cell Biol — yeast/worm/fly lifespan; 1,586 citations; bronze OA; archive: download failed)
  • 10.1038/nm.4222 (Eisenberg 2016, Nat Med — mouse lifespan + cardio; 1,060 citations; green OA via HAL; archive: pending)
  • 10.1038/cdd.2014.215 (Pietrocola 2015, Cell Death Differ — EP300/HAT mechanism; 300 citations; bronze OA; archive: pending)
  • 10.1126/science.aan2788 (Madeo 2018, Science — review; 1,012 citations; closed-access; archive: not_oa; no-fulltext-access)
  • 10.1093/ajcn/nqy102 (Kiechl 2018, Am J Clin Nutr — Bruneck dietary spermidine + mortality; 225 citations; bronze OA; archive: pending)
• implicit stubs created (new wikilinks to non-existent pages): ep300, caloric-restriction (may already exist as intervention page — check)
• gaps surfaced:
  • Mouse lifespan (Eisenberg 2016) not yet independently replicated — needs-replication
  • No RCT with longevity/healthspan endpoint — needs-human-replication
  • Dose-response in humans unknown; dietary vs supplement doses differ ~10× — dose-response-unclear
  • HAT-broad vs EP300-specific mechanism not fully resolved in human tissues — no-mechanism
  • Tissue-level endogenous spermidine decline with age — sparse quantitative human data — needs-replication
  • Long-term supplementation safety data absent — long-term-unknown
  • ClinicalTrials.gov API returned 404 at seeding — NCT03847870 status unverified — unsourced
• schema decisions: no novel frontmatter fields required; standard compound schema used; measured-targets field used for EP300 (most quantitatively supported target); targets field used for pathway-level links; no escalation needed
• ROADMAP updated: Compounds section spermidine marked [x] drafted 2026-05-04; Tier A queue row updated with ✓ seeded R6d
• suggested verification priority: Eisenberg 2016 (green OA via HAL, should download — hal-04702765) — verify mouse lifespan curve quantification, cardiac endpoint values, n’s, and autophagy-dependence epistasis data first; Kiechl 2018 (bronze OA, should download) — verify Bruneck Study n, HR values, tertile cutpoints, and confounders; Pietrocola 2015 (bronze OA, should download) — verify EP300 IC50 values, specificity vs other HATs, and in vivo acetylome data; Eisenberg 2009 (bronze OA) — re-attempt download and verify that no mouse lifespan data exist (confirms autophagy.md verifier finding)

[2026-05-04] ingest | round-6c-variants-receptors

bnip3

• added: molecules/proteins/bnip3.md
• entity: protein (BNIP3 / NIP3 / BCL2/adenovirus E1B 19 kDa protein-interacting protein 3); UniProt Q12983; NCBI Gene 664; HGNC 1084
• inbound context: 2 inbound references per ROADMAP Tier D; referenced from mitophagy (verified-partial) and bcl-2-family-signaling (verified-partial); round 6c per ROADMAP
• canonical sources pulled: UniProt Q12983 (194 aa, tail-anchored, BNIP3_HUMAN); NCBI Gene 664; HGNC 1084; UniProt REST API returned 500 error at seeding — identity fields from task briefing, not live-verified; flagged for lint pass
• DOI corrections applied at seeding (3 of 6 task-supplied DOIs were BUG-2 mismatches):
  • Boyd 1994: task brief supplied 10.1038/371375a0 (archive title: “How to stop bad B cells” — wrong paper; BUG-2); correct DOI for Boyd 1994 Nature 371:382-385 not confirmed; excluded from citations; flagged needs-canonical-id
  • Sandoval 2008: task brief supplied 10.1038/nature06834 (archive title: “Chemical compass model of avian magnetoreception” — wrong paper; BUG-2); corrected to 10.1038/nature07006 (“Essential role for Nix in autophagic maturation of erythroid cells”) per PMID 18454133 + DOI lookup (1,133 citations; LOCAL PDF available)
  • Schweers 2007: task brief supplied 10.1073/pnas.0709047104 (not found in archive); corrected to 10.1073/pnas.0708818104 (“NIX is required for programmed mitochondrial clearance during reticulocyte maturation”) per Crossref search + DOI lookup (886 citations; green OA, pending download)
  • Hanna 2012: task brief supplied 10.1074/jbc.M111.332155 (not found in archive); corrected to 10.1074/jbc.m111.322933 (“Microtubule-associated Protein 1 Light Chain 3 (LC3) Interacts with Bnip3 Protein…”) per Crossref search + PMID 22505714 + DOI lookup (716 citations; hybrid OA, pending download)
  • Quinsay 2010: task brief supplied 10.1080/15548627.2010.10.4.13196 (not found in archive); corrected to 10.4161/auto.6.7.13005 per PubMed PMID 20668412 + DOI lookup (231 citations; bronze OA, pending download)
  • Chen 1997: task brief supplied 10.1084/jem.186.12.1975 — confirmed correct in archive (331 citations; bronze OA, pending download)
  • Bruick 2000 (new — not in task brief): added 10.1073/pnas.97.16.9082 for HIF-1α transcriptional regulation (762 citations; not_oa)
  • Rikka 2011 (new — not in task brief): added 10.1038/cdd.2010.146 for cardiac BNIP3 bioenergetics/turnover (286 citations; bronze OA, pending download)
• primary sources cited (footnotes) — all confirmed via :
  • 10.1084/jem.186.12.1975 (Chen 1997, J Exp Med — NIP3/BNIP3 dimeric pro-apoptotic OMM protein; 331 citations; bronze OA, pending download)
  • 10.1074/jbc.m111.322933 (Hanna 2012, J Biol Chem — BNIP3 LIR W18-V-V-L21 + LC3/ER/mito autophagy; 716 citations; hybrid OA, pending download)
  • 10.4161/auto.6.7.13005 (Quinsay 2010, Autophagy — BNIP3 mitophagy mPTP-independent; 231 citations; bronze OA, pending download)
  • 10.1073/pnas.0708818104 (Schweers 2007, PNAS — NIX required for reticulocyte mitochondrial clearance; 886 citations; green OA, pending download)
  • 10.1038/nature07006 (Sandoval 2008, Nature — NIX as direct autophagy receptor in erythroid maturation; 1,133 citations; LOCAL PDF available)
  • 10.1073/pnas.97.16.9082 (Bruick 2000, PNAS — BNIP3 HIF-1-induced by hypoxia; 762 citations; not_oa)
  • 10.1038/cdd.2010.146 (Rikka 2011 [pub. year 2010 online], Cell Death Differ — BNIP3 impairs bioenergetics + stimulates mitochondrial turnover; 286 citations; bronze OA, pending download)
  • doi:10.1016/j.bbadis.2018.09.034 (Moyzis & Gustafsson 2019, BBA-Mol Basis Dis — review; cited for cardiac aging context; unsourced for primary BNIP3 aging-decline quantification)
• implicit stubs created (new wikilinks to non-existent pages): fundc1, nix (BNIP3L paralog page — distinct from bnip3.md)
• gaps surfaced:
  • Boyd 1994 cloning DOI unverified (BUG-2) — needs-canonical-id
  • UniProt Q12983 live API not verified at seeding — needs-canonical-id (lint pass recommended)
  • GenAge entry for BNIP3 not confirmed — needs-canonical-id
  • BNIP3 LIR affinity values (Kd to LC3A/B/C and GABARAP family) not quantified in archived sources — needs-replication
  • BNIP3 BH3-like motif affinity (Kd to BCL-2/BCL-xL/MCL-1) not quantified — needs-replication
  • Arm-switch mechanism (mitophagy vs. apoptosis) not mechanistically defined — no-mechanism
  • Quantitative BNIP3 protein decline with age in human cardiac tissue — no confirmed primary source; most data from rodent models — needs-human-replication
  • Bnip3 x Bnip3l double-KO aging phenotype — not characterized — needs-replication
  • Receptor-dependent (BNIP3/NIX) vs. ubiquitin-dependent (PINK1/Parkin) mitophagy partitioning in aged tissue — unmeasured — no-mechanism
• schema decisions: no novel frontmatter fields required; standard protein schema used; sens-categories: [] (BNIP3 does not map cleanly to MitoSENS — it is a quality control mechanism, not a damage category); no escalation needed
• ROADMAP updated: Tier D queue — bnip3 marked ✓ seeded R6c 2026-05-04
• suggested verification priority: Sandoval 2008 LOCAL PDF available — verify NIX LIR mechanism, reticulocyte mitophagy defect quantification (% mitochondria-retaining erythrocytes), and erythroid specificity; Hanna 2012 (hybrid OA, should download) — verify LIR residue assignments (W18-V-V-L21) and LC3/GABARAP binding specificity data; Rikka 2011 (bronze OA, should download) — verify BNIP3 direction of effect on OCR and cardiomyocyte phenotype numbers; Quinsay 2010 (bronze OA, should download) — verify mPTP independence claim and cardiomyocyte assay conditions

[2026-05-04] verify | molecules/proteins/beclin-1.md

• page: molecules/proteins/beclin-1.md
• sources verified (PDFs read end-to-end):
  • Pattingre 2005 (10.1016/j.cell.2005.07.002) — downloaded from Cell Press via camoufox; full 10-page PDF read
  • Wei 2008 (10.1016/j.molcel.2008.06.001) — downloaded from Cell Press via camoufox; full 11-page PDF read
  • Fernández 2018 (10.1038/s41586-018-0162-7) — downloaded from PMC5992097; full 19-page PDF (author manuscript) read
  • Yue 2003 (10.1073/pnas.2436255100) — downloaded from PNAS; full 6-page PDF read
• sources unverifiable (closed-access, not_oa):
  • Liang 1999 (10.1038/45257) — tagged no-fulltext-access
  • Zalckvar 2009 (10.4161/auto.5.5.8625) — tagged no-fulltext-access
• corrections made (5 substantive):
  1. Domain table: “F121A abolishes interaction” → “mouse F121A (= human F123A) decreases interaction” (Fernández 2018 says “decreases”; Pattingre uses F123A human numbering throughout)
  2. BH3 binding domain note added: Pattingre 2005 characterises BCL-2-binding domain as aa 88–150 and uses F123A; Fernández uses mouse F121A equivalent
  3. F121A lifespan claim: “~10% median lifespan extension” → exact medians: WT=26 mo (n=68) vs KI=29 mo (n=102), p<0.0001 log-rank; females WT=27 vs KI=30 mo p=0.0004; males WT=25 vs KI=28 mo p=0.0004. Max lifespan WT=36 mo vs KI=39 mo.
  4. “improved metabolic fitness (reduced adiposity, better glucose tolerance)” removed — these outcomes not reported in Fernández 2018; replaced with actual outcomes: delayed renal/cardiac aging and reduced non-lymphoid spontaneous tumorigenesis
  5. Beclin-1 age-decline attribution removed from Fernández 2018 — that paper does not report age-related Beclin-1 protein level changes; claim tagged unsourced pending primary source identification
• verified-scope: Pattingre 2005, Wei 2008, Fernández 2018, Yue 2003 verified against PDFs; Liang 1999 and Zalckvar 2009 not_oa; UniProt canonical identity fields not re-verified against database

---

[2026-05-04] verify | tfeb

Pages verified: 1 (partial — 3 of 6 sources not_oa or download-failed)

• molecules/proteins/tfeb.md — corrections made (see below); verified: true (partial scope — see verified-scope in frontmatter)

Sources checked:

• 10.1126/science.1174447 (Sardiello 2009, Science) — NOT_OA. Abstract verified via PubMed PMID 19556463. Gene-count numerics (471 predicted / 96 confirmed) not in abstract; cannot be verified without full PDF. Tagged unsourced no-fulltext-access.
• 10.1126/science.1204592 (Settembre 2011, Science) — DOWNLOADED (PMC3638014). VERIFIED against full PDF. Critical finding: ERK2 (MAPK pathway) is the primary TFEB kinase identified in this paper (phosphorylates Ser142); rapamycin had minimal effect on TFEB localization. Pearson r=0.42 between TFEB-OE and starvation-induced autophagy gene expression (P=0.001), n=51 autophagy genes, 11 significantly upregulated.
• 10.1038/emboj.2012.32 (Settembre 2012, EMBO J) — DOWNLOADED (PMC3298007). VERIFIED against full PDF. Critical finding: mTORC1 primary phosphosite is Ser142 (confirmed by phospho-antibody and in-vitro kinase assay); Ser211 is an additional site (S211A causes constitutive nuclear TFEB by mutagenesis). Both are rapamycin-resistant. Torin1 EC50 ≈ 148 nM vs rapamycin EC50 ≈ 104,322 nM for TFEB nuclear translocation. Rag GTPases necessary and sufficient. Mouse hepatocytes (Tcfeb conditional KO) showed TFEB-dependent transcriptional response to Torin1 and chloroquine.
• 10.1126/scisignal.2002790 (Roczniak-Ferguson 2012, Sci Signal) — NOT_OA. Abstract verified via PubMed PMID 22692423. Ser211 is confirmed as the mTOR-dependent phosphosite creating 14-3-3 docking site. MITF and TFE3 similarly regulated.
• 10.1038/ncb3114 (Medina 2015, Nat Cell Biol) — NOT_OA. Abstract verified via PubMed PMID 25720963. MCOLN1 → calcineurin → TFEB dephosphorylation confirmed. Calmodulin as intermediate not confirmed from abstract.
• 10.1038/ncomms3267 (Lapierre 2013, Nat Commun) — DOWNLOAD FAILED (bronze OA but URL filter excluded). Abstract verified via PubMed PMID 23925298. Six longevity models (not just eat-2/daf-2) confirmed. Mouse liver nuclear TFEB elevation under DR confirmed.

Corrections made to molecules/proteins/tfeb.md:

1. key-ptms order: Ser211 first → Ser142 first (Settembre 2012 identifies S142 as primary mTORC1 substrate by kinase assay and phosphoantibody; S211 secondary)
2. mTOR phospho-switch claim: “mTORC1 phosphorylates TFEB at Ser211 (and also at Ser142) [settembre2012] [roczniak-ferguson2012]” → corrected to accurately reflect that Settembre 2012 identifies S142 as primary kinase-assay site and S211 as additional mutagenesis-confirmed site; Roczniak-Ferguson 2012 identifies Ser211 as the 14-3-3 docking site. These two papers are in agreement but focused on different aspects.
3. Rapamycin claim: “rapamycin” listed as TFEB activator alongside Torin1 → corrected to show rapamycin is a poor TFEB activator (EC50 ≈ 104,322 nM vs Torin1 EC50 ≈ 148 nM per Settembre 2012 high-content assay). S142 and S211 are rapamycin-resistant mTORC1 phosphosites.
4. Settembre 2011 kinase attribution: wiki body implied mTOR inhibition → TFEB nuclear translocation → corrected to reflect Settembre 2011 identified ERK2 (MAPK) as the primary kinase in that paper, with rapamycin having minimal effect.
5. Lapierre 2013 longevity scope: “suppresses longevity of eat-2 and daf-2 mutants” → corrected to “essential for extended lifespan in six mechanistically distinct longevity models” (per abstract).
6. Domain description: “C-terminal region contains the principal mTORC1-regulated Ser211 phosphorylation site” → corrected per Settembre 2012 Fig. 3C domain map (Ser142 in TACD region, Ser211 between TACD and HLH).
7. Regulation table: updated to accurately represent both papers’ findings, including ERK2 row (Settembre 2011) and correct rapamycin-resistance annotation.
8. Calmodulin intermediate: flagged as unverifiable from Medina 2015 abstract; note added.
9. Sardiello 2009 gene counts: 471/96 figures flagged as unverifiable from abstract; unsourced and no-fulltext-access added.
10. All 6 footnotes updated with accurate model systems, quantitative details, and access status.
11. Banner removed; verified flag flipped to true (partial scope).

Pages unverifiable (closed-access):

• 10.1126/science.1174447 (Sardiello 2009) — not_oa; 471/96 gene counts unverified
• 10.1126/scisignal.2002790 (Roczniak-Ferguson 2012) — not_oa; abstract verified only
• 10.1038/ncb3114 (Medina 2015) — not_oa; abstract verified only; calmodulin intermediate unverified

Downstream pages that may need updates (for main agent — not propagated here):

• processes/autophagy.md — carries #gap/unsourced “~500 CLEAR genes” note; this page partially resolves it but confirms the 471/96 figures are themselves unverified; autophagy.md gap tag should remain and note Sardiello 2009 is not_oa
• Any other entity page citing mTOR→TFEB→Ser211 as the sole phosphosite without mentioning Ser142 should be updated to reflect dual-site model from Settembre 2012
• processes/lysosomal-biogenesis.md (if exists) — may carry Sardiello 2009 gene-count claims

[2026-05-04] verify | atg12

Pages verified: 1 (partial — Mizushima 1998 Nature not_oa; Mizushima 1998 JBC download failed)

• molecules/proteins/atg12.md — corrections made; verified: true with partial scope

Sources checked:

• 10.1016/j.molcel.2011.10.014 (Rubinstein 2011, Mol Cell) — local PDF. VERIFIED end-to-end. BH3-like motif residue range corrected; conjugation-masking claim corrected.
• 10.1091/mbc.e07-12-1257 (Fujita 2008, MBC) — local PDF. VERIFIED end-to-end. ATG16L1 coiled-coil domain location corrected (central, not C-terminal).
• 10.1074/jbc.C700195200 (Hanada 2007, JBC) — local PDF. VERIFIED end-to-end. Yeast-only system clarified in footnote.
• 10.1038/nature03029 (Kuma 2004, Nature) — local PDF. VERIFIED end-to-end. Neonatal lethality timing confirmed (~24 h / 12.4±1.3 h). Mechanistic analogy for Atg12 KO retained.
• 10.1371/journal.pone.0023367 (Kang 2011, PLoS ONE) — local PDF. VERIFIED end-to-end. Premature senescence phenotype (SA-β-gal, p21waf1, ROS, lipofuscin, p53-dependent) confirmed; cell models (IMR90, HDF) added to footnote.
• 10.1038/ncomms3300 (Pyo 2013, Nat Commun) — local PDF. VERIFIED end-to-end. n=65+70, χ²=17.32, p<0.001, ~17.2% median lifespan extension all confirmed.
• 10.1038/26506 (Mizushima 1998, Nature) — UNVERIFIABLE. not_oa. Conjugation cascade claims attributed to this paper cannot be verified. Tagged no-fulltext-access.
• 10.1074/jbc.273.51.33889 (Mizushima 1998, JBC) — UNVERIFIABLE. Download failed (no PMC mapping, publisher URL inaccessible). Tagged no-fulltext-access.

Corrections made to molecules/proteins/atg12.md:

1. BH3-like motif residue range: “residues 136–149, overlapping the C-terminal Gly140 region” → “approximately residues 62–76, centered on D64” (Rubinstein 2011 Fig. 3A sequence alignment; D64S/D64N point mutations abolish Bcl-2/Mcl-1 binding, Fig. 3H). The motif is in the N-terminal disordered segment, not near Gly140.
2. ATG16L1 coiled-coil domain: “C-terminal coiled-coil domain” → “central coiled-coil domain” with clarification of full ATG16L1 domain organization (N-terminal ATG5-binding, central coiled-coil for self-multimerization, C-terminal WD repeat) per Fujita 2008.
3. Conjugation-masking claim: Rewritten to accurately reflect Rubinstein 2011 results — conjugation-deficient Atg12ΔG140 retains Bcl-2 binding (Fig. 4B); it is the low abundance of free ATG12 at steady state (not steric masking by conjugation) that limits pro-apoptotic activity.
4. Hanada 2007 E3-like activity: Footnote updated to specify purified S. cerevisiae protein system; added note that conservation in mammals is inferred, not independently demonstrated in that paper.
5. Kang 2011 footnote: Cell models specified (IMR90, HDF); “shRNA knockdown corroborated siRNA” added; local PDF path added.
6. Fujita 2008 footnote: ATG16L1 domain organization detail added; local PDF path added.
7. Rubinstein 2011 footnote: BH3-like motif residue detail and free-ATG12 pool framing updated; local PDF path added; BUG-2 DOI mismatch note retained.
8. Mizushima 1998 Nature footnote: Updated from “locally: not downloaded (closed OA)” to “not_oa — closed-access; cannot verify”.
9. Mizushima 1998 JBC footnote: Updated from “pending” to “download failed — no PMC mapping found, publisher URL inaccessible”.
10. ⚠️ banner removed; verified flag flipped to true (partial scope).

Pages unverifiable (closed-access / failed download):

• 10.1038/26506 (Mizushima 1998, Nature) — not_oa; conjugation reaction E1/E2 claims cannot be confirmed no-fulltext-access
• 10.1074/jbc.273.51.33889 (Mizushima 1998, JBC) — download failed; human ATG12 140 aa, Lys130 isopeptide claims cannot be confirmed no-fulltext-access

Downstream pages that may need updates (for main agent — not propagated here):

• molecules/proteins/atg5.md line 62: “N-terminal coiled-coil domain of ATG16L1” — should be “central coiled-coil domain” per Fujita 2008 (same correction as ATG12)
• molecules/proteins/atg5.md — ATG5 page is already verified; the coiled-coil domain description propagation is a targeted single-line fix
• processes/autophagy.md — no corrected claims found in scan; no action needed
• molecules/proteins/bnip3.md, fundc1.md — reference atg12 only via wikilinks; no corrected claims found in scan

[2026-05-04] round-6 summary

Round 6 (Autophagy machinery completion) — 15 new pages drafted + verified. Sub-rounds:

• R6a (core autophagy): beclin-1, lc3, tfeb, atg7
• R6b (PINK1-Parkin axis + cargo receptor): pink1, parkin, p62, pink1-parkin-pathway
• R6c (variants + receptors): lipophagy, chaperone-mediated-autophagy, bnip3, fundc1
• R6d (autophagy compound + conjugation): spermidine, atg5, atg12

Major corrections from verifier batches:

• Hara 2006 misattribution surfaced as systemic error — DOI 10.1038/nature04724 is an ATG5 paper (Atg5^flox/flox × nestin-Cre), not an Atg7 paper. Wiki had it consistently misattributed across atg7.md, lc3.md, atg5.md, and any references. All corrected; study slug needs renaming from hara-2006-neuronal-atg7-ko-neurodegeneration to -atg5-.
• PINK1 PARL contradiction — Narendra 2010 (PMC) explicitly shows PARL is dispensable for PINK1 cleavage in mammalian cells (PARL shRNA + PARL-/- MEFs do not change PINK1 cleavage); contradicts the Kazlauskaite/Kane 2014 framing. Now flagged contradictory-evidence on pink1.md and pink1-parkin-pathway.md.
• Parkin pUb binding site corrected — pUb binds RING1 at His302/Arg305 (not Ubl-RING0 interface) per Wauer 2015 Fig 4. H302A mutation reduces affinity 60-fold by ITC. Corrected on parkin.md and pink1-parkin-pathway.md.
• Parkin domain boundaries all wrong — corrected: RING0 ~141-225, linker 77-140 (disordered), RING1 ~226-326, IBR ~327-378, REP ~378-410, RING2 ~410-465 per Wauer 2015 Fig 1A.
• PINK1 domain boundaries — TM 85-110 (not 94-110), kinase domain 156-511 (not 509) per Onishi 2021 Fig 3A.
• TFEB primary mTOR site is Ser142 (kinase-assay confirmed) per Settembre 2012; Ser211 is the 14-3-3 docking site per Roczniak-Ferguson 2012. Wiki had Ser211 as primary.
• Rapamycin is a poor TFEB activator (EC50 ~104 µM vs Torin1 ~148 nM) — Settembre 2012 Fig 2D. S142/S211 are rapamycin-resistant mTORC1 phosphosites.
• TFEB CLEAR network: 471 predicted / 96 confirmed gene counts are NOT in Sardiello 2009 abstract; paper is closed-access. Tagged needs-replication.
• FUNDC1 LIR motif corrected — Y18-E19-V20-L21 (Y18EVL), NOT “Y-x-x-L” or “Y18-D19-D20-L21”. Critical for designing peptide inhibitors.
• Kuang 2016 ITC values added/corrected — pTyr18 weakens LC3 binding ~4-fold (0.40→1.72 µM); pSer13 small effect (0.40→0.60 µM); Ser17-phospho Kd was NEVER measured by Kuang 2016. The wiki’s “~10-fold Ser17 affinity enhancement” claim was fabricated.
• BNIP3 LIR is W18-V19-E20-L21 (WVEL) — not WVVL. BNIP3 binds LC3 but NOT GABARAP (NIX shows opposite preference).
• Schneider 2014 was a CMA liver-KO paper (not LAMP-2A overexpression rescue). The actual rescue paper is Zhang & Cuervo 2008 (10.1038/nm.1851) — verified in FVB-bg double Tg, aged 22-26 mo. Major corrections to chaperone-mediated-autophagy.md.
• Spermidine mechanism: EP300-specific, NOT HAT-broad — per Pietrocola 2015 (full PDF verified). C646 (EP300-selective) sufficient. The autophagy.md verifier’s “HAT-broad” framing was an over-extrapolation. Spermidine.md now correctly cites EP300 inhibition.
• Eisenberg 2009 lifespan was yeast/worm/fly only; mouse lifespan extension came from Eisenberg 2016 (Nat Med, life-long N=40/41 + late-life N=91/86, p=0.018 / p=0.012).
• Beclin-1 F121A lifespan extended ~12% combined (WT 26 mo / KI 29 mo p<0.0001 log-rank), NOT “~10%”. Specific medians + full sex-stratified data added per Fernández 2018.
• Settembre 2011 primary kinase is ERK2 (MAPK) — not mTOR. mTOR-driven TFEB translocation comes from Settembre 2012.
• Yang et al. 2006 (NOT “Clark et al. 2006”) for Drosophila parkin RNAi epistasis (DOI 10.1073/pnas.0602493103).
• Lipophagy correction: AMPK-PLIN2 causal order INVERTED. Hsc70 binds PLIN2 first; AMPK phosphorylates after. Specific Ser492/Ser494 residues were never reported in the primary papers.
• Multiple BUG-2 DOI mismatches caught in seeder briefs (BNIP3 had 5 wrong DOIs; FUNDC1 had 3; CMA had 4; spermidine had 2; ATG12 had 1). All corrected during verification.

Archive defects surfaced: BUG-3 (Komatsu 2007 PDF was a Cell Press recruitment kit), BUG-4 (Wu 2016 Autophagy PDF was a Portuguese physiology thesis), BUG-5 (Schneider 2014 PDF same as BUG-3 pattern). Filed to a local paper archive/FEATURE_REQUESTS.md.

Coverage delta: +15 protein/process/pathway/compound pages; key Tier A/B implicit stubs closed (beclin-1 9 inbound, spermidine 9 inbound, p62 3, etc.). Autophagy/mitophagy cluster substantially complete.

---

[2026-05-04] ingest | round-6e-followup

nix

• added: molecules/proteins/nix.md (NIX / BNIP3L — mitophagy receptor, BNIP3 paralog)
• updated: ROADMAP.md — marked [[nix]] as drafted R6e 2026-05-04; resolved [[bnip3l]] alias in ROADMAP entry
• type: protein
• canonical IDs confirmed: UniProt O60238, NCBI Gene 665, HGNC 1085 (from task brief; not live-verified against databases this pass)
• key sources cited: Chen 1999 (10.1074/jbc.274.1.7 — archive pending), Sowter 2001 (PMID 11559532 — no confirmed DOI), Hanna 2012 (10.1074/jbc.m111.322933 — downloaded), Novak 2010 (10.1038/embor.2009.256 — archive pending), Schweers 2007 (10.1073/pnas.0708818104 — downloaded), Sandoval 2008 (10.1038/nature07006 — downloaded)
• critical content: NIX LIR preferentially binds GABARAP/GABARAPL1/L2 (not LC3); BNIP3 LIR preferentially binds LC3 (not GABARAP) — differential specificity per Hanna 2012 (verified on bnip3.md, carried forward here); Nix-/- erythroid mitophagy defect (Schweers 2007: ~40% mito-retaining total erythrocytes vs ~1% WT; Sandoval 2008: 28% at 3 weeks, 12% at 6 weeks); BNIP3 cannot rescue NIX deficiency in erythrocytes
• gaps surfaced: Sowter 2001 has no confirmed DOI (#gap/needs-canonical-id); Chen 1999 and Novak 2010 PDFs pending download; exact NIX LIR residue positions are approximate; cardiac aging primary data for NIX protein-level trajectories unsourced (#gap/unsourced); GenAge entry ID unknown (#gap/needs-canonical-id); FAD-binding hypothesis disproof source unknown (#gap/unsourced)
• implicit stubs created: gabarap (linked from key interactors — page may not exist), chen-1999-nix-bnip3l-cloning, sowter-2001-nix-bnip3-hif1-hypoxia, novak-2010-nix-autophagy-receptor
• verification priority: HIGH — Schweers 2007 and Sandoval 2008 PDFs are downloaded (verified on bnip3.md); Hanna 2012 is downloaded; Chen 1999 and Novak 2010 need download before full verification. Sowter 2001 has no DOI — cite permanently by PMID. Suggest verifying LIR residue positions against Novak 2010 or Hanna 2012 PDFs as first priority.

[2026-05-04] ingest | round-6e-followup

xenophagy

• added: processes/xenophagy.md
• entity type: process (selective autophagy of intracellular pathogens)
• verified: false (auto-extracted; no primary PDFs read end-to-end)

DOIs cited (8 total):

DOI	Title	Archive status
10.1016/j.cell.2004.11.038	Gutierrez 2004 Cell — founding xenophagy paper (M. tuberculosis)	bronze OA, pending download
10.1038/ni.1800	Thurston 2009 Nat Immunol — NDP52 in Salmonella xenophagy	local PDF available
10.1126/science.1205405	Wild 2011 Science — TBK1/OPTN Ser177 phosphorylation	not_oa
10.1038/nature10744	Thurston 2012 Nature — galectin-8 damage sensor	not_oa
10.1038/nature09782	Levine 2011 Nature — autophagy in immunity (review)	local PDF available
10.1371/journal.ppat.1005174	Tumbarello 2015 PLoS Pathog — TAX1BP1/myosin VI	gold OA, pending download
10.4161/auto.5.8.10177	Bhatt 2009 Autophagy — Listeria ActA evasion	bronze OA, pending download
10.1080/15548627.2021.1926656	Chua 2022 Autophagy — TBK1/ALS review	pending download

BUG-2 mismatch detected and corrected:

• Task-supplied DOI for Levine 2011 was 10.1016/j.cell.2010.12.024 — this resolves to “SnapShot: Chromatin Remodeling: INO80 and SWR1” (Cell 2011, 54 citations). The correct DOI for the xenophagy/immunity review is 10.1038/nature09782 (“Autophagy in immunity and inflammation,” Nature 2011, 3,189 citations, local PDF available). Corrected on the page and noted in the auto-extraction banner.

Implicit stubs created (new wikilinks to non-existent pages):

• tbk1 — TANK-binding kinase 1; master xenophagy/mitophagy kinase; ALS/FTD Mendelian gene
• tax1bp1 — fourth canonical xenophagy receptor; myosin VI co-factor
• galectin-8 — early vacuole-damage sensor; recruits NDP52
• immunosenescence — aging immune decline phenotype; encompasses xenophagy failure
• inflammaging — chronic aging inflammation (cross-linked from chronic-inflammation and sasp)

Gaps surfaced:

• needs-human-replication — xenophagy flux in aged primary human macrophages not directly measured; epidemiological TB/Listeria patterns are circumstantial
• no-mechanism — TBK1 → OPTN axis in aged macrophages not characterized; IFN-γ → TBK1 pathway not mapped in senescent macrophages
• needs-replication — receptor dominance in primary human macrophages (vs. HeLa/U2OS cell lines) not established for most pathogens
• long-term-unknown — no controlled human trial of xenophagy-boosting interventions (mTOR inhibitors as TB adjunct, etc.)

ROADMAP.md updated: xenophagy marked as [x] (drafted R6e 2026-05-04, verified: false) under Round 6 Variants.

Suggested verification priority:

• High: Gutierrez 2004 (founding paper; bronze OA — downloadable; most mechanistic claims traced here)
• High: Thurston 2009 (local PDF; NDP52 quantitative data)
• High: Levine 2011 (local PDF; review — verify which claims attributed to it vs. primary sources)
• Medium: Wild 2011 (not_oa; OPTN Ser177 data — core mechanistic claim)
• Medium: Thurston 2012 (not_oa; galectin-8 data)
• Low: Tumbarello 2015, Bhatt 2009, Chua 2022 (pending; supplementary evidence)

[2026-05-04] verify | optn

• page: molecules/proteins/optn.md
• verified: true (partial scope — see verified-scope field)
• sources checked:
  • Maruyama 2010 (10.1038/nature08971) — local PDF read in full; VERIFIED
  • Wong 2014 (10.1073/pnas.1405752111) — downloaded from PMC 2026-05-04; local PDF read in full; VERIFIED
  • Rezaie 2002 (10.1126/science.1066901) — not_oa; UNVERIFIABLE
  • Wild 2011 (10.1126/science.1205405) — not_oa; UNVERIFIABLE
  • Lazarou 2015 (10.1038/nature14893) — not_oa; UNVERIFIABLE
• corrections made (Maruyama 2010):
  • “exome sequencing and linkage analysis” → “homozygosity mapping (Affymetrix SNP 500K array) and candidate-region sequencing” — method was mischaracterized; paper uses SNP-array homozygosity mapping, not whole-exome sequencing
  • Q398X: clarified mutation nomenclature (c.1502C>T, exon 12), affected subjects and families; noted co-occurrence with POAG in homozygous subject 3
  • E478G: clarified mutation nomenclature (c.1743A>G, exon 14), UBAN/DFxxER domain designation, four affected individuals in two families; confirmed NF-κB loss-of-inhibition (luciferase assay, NSC-34 cells, Maruyama Fig. 2A)
  • ~1% of familial ALS: flagged as a derived estimate not stated in the Maruyama paper; added note distinguishing the paper’s cohort data from the population-level estimate
  • Maruyama footnote n= updated: “n=ALS families” → “n=6 initial families; extended to 76 FALS + 597 SALS”
• corrections made (Wong 2014):
  • Footnote cell line: “HeLa + MEF” → “HeLa cells only” — MEF cells are not used in Wong 2014; this was an error in the seeder
  • Footnote download status: “pending” → “local PDF available at archive (downloaded 2026-05-04)”
  • E478G section: “dominant-negative-like mitophagy defect” → “loss-of-stable-recruitment allele” — the paper characterizes E478G as failing to stably associate with mitochondria, not as dominant-negative
  • Added exact quantitative data from Wong 2014 Fig. 4A and Fig. 6H/I: ~1% mito surrounded (E478G) vs ~60% (WT); ~3% LC3-positive engulfment (E478G) vs ~30% (WT); p<0.001
  • Added clarification that LIR function is tested with F178A mutant (not E478G) — important for separating UBAN vs LIR step mechanistically
  • Added that p62 dispensability in Wong 2014 independently corroborates the Lazarou 2015 claim (not_oa)
  • CCCP concentration added to footnote: 20 µM
• banner updated: ⚠️ auto-extracted → ⚠️ partially verified with scope description
• gap tags added: no-fulltext-access on Rezaie 2002, Wild 2011, and Lazarou 2015 derived claims
• downstream pages potentially affected: mitophagy, pink1-parkin-pathway (both cite Lazarou 2015 which remains not_oa; Wong 2014 p62 corroboration now added to optn.md). No corrections to propagate to those pages — the mitophagy mechanism description on optn.md still aligns with those pages’ content.
• Low: Borkhardt 1997 (10.1038/sj.onc.1200814) — not_oa; AFX/FOXO4 discovery claim widely corroborated in subsequent literature; low priority for senolytic-focused wiki use

[2026-05-04] verify | nix

• page: molecules/proteins/nix.md
• verified: true (partial scope — Hanna 2012, Novak 2010, Schweers 2007, Sandoval 2008 PDF-verified; Chen 1999 and Sowter 2001 unverifiable; UniProt identity fields not re-checked against live DB)
• sources checked:
  • Hanna 2012 (10.1074/jbc.m111.322933) — local PDF; VERIFIED
  • Novak 2010 (10.1038/embor.2009.256) — downloaded 2026-05-04; VERIFIED
  • Schweers 2007 (10.1073/pnas.0708818104) — local PDF; VERIFIED
  • Sandoval 2008 (10.1038/nature07006) — local PDF; VERIFIED
  • Chen 1999 (10.1074/jbc.274.1.7) — download failed (hybrid OA; 0 OA URLs after PMC lookup); UNVERIFIABLE no-fulltext-access
  • Sowter 2001 (PMID 11559532) — no DOI; not in archive; UNVERIFIABLE no-fulltext-access
• corrections made:
  • LIR W36–L39 → W35–L38: Novak 2010 Fig. 2A alignment shows W at position 35 in hNIX; NixW35A is the loss-of-function mutant used in both Novak 2010 and Hanna 2012. Domain table, comparison table, and limitations section all corrected.
  • Intro erythroid figures disambiguated: “28–40% vs ~1–3% WT” was a two-study conflation. Corrected to: 40% vs 1% (Schweers 2007, MitoTracker Red, total erythrocytes, Fig. 2G); 28% at 3 wks vs near-zero (Sandoval 2008, Mito+CD71+ of total RBCs).
  • Second LIR (LIR-W139/143) added: Novak 2010 characterizes a secondary LIR at W139/W143 (DW motif adjacent to BH3-like domain); ITC Kd 670 μM (LC3B) / 130 μM (LC3A). Was entirely absent from seeded page.
  • NIX–LC3 nuance corrected: “not a NIX LIR binding partner” overstated per Novak 2010 ITC (Kd 91 μM LC3B, 28 μM LC3A). Updated: GABARAP/GABARAPL1 preferred in vivo; LC3 binding is weak and not functionally dominant.
  • Novak 2010 footnote updated with ITC Kd values and dual-LIR summary.
  • Chen 1999 gap marker updated from pending to no-fulltext-access.
  • ⚠️ banner removed; verified: true, verified-date: 2026-05-04, verified-by: claude set.
• accidental stub deleted: (0 bytes at wiki root — caused link-resolution ambiguity with canonical molecules/proteins/nix.md).
• unverifiable claims: Chen 1999 cloning/structural claims (OMM topology, FAD-binding hypothesis disproof, NIX naming); Sowter 2001 HIF-1α induction quantitative claims. Both tagged no-fulltext-access.
• downstream propagation needed (for main agent):
  • bnip3 — comparison table may carry W36 for NIX LIR position; propagate W35 correction and LIR-W35/W139 distinction
  • mitophagy — NIX receptor mechanism section; check for W36 reference and “not LC3” absolutism
  • hanna-2012-bnip3-lc3-mitophagy — study page may reference NIX W36; check
  • novak-2010-nix-autophagy-receptor — study page should be created/updated with verified ITC Kd values and dual-LIR characterization

[2026-05-04] verify | ndp52

• page: molecules/proteins/ndp52.md
• verified: true (partial scope — see verified-scope field)
• sources checked:
  • Thurston 2009 (10.1038/ni.1800) — local PDF read in full; VERIFIED
  • von Muhlinen 2012 (10.1016/j.molcel.2012.08.024) — downloaded (hybrid OA) during session; local PDF read in full; VERIFIED
  • Tumbarello 2012 (10.1038/ncb2589) — downloaded (green OA / PMC) during session; local PDF read in full; VERIFIED
  • Lazarou 2015 (10.1038/nature14893) — not_oa; UNVERIFIABLE; no-fulltext-access retained on Lazarou-derived claims
• corrections made (Thurston 2009):
  • CRITICAL: galectin-8 attribution removed from Thurston 2009. Thurston 2009 contains no galectin-8 data — it covers ubiquitin-coated cytosolic bacteria. Galectin-8 is from Thurston 2012 (Nature 482:414–418; doi:10.1038/nature10744). New [^thurston2012gal8] footnote added; xenophagy section restructured to separate the two mechanisms.
  • Zinc finger domain boundary corrected: wiki said “~411–446” → source paper (von Muhlinen Fig 3A + Thurston 2009 data) indicates aa 421–446.
  • Footnote model system corrected: removed incorrect “RAW264.7 macrophages”; core experiments in 293ET cells and HeLa cells. Authors named in full from byline.
• corrections made (von Muhlinen 2012):
  • CRITICAL: CLIR sequence corrected. Wiki said “LVSIVDV, around residues 133–141.” Source paper identifies core tripeptide as Leu134-Val135-Val136 (Leu-Val-Val) within ENEEDILVVTTQGE. “LVSIVDV” does not appear in the paper.
  • Domain residue boundaries corrected: coiled-coil corrected from “~130–350” to 140–420; zinc finger corrected from “~411–446” to 421–446; “LIM-like zinc finger” descriptor removed (not used by either source paper).
  • PDB entries confirmed: 3VVV (SKICH alone, 1.35 Å) and 3VVW (NDP52 21–141 + LC3C, 2.5 Å); prior needs-canonical-id resolved.
  • CLIR structural mechanism added (flat hydrophobic patch on LC3C; Lys32, Phe33, Leu64, Phe69; CLIR β-strand).
  • Domain table restructured: separate “Linker/CLIR” row added (127–140); “LIM-like zinc finger” → “UBZ (zinc finger)” at 421–446.
• corrections made (Tumbarello 2012):
  • Tom1 mechanism inverted. Wiki said NDP52/OPTN/TAX1BP1 link autophagosomes to myosin VI “via the scaffolding protein Tom1.” Source paper: Tom1 recruits myosin VI to endosomes (WWY motif); NDP52/OPTN/T6BP recruit myosin VI to autophagosomes (RRL motif). Myosin VI bridges Tom1-positive endosomes to NDP52/OPTN/T6BP-positive autophagosomes. Section rewritten.
  • TAX1BP1/T6BP: paper uses T6BP; TAX1BP1 alias noted.
  • Model systems added: MEFs, primary cortical neurons (Snell’s Waltzer MYO6-KO mouse). Footnote updated with full author list and citation.
• frontmatter: key-domains updated; verified flipped to true; verified-date and verified-scope set.
• downstream pages to check (main agent):
  • processes/xenophagy.md — if it cites Thurston 2009 for galectin-8, correct to Thurston 2012
  • studies/thurston-2009-ndp52-xenophagy.md — ensure galectin-8 not attributed to this study page
  • studies/thurston-2012-galectin8-xenophagy.md — does not exist; create + queue doi:10.1038/nature10744 download

[2026-05-04] round-6e + round-7d follow-up summary

Closed out the deferred items from Rounds 6 and 7 in two parallel passes — 13 additional pages drafted + verified (7 R6e + 6 R7d).

Round 6e (autophagy follow-up): atg13, atg101, ndp52, optn, nix, bcl2l13, xenophagy

Round 7d (IIS follow-up): sgk1, daf-2, irs2, foxo-transcription-factors, growth-hormone, foxo4

Critical corrections caught by verifiers:

• DAF-2 UniProt fix (CRITICAL): Q967D7 was WRONG — resolves to Drosophila Turtle/TUTL. Correct C. elegans DAF-2 accession is Q968Y9. Protein length 1846 aa confirmed. Mammalian ortholog count corrected: DAF-2 is ortholog of INSR + IGF1R + IRR (three paralogs, not two), per Kimura 1997.
• ATG13 AMPK biology REVERSED: Wiki said AMPK-mediated dephosphorylation activates ATG13 → ULK1. Per Puente 2016 verified: AMPK is the DIRECT KINASE that installs Ser-224 phospho-brake (using kinase-dead ULK1 mutant + Compound C). AMPK activity actually DECREASES under AA starvation; release of pSer-224 is via reduced kinase activity, not increased. Major mechanism inversion corrected.
• ATG101 WF finger conformation INVERTED: Wiki said constitutively open. Qi 2015 PDB 5C50 shows WF finger is in CLOSED/INACTIVE conformation in human heterodimer (Trp110/Phe112 folded into hydrophobic pocket); the open projection seen in S. pombe is absent in human structure. Major structural correction.
• NDP52 galectin-8 attribution corrected: Galectin-8 SCV recognition belongs to Thurston 2012 (Nature 482), NOT Thurston 2009 (Nat Immunol). Wiki had it wrong on multiple pages; new footnote created.
• NDP52 CLIR sequence corrected: LVSIVDV was fabricated. Actual CLIR core is L134-V135-V136 within ENEEDILVVTTQGE (von Muhlinen 2012 NMR + crystal Fig 3-4). PDB 3VVV/3VVW confirmed.
• NDP52 ubiquitin-binding corrected: Wiki said CLIR motif binds ubiquitin (it binds LC3). Ubiquitin-binding domain is the C-terminal ZnF (zinc finger), NOT CLIR.
• OPTN E478G mechanism corrected: Wiki said dominant-negative-like. Wong 2014 verified: E478G is loss-of-stable-recruitment (transient ~6 min mitochondrial contacts vs WT stable rings; ~1% mitochondria surrounded vs ~60% WT, p<0.001).
• NIX LIR motif corrected: W36-V-E-L → W35-V-E-L per Novak 2010 + Hanna 2012 mutagenesis (NixW35A abolishes binding). NIX has TWO LIRs: LIR-W35 (dominant) + LIR-W139/143 (secondary, weaker — Novak 2010 ITC: LIR-W35 Kd 91 μM LC3B/28 μM LC3A; LIR-W139/143 Kd 670 μM LC3B/130 μM LC3A).
• BCL2L13 dramatically corrected (3 major errors): (1) LIR sequence WXXI → WQQI, residues 286-289 (human) / 273-276 (mouse) per Kataoka 2022 + Murakawa 2015; (2) ALL 4 BH domains required for fragmentation (not just BH1+BH4); (3) BCL2L13 does NOT bind any anti-apoptotic or pro-apoptotic BCL-2 family members (wiki incorrectly classified as ‘weak sensitizer’ with BCL-2/xL interactions).
• Xenophagy corrections: Gutierrez 2004 used RAW 264.7 (NOT J774); ESX-1/EsxA/SNARE attribution removed (not in Gutierrez 2004); TAX1BP1 ZF2 binds K63-Ub (not UBA domain); myosin VI mechanism corrected (Tom1 recruits to endosomes, NDP52/OPTN/T6BP recruit to autophagosomes — not ‘Tom1 scaffolds NDP52-myosin VI’).
• Hertweck 2004 SGK-1 +63% CONFIRMED: Independent verification on sgk1.md confirms the longevity-kinase finding from akt.md verifier batch. n=147, 14.7→24.0 days, p<0.0001. Plus akt-1 alone n=100 NS p=0.1642; akt-2 alone n=101 NS p=0.3717; akt-1;akt-2 double 17.5±0.6d ~19% extension p=0.0005 n=132.
• DAF-2 ins peptide count corrected: ~40 → exactly 37 ins genes (ins-1 through ins-37) per Pierce 2001. DAF-28 NOT in Pierce 2001 — characterized by later Li 2003 / Murphy 2003.
• Gems 1998 DAF-2 alleles refined: Class structure is 1A-1D + 2A-2E (not just Class I/II); e1370 is specifically Class 2C; max ~3-fold lifespan extension at 15°C (Gems Table 12; Kenyon’s ‘~2-fold’ is e1370 at permissive temp).
• FOXO triple-KO attribution definitively resolved: Hosaka 2004 has only single KOs (Foxo1-/- E10.5 angiogenesis defect; Foxo3-/- POF; Foxo4-/- normal). Triple KO hemangiomas/lymphomas = Paik 2007 (Cell, 100% penetrance, n=22, 9% angiosarcoma progression) + Tothova 2007 (Cell, HSC defect: ROS 2.5×, apoptosis 4×, LSK depletion 4.6× p<0.0001).
• FOXO4-DRI peptide corrections: Domain residues 88-191 → 86-206 per Baar 2017 STAR Methods; mouse models corrected (XpdTTD/TTD fast-aging model + p16::3MR aged + dox toxicity — three distinct arms, NOT one model); endpoints: plasma urea (primary) + grip strength removed (not measured); UniProt phospho sites Thr32/Ser197/Ser262 confirmed (brief had Thr28/Ser193/Ser258 wrong).
• IRS-2 inhibitory serine sites corrected: Wiki had Ser-303/Ser-675 (wrong — those are IRS-1 sites or unattributed). Per Boucher 2014 verified: IRS-2 mouse sites are Thr-347, Ser-484, Ser-488. KRLB term attribution to Sun 1995 is also wrong (term coined by Sawka-Verhelle 1996/1997).
• Multiple BUG-2 DOI corrections caught (~10+ across both rounds, including: Mercer 2009.8504→.8249, Suzuki 2015 nsmb.3162→.3036, von Muhlinen 2012.04.014→.08.024, Wong 2014 ncomms3537→pnas.1405752111, Levine 2011 cell.2010.12.024→nature09782, Borkhardt 1997 Blood→Oncogene.1200814, Niall 1971 PMID 5159922→DOI 10.1073/pnas.68.4.866, Kataoka 2001 curr-biol→jbc, Paik 2007.2007.01.034→.2006.12.029).

Cleanup: Deleted accidental empty (0 bytes) at wiki root.

Coverage delta: Rounds 5/6/7 + follow-ups = 53 new pages drafted + verified in this session. Tier A/B/C/D implicit-stub queue substantially closed for apoptosis, autophagy/mitophagy, and PI3K/AKT/IIS clusters. Round 6 + Round 7 are now fully closed — no remaining “deferred to follow-up” items.

[2026-05-04] round-6e + round-7d follow-up summary

Closed out the deferred items from Rounds 6 and 7 in two parallel passes — 13 additional pages drafted + verified (7 R6e + 6 R7d).

Round 6e (autophagy follow-up): atg13, atg101, ndp52, optn, nix, bcl2l13, xenophagy

Round 7d (IIS follow-up): sgk1, daf-2, irs2, foxo-transcription-factors, growth-hormone, foxo4

Critical corrections caught by verifiers:

• DAF-2 UniProt fix (CRITICAL): Q967D7 was WRONG — resolves to Drosophila Turtle/TUTL. Correct C. elegans DAF-2 accession is Q968Y9. Protein length 1846 aa confirmed. Mammalian ortholog count corrected: DAF-2 is ortholog of INSR + IGF1R + IRR (three paralogs, not two), per Kimura 1997.
• ATG13 AMPK biology REVERSED: Wiki said “AMPK-mediated dephosphorylation activates ATG13 → ULK1.” Per Puente 2016 verified: AMPK is the DIRECT KINASE that installs Ser-224 phospho-brake (using kinase-dead ULK1 mutant + Compound C). Major mechanism inversion corrected.
• ATG101 WF finger conformation INVERTED: Wiki said “constitutively open.” Qi 2015 PDB 5C50 shows WF finger is in CLOSED/INACTIVE conformation in human heterodimer; the open projection seen in S. pombe is absent in human structure.
• NDP52 galectin-8 attribution corrected: Galectin-8 SCV recognition belongs to Thurston 2012 (Nature 482), NOT Thurston 2009. NDP52 ubiquitin-binding domain is C-terminal ZnF, NOT CLIR. CLIR sequence “LVSIVDV” was fabricated; actual is L134-V135-V136 within ENEEDILVVTTQGE per von Muhlinen 2012.
• OPTN E478G mechanism corrected: Wiki said “dominant-negative-like.” Wong 2014 verified: E478G is loss-of-stable-recruitment (transient ~6 min mitochondrial contacts vs WT stable rings; ~1% mitochondria surrounded vs ~60% WT, p<0.001).
• NIX LIR motif corrected: W36-V-E-L → W35-V-E-L per Novak 2010 mutagenesis. NIX has TWO LIRs: LIR-W35 (dominant) + LIR-W139/143 (secondary).
• BCL2L13 dramatically corrected: (1) LIR sequence WXXI → WQQI residues 286-289 (human); (2) ALL 4 BH domains required for fragmentation (not BH1+BH4); (3) BCL2L13 does NOT bind any anti-apoptotic or pro-apoptotic BCL-2 family members.
• Xenophagy corrections: Gutierrez 2004 used RAW 264.7 (NOT J774); ESX-1/EsxA/SNARE attribution removed; TAX1BP1 ZF2 binds K63-Ub (not UBA); myosin VI mechanism corrected.
• Hertweck 2004 SGK-1 +63% CONFIRMED: Independent verification on sgk1.md. n=147, 14.7→24.0 days, p<0.0001.
• DAF-2 ins peptide count corrected: ~40 → exactly 37 ins genes per Pierce 2001. DAF-28 NOT in Pierce 2001.
• Gems 1998 DAF-2 alleles refined: Class structure is 1A-1D + 2A-2E; e1370 is specifically Class 2C; max ~3-fold lifespan extension at 15°C.
• FOXO triple-KO attribution definitively resolved: Hosaka 2004 has only single KOs. Triple KO hemangiomas/lymphomas = Paik 2007 + Tothova 2007.
• FOXO4-DRI peptide corrections: Domain residues 86-206 (not 88-191); mouse models XpdTTD/TTD + p16::3MR + dox toxicity (three distinct); endpoints plasma urea (grip strength NOT measured); UniProt phospho sites Thr32/Ser197/Ser262 confirmed.
• IRS-2 inhibitory serine sites corrected: Thr-347, Ser-484, Ser-488 (not Ser-303/Ser-675 which were IRS-1 sites). KRLB term not coined by Sun 1995.
• Multiple BUG-2 DOI corrections caught (~10+ across both rounds).

Cleanup: Deleted accidental empty nix-md (0 bytes) at wiki root.