🧬 Aging Wiki

R5

47 min read

log/R5.md — Round 5 entries

Sub-file of log — see parent for index.

[2026-05-04] verify | processes/apoptosis.md

Pages verified: 1 (full — 4 sources read end-to-end as local PDFs; 1 confirmed present only)

Sources checked:

  • Kerr 1972 (10.1038/bjc.1972.33) — downloaded this pass (PMC); full 19 pages read
  • Galluzzi 2018 (10.1038/s41418-017-0012-4) — downloaded this pass (PMC); full 56 pages read
  • Bergmann 2015 (10.1016/j.cell.2015.05.026) — local PDF; all 11 pages read
  • Yosef 2016 (10.1038/ncomms11190) — local PDF; all 11 pages read
  • Lopez-Otin 2013 (10.1016/j.cell.2013.05.039) — local PDF confirmed; no detailed claims on this page; not read end-to-end

Corrections applied (10 substantive):

  1. Bergmann 2015 footnote n/age-range: “n=36, age 19-74 yr” → “n=29 stereology (1mo-73yr)/n=51 C14 (8-75yr)”
  2. CM number trajectory: “constant from birth through early adulthood, then declines” → “constant throughout adult lifespan (R=0.01, p=0.96)”
  3. IMR90 SCAP: “BCL-2+BCL-xL+BCL-w triple, all three required” → “BCL-W+BCL-XL dual; BCL-2 minor/additive only (ABT-199 alone does not kill DIS IMR-90 cells)”
  4. Preadipocyte SCAP row attribution: Yosef 2016 → Zhu 2015 (Yosef 2016 does not study preadipocytes)
  5. Kerr 1972 etymology: “falling off, as leaves fall from a tree” → source verbatim: “dropping off or falling off of petals from flowers, or leaves from trees”
  6. Added Kerr 1972 two-stage morphological description to Discovery section
  7. Necroptosis executioner table row: “RIPK3 → MLKL” → “RIPK1 → RIPK3 → MLKL”
  8. Pyroptosis executioner: “Caspase-1/4/5/11 → Gasdermin D” → “inflammatory caspases → gasdermin family (predominantly GSDMD)” (NCCD 2018: not always inflammatory-caspase-activated; family not limited to GSDMD)
  9. Yosef 2016 compound in footnote: navitoclax → ABT-737 (ABT-737 is BCL-2/BCL-W/BCL-XL; navitoclax is ABT-263)
  10. Bergmann 2015 limitations description: “longitudinal” → correctly characterized as cross-sectional retrospective birth-dating

Unverifiable:

  • SN dopaminergic neuron ~50% lifetime loss — unsourced retained
  • Caloric restriction apoptosis mechanism — no-mechanism retained

Downstream propagation needed:

  • apoptosis-pathway.md — check IMR90 SCAP triple claim consistency
  • senolytics.md — check IMR90 SCAP table

[2026-05-04] verify | molecules/compounds/a1331852.md

Pages verified: 1 (partial scope — Leverson 2015 PDF unavailable)

Sources checked:

  • Koehler 2014 (10.1021/ml500030p) — downloaded and verified end-to-end (PMC PDF)
  • Zhu 2016 (10.1111/acel.12445) — local PDF, verified end-to-end
  • Yosef 2016 (10.1038/ncomms11190) — local PDF, verified pages 1–6
  • Khan 2019 (10.1038/s41591-019-0668-z) — downloaded and verified pages 1–5
  • Leverson 2015 (10.1126/scitranslmed.aaa4642) — download failed; tagged no-fulltext-access
  • ChEMBL API verified: CHEMBL3793424 correct; CHEMBL3753332 wrong (potassium polysulfide)

Key corrections:

  1. Zhu 2016 agent: navitoclax (ABT-263), not A-1331852 — all senolytic cell-type data (HUVECs, IMR90, MEFs, preadipocytes) reattributed to navitoclax
  2. Yosef 2016 agent: ABT-737, not A-1331852 — all Yosef 2016 → A-1331852 senolytic attribution removed; in vivo model clarified (DNA damage lung + p14ARF skin senescence, ABT-737)
  3. DT2216 warhead: ABT263/navitoclax BCL-2/BCL-xL binding moiety (per Khan 2019) — not A-1331852; DC50 63 nM, Dmax 90.8% added; PROTAC section rewritten
  4. Koehler 2014: does not report A-1331852 Ki values — reports SAR scaffold series precursors only; Ki table tagged needs-replication
  5. ChEMBL CHEMBL3753332 confirmed wrong; CHEMBL3793424 confirmed correct; needs-canonical-id removed
  6. Leverson 2015 claims tagged no-fulltext-access throughout
  7. verified: true (partial scope); banner removed

Downstream pages that may need updates:

  • bcl-xl — check for inherited A-1331852 senolytic efficacy claims from Zhu 2016/Yosef 2016
  • bax — “primary agent A1331852/A1155463” claim in ROADMAP sourced to Zhu 2017 (separate paper); confirm Zhu 2017 DOI is distinct from Zhu 2016
  • senolytics — may cite A-1331852 data from Zhu 2016/Yosef 2016; propagation needed

[2026-05-04] verify | molecules/proteins/caspase-8.md

Pages verified: 1

  • molecules/proteins/caspase-8.md — corrections: 8 substantive; set verified: true (partial scope — UniProt identity fields and PTM primary sources not independently re-checked)

Sources checked (all 6 PDFs read end-to-end):

  • Muzio 1996 (10.1016/s0092-8674(00)81266-0) — local PDF; FLICE/CASP8 = CAP4, 479 aa, ~55 kDa, FADD co-IP identification; active-site QACQG pentapeptide by ICE crystal-structure homology; granzyme B cleaves to p20+p10
  • Boldin 1996 (10.1016/s0092-8674(00)81265-9) — local PDF; MACH α1 = 479 aa; His317 + Cys360 explicitly labeled in Figure 2A sequence alignment; Asp216 + Asp374 cleavage sites confirmed; Cys360S mutant used (not Asp mutant)
  • Scaffidi 1998 (10.1093/emboj/17.6.1675) — downloaded during this pass; type I = SKW6.4 + H9; type II = CEM + Jurkat (Jurkat is type II, not type I as wiki had)
  • Varfolomeev 1998 (10.1016/s1074-7613(00)80609-3) — downloaded during this pass; Immunity 9:267–276; lethality phenotype apparent at E11.5 (no abnormalities at E10.5); impaired heart muscle + congested erythrocytes; fibroblasts resistant to anti-Fas/TNF/DR3 but not UV/etoposide/staurosporine/ceramide/serum starvation/VSV
  • Oberst 2011 (10.1038/nature09852) — local PDF; Nature 471:363–367 (not 480:273 as wiki had); caspase-8^CA mutant is C360A (not C360S); FLIP_L = CFLAR_L
  • Kaiser 2011 (10.1038/nature09857) — local PDF; Nature 471:368–372 (not 480:269 as wiki had); lethality E10.5–E11.5 (not “~E10.5”); DKO mice viable + fertile confirmed; immunocompetent but accumulate abnormal T cells with age

Key corrections:

  1. Jurkat misclassified as type I in extrapolation table — Scaffidi 1998 uses Jurkat as the prototypical type II cell throughout; corrected to SKW6.4+H9 (type I) vs CEM+Jurkat (type II)
  2. Oberst 2011: Nature 480(7376):273–277 → Nature 471:363–367 (volume and pages both wrong)
  3. Kaiser 2011: Nature 480(7376):269–273 → Nature 471:368–372 (volume and pages both wrong)
  4. Active-site mutant: “C360S” → “C360A (caspase-8^CA)” per Oberst 2011 methods
  5. Embryonic lethality: “~E10.5” → “E10.5–E11.5” per Kaiser 2011 abstract
  6. Varfolomeev defect description: “cardiac malformations and erythroblast abundance in the bloodstream” → “impaired heart muscle development and congested accumulation of erythrocytes” (paper’s exact language)
  7. TNF treatment: “TNF + cycloheximide” → “human TNF” alone (paper’s fibroblast assay did not use cycloheximide)
  8. Bogus “Bhatt V” author triplicates removed from varfolomeev1998 footnote (auto-extraction artifact)
  9. Scaffidi 1998 footnote: “MCF-7” removed as main cell line (it is not the primary cell line in that paper); status updated from “pending download” to “locally available”
  10. BUG-2 note removed from Limitations (DOI corrections already applied; note is now historical)

Downstream propagation needed (main agent):

  • No other wiki pages currently cite [^scaffidi1998] or reference Jurkat as type I
  • Any page citing Oberst 2011 or Kaiser 2011 by journal/volume should have those citations checked (no other pages currently contain “Nature 480” for these papers based on the caspase-8 page being the only one that has them)

[2026-05-04] verify | molecules/proteins/caspase-9.md

Pages verified: 1

  • molecules/proteins/caspase-9.md — corrections: 6 substantive; set verified: true (partial scope — Salvesen 1999 and Cardone 1998 not_oa; UniProt identity fields not live-checked)

Sources checked:

  • Li 1997 (10.1016/s0092-8674(00)80434-1) — local PDF, verified end-to-end
  • Srinivasula 1998 (10.1016/s1097-2765(00)80095-7) — downloaded 2026-05-04, verified end-to-end
  • Renatus 2001 (10.1073/pnas.231465798) — downloaded 2026-05-04, verified end-to-end
  • Hakem 1998 (10.1016/s0092-8674(00)81477-4) — local PDF, verified end-to-end
  • Kuida 1998 (10.1016/s0092-8674(00)81476-2) — local PDF, verified end-to-end
  • Salvesen & Dixit 1999 (10.1073/pnas.96.20.10964) — not_oa; unverifiable
  • Cardone 1998 (10.1126/science.282.5392.1318) — not_oa; unverifiable

Key corrections:

  1. Li 1997 model system corrected: “Xenopus egg extract + recombinant human proteins” → “HeLa S-100 cytosolic extract + purified recombinant APAF-1, caspase-9 (APAF-3), caspase-3 + dATP”. Li 1997 uses HeLa fractionation, not Xenopus.
  2. Apoptosome architecture (7-spoke wheel, ~280 Å) attribution corrected: this structural description is NOT in Li 1997, which only demonstrated functional reconstitution. Attributed to subsequent cryo-EM work (Acehan 2002, Riedl 2005); tagged needs-replication with secondary citations noted.
  3. Kuida 1998 intrinsic triggers corrected: “staurosporine or dexamethasone” → “etoposide, dexamethasone, and gamma-irradiation”. Staurosporine is not mentioned in Kuida 1998; the tested intrinsic triggers in thymocytes were etoposide, dexamethasone, and gamma-irradiation.
  4. Hakem 1998 lethality and phenotype expanded: “E16.5 embryonic lethal; brain overgrowth” → “perinatal lethal; brain malformation (protrusion, ventricular stenosis, heterotopias)”; additional phenotypes in ES cells/MEFs added from paper.
  5. Kuida 1998 lethality expanded: added onset of brain malformation at ~E12.5 and <3% postnatal survival figure from Table 1.
  6. Cys287 numbering note added: Renatus 2001 uses caspase-1 numbering convention (Cys-285) for the same residue called Cys-287 in UniProt P55211 numbering; clarified in both domain structure section and Renatus footnote.

Downstream pages that may need review:

  • Any page citing caspase-9 for apoptosome architecture (check if 7-spoke / 280 Å claim appears elsewhere)
  • apoptosis-pathway — may inherit the HeLa vs Xenopus model system description
  • apaf-1 — may describe the Li 1997 reconstitution system

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

Pages verified: 1

  • molecules/proteins/bim.md — corrections: 8 substantive; set verified: true (partial scope — Bouillet 1999 not_oa; canonical-DB identity fields not live-checked)

Sources checked:

  • O’Connor 1998 (10.1093/emboj/17.2.384) — downloaded 2026-05-04 (PMC via bronze OA), verified end-to-end
  • Dijkers 2000 (10.1016/s0960-9822(00)00728-4) — downloaded 2026-05-04, verified end-to-end
  • Chen 2005 (10.1016/j.molcel.2004.12.030) — downloaded 2026-05-04 (previously pending, now available), verified end-to-end
  • Ley 2003 (10.1074/jbc.m301010200) — downloaded 2026-05-04 (hybrid OA), verified end-to-end
  • Opferman 2003 (10.1038/nature02067) — local PDF, verified end-to-end
  • Gavathiotis/Walensky 2010 (archived as “Chipuk 2010” DOI 10.1016/j.molcel.2010.10.019) — local PDF, verified end-to-end; paper identity corrected
  • Vogler 2025 (10.1038/s41392-025-02176-0) — local PDF, verified end-to-end
  • Bouillet 1999 (10.1126/science.286.5445.1735) — not_oa; unverifiable; claims remain tagged no-fulltext-access

Key corrections:

  1. Chipuk 2010 DOI misidentified: 10.1016/j.molcel.2010.10.019 resolves to Gavathiotis/Walensky 2010 (structural BAX NMR study, Mol Cell 40:481–492), NOT a Chipuk review. Citation label corrected in footnote and body text. Fig 1C of that paper is about BAX tethered-loop conformation, not the direct activator/sensitizer framework.
  2. ERK phosphorylation sites removed: Ser69/Ser77/Ser87 NOT stated in Ley 2003 (paper shows phosphorylation by gel shift but does not name specific serine residues). TRIM2 attribution to Ley 2003 is wrong (paper does not name any E3 ligase). Sites updated to Ser65/Ser55/Ser73 per Vogler 2025; uncertainty noted throughout. key-ptms frontmatter corrected accordingly.
  3. DLC1 attribution corrected: O’Connor 1998 shows cytoskeletal/membrane BIM localization but does not identify DLC1 as the binding partner — that attribution belongs to Puthalakath et al. 1999 (Mol Cell 3:287–296), which is not in the footnotes.
  4. Dijkers 2000 cell model corrected: Ba/F3 mouse pro-B cells + mouse fetal liver cells (not “primary human T cells” as stated in the original footnote). Ser315 removed from Dijkers 2000 attribution — paper only documents Thr32 and Ser253 phosphorylation sites.
  5. BH3 affinity table corrected: Chen 2005 IC50 values added from Fig 3A (BCL-2: 2.6 nM, BCL-xL: 4.6 nM, BCL-w: 4.3 nM, MCL-1: 2.4 nM, A1: 4.3 nM). Methodological incompatibility with Opferman 2003 FP Kd (74 nM for MCL-1) explained — these are different assay types, not contradictory values. Prior claim that BIM has “substantially lower” affinity for MCL-1 than for BCL-2/BCL-xL was based on conflating these two different assays.
  6. BIM-S potency corrected: O’Connor 1998 Table I and colony suppression data show BIM-S is the MOST potent isoform; BIM-EL is the LEAST potent (not BIM-S as smallest = least active as implied by prior table).
  7. Isoform sizes added from O’Connor 1998 Fig 1B: human BIM-EL 198 aa, BIM-L 138 aa, BIM-S 110 aa (mouse values: 196/140/110).
  8. FOXO3 axis diagram: Ser315 removed from Dijkers 2000 attribution; text notes this comes from other studies.

Pages unverifiable (closed-access):

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

  • mcl-1 — the affinity-method caveat (Chen 2005 IC50 2.4 nM vs Opferman 2003 FP 74 nM) should be added; the mcl-1.md verifier previously treated 74 nM as the single definitive figure without noting the methodological distinction
  • bak — cites “Chipuk 2010” by DOI 10.1016/j.molcel.2010.10.019 for PUMA classification; that DOI is Gavathiotis/Walensky 2010; citation identity correction needed
  • apoptosis-pathway — same Chipuk DOI issue likely applies; BIM-S potency claim and DLC1 attribution may be inherited
  • bcl-2-family-signaling — affinity table for BIM vs anti-apoptotic proteins may inherit the conflated values

[2026-05-04] verify | molecules/compounds/navitoclax.md

Pages verified: 1

  • molecules/compounds/navitoclax.md — verified-partial (6/7 sources; Chang 2016 permanently unverifiable)

Sources checked:

  • Tse 2008 (10.1158/0008-5472.CAN-07-5836) — newly downloaded; PDF read end-to-end; VERIFIED
  • Zhu 2016 (10.1111/acel.12445) — local PDF; read end-to-end; VERIFIED
  • Yosef 2016 (10.1038/ncomms11190) — local PDF; read end-to-end; VERIFIED
  • Roberts 2012 (10.1200/JCO.2011.34.7898) — newly downloaded PDF is supplementary appendix only; key numbers cross-checked against Crossref abstract; PARTIAL
  • Roberts 2016 (10.1056/NEJMoa1513257) — local PDF; read end-to-end; VERIFIED
  • He 2020 (10.1038/s41467-020-15838-0) — newly downloaded; PDF read end-to-end; VERIFIED
  • Chang 2016 (10.1038/nm.4010) — PERMANENTLY UNVERIFIABLE; 0 OA candidate URLs; no-fulltext-access

Corrections made (7 substantive):

  1. BCL-2 Ki: <0.5 nM → ≤1 nM (Tse 2008 Table 1; BCL-xL ≤0.5 nM; BCL-2 and BCL-w are ≤1 nM; A1 IC50 = 354 nM not “>1 µM”)
  2. MCL-1/A1 values: reclassified from Ki to IC50 per Tse 2008 Table 1 footnote
  3. Yosef 2016 attribution: paper uses ABT-737 not navitoclax; A1331852 is NOT in Yosef 2016; BCL-W+BCL-xL ~53% dual-knockdown senolysis confirmed
  4. He 2020 senolytic window: ~5× → >37-fold (paper’s explicit statement; EC50 ratio PLTs/IR-SCs: PZ=10.3 vs ABT-263=0.28); PROTAC = PZ15227, CRBN E3 ligase, IP-only
  5. Roberts 2016: n=107/17p-del Phase III → n=116 Phase I CLL/SLL (ORR 79%; CR 20%); not Phase III
  6. Roberts 2012: n enrolled = 29 added to footnote; DLT peripheral-destruction mechanism confirmed
  7. Preadipocyte SCAP: “ephrins/PI3Kδ/p21/PAI-2 [Zhu 2016]” → corrected; Zhu 2016 names only “tyrosine kinase pathways”; specific components are from Zhu 2015

Pages unverifiable (permanently blocked): Chang 2016 — no-fulltext-access

Downstream propagation needed:

  • molecules/proteins/bcl-xl.md — check A1331852 attribution to Yosef 2016 (A1331852 not in Yosef 2016)
  • molecules/compounds/a1331852.md — correct “Yosef 2016 senolytic activity” attribution for A1331852
  • studies/roberts-2016-venetoclax-cll.md — correct n (107→116) and study type (Phase II→Phase I)
  • studies/yosef-2016-bcl-w-bcl-xl-senescence.md — note ABT-737 vs navitoclax distinction

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

Pages verified: 1

  • molecules/proteins/noxa.md — corrections: 10 substantive (see below); set verified: true (partial scope — Villunger 2003 not_oa; Shahbandi 2020 download failed; Hijikata 1990 pre-DOI unresolvable; UniProt identity fields not live-checked)

Sources checked:

  • Oda 2000 (10.1126/science.288.5468.1053) — local PDF, verified end-to-end
  • Chen 2005 (10.1016/j.molcel.2004.12.030) — downloaded 2026-05-04 (bronze OA via camoufox), verified end-to-end
  • Shibue 2003 (10.1101/gad.1103603) — downloaded 2026-05-04 (diamond OA), verified end-to-end
  • Tyner 2002 (10.1038/415045a) — local PDF, previously verified; invoked here as indirect evidence only; no corrections needed
  • Villunger 2003 (10.1126/science.1090072) — not_oa; unverifiable; claims remain tagged no-fulltext-access
  • Shahbandi 2020 (10.1038/s41418-020-0564-6) — download failed (bronze OA); unverifiable; claim remains tagged no-fulltext-access

Key corrections:

  1. p53RE coordinates: “−172 to −151” → “−155 to −174” (murine Noxa promoter per Oda 2000 Fig 2A); human PMAIP1 p53RE coordinates not stated in Oda 2000 (tagged gap)
  2. Murine Noxa protein length: added clarification — mouse Noxa = 103 aa with TWO BH3 motifs (A and B); human NOXA = 54 aa with ONE BH3 motif (residues 29–37); these are structurally distinct orthologs
  3. Shibue 2003 oncogene model: “c-Myc-induced apoptosis” → “E1A (adenoviral oncoprotein)-induced apoptosis” — Shibue 2003 used E1A, not c-Myc
  4. Shibue 2003 quantitative survival data added: Noxa−/− mean survival 17.5±9.5 d vs WT 9.1±3.3 d (P<0.01), p53−/− 10.7±7.9 d (P<0.05); intestinal ESS apoptosis suppressed ~2-fold
  5. Shibue 2003 thymocyte correction: Noxa−/− thymocytes undergo radiation-induced apoptosis as efficiently as WT — NOXA is dispensable in thymocytes (contradicts the prior “NOXA contributes to lymphoid progenitor apoptosis” claim from unverified Villunger 2003 abstract)
  6. PUMA MCL-1 affinity: “Moderate” → “High (5.0 nM)” per Chen 2005 Fig 3A; PUMA binds all five prosurvival proteins at similar high affinity (~3–7 nM)
  7. PUMA BCL-w affinity: “Moderate” → “High (5.1 nM)” per Chen 2005 — also corrected in complementarity tables
  8. Rapid proteasomal turnover of NOXA: removed false attribution to Oda 2000 — this mechanism is not in that paper; tagged unsourced
  9. MCL-1 co-degradation claim: removed false attribution to Oda 2000 — not in that paper; tagged unsourced; primary citation needed
  10. Hijikata 1990 journal: “Molecular and Cellular Biology” and DOI 10.1128/mcb.10.6.2879 → “Journal of Virology 64(10):4632–4641” per Oda 2000 ref 25; old DOI was wrong
  11. Proteasome-inhibitor-driven NOXA induction: removed false citation to Oda 2000 — that paper does not discuss this mechanism; tagged unsourced
  12. A1 affinity gap resolved: Chen 2005 IC50 = 180 nM (vs MCL-1 = 24 nM)
  13. Discovery method: “microarray-based screening” → “mRNA differential display” per Oda 2000 Methods
  14. Added quantitative IC50 values throughout binding spectrum tables (from Chen 2005 Fig 3A and Fig 1C)

Pages unverifiable (closed-access or failed):

Downstream pages that may need updates:

  • puma — PUMA MCL-1/BCL-w affinity claims may be sourced from Chen 2005; verify consistency with “High” (not “Moderate”) characterization
  • mcl-1 — any claim about NOXA MCL-1-co-degradation mechanism sourced without primary citation should be re-checked
  • Any study page for Villunger 2003 if it exists (check with find -name "*villunger*")

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

Pages verified: 1

  • molecules/proteins/bad.md — corrections: 8 substantive (see below); set verified: true (partial scope)

Sources checked:

  • Datta 1997 (10.1016/s0092-8674(00)80405-5) — local PDF, verified end-to-end
  • Zha 1996 (10.1016/s0092-8674(00)81382-3) — local PDF, verified end-to-end
  • Harada 1999 (10.1016/s1097-2765(00)80469-4) — downloaded 2026-05-04, verified end-to-end
  • Yang 1995 (10.1016/0092-8674(95)90411-5) — download failed (bronze OA, no PMC copy); DOI confirmed correct via DOI lookup title; unverifiable from PDF

Key corrections:

  1. Datta 1997 footnote cell lines: “NIH3T3” → “Balb/c 3T3, HEK293, COS-7” (NIH3T3 not used in that paper)
  2. Datta 1997 neuron model clarified: rat cerebellar granule neurons DIV 4-5 (not vague “primary neurons”)
  3. Ser136 kinase table: removed RSK, CaMKII, p70S6K as co-equal Ser136 kinases — Datta 1997 shows AKT is the primary in-vivo Ser136 kinase; p70S6K explicitly ruled out (rapamycin does not block); RSK/CaMKII are speculative in-vitro observations only (Datta Discussion); AKT does not significantly phosphorylate Ser-112 in vivo per Datta 1997
  4. Harada 1999 model system: “COS7/293T” → FL5.12 BCL-xL/BAD cells (IL-3-dependent murine pro-B hematopoietic line), same model as Zha 1996
  5. Ser155/Ser118 (BH3-motif PKA site): de-attributed from Harada 1999 — that paper is exclusively about Ser112 and does not address Ser155; tagged needs-primary-source throughout
  6. Zha 1996 “direct quote” converted to paraphrase (original text was paraphrase in quotation marks)
  7. 14-3-3 binding mechanism: corrected from “via phospho-Ser136” to “via either Ser-112 OR Ser-136 alone — double S112A/S136A mutant required to abolish binding” (Zha 1996 Fig 7B)
  8. Harada 1999 PKA paragraph expanded with actual experimental evidence: AKT immunodepletion result; H-89/Rp-cAMP block Ser112 not Ser136; Ht31 AKAP-disrupting peptide ~50% reduction; kinase identity by mass spec + Western

Pages unverifiable:

  • Yang 1995 — tagged no-fulltext-access in Limitations section and footnote; DOI 10.1016/0092-8674(95)90411-5 confirmed correct

Downstream pages that may need review:

  • Any page citing BAD Ser-155/Ser-118 as a PKA phosphorylation site should be flagged needs-primary-source
  • bcl-2-family-signaling.md — if it repeats kinase assignments for BAD phospho-sites, cross-check against corrected table
  • apoptosis-pathway.md — if it describes BAD regulation mechanism, verify 14-3-3 binding is correctly attributed to either-site (not Ser136 only)

[2026-05-04] verify | molecules/proteins/cytochrome-c

Page verified: molecules/proteins/cytochrome-c.md

Sources checked (5 verified / 2 unverifiable):

  • Liu 1996 (10.1016/s0092-8674(00)80085-9) — local PDF, verified
  • Kluck 1997 (10.1126/science.275.5303.1132) — local PDF, verified
  • Goldstein 2000 (10.1038/35004029) — local PDF, verified
  • Bossy-Wetzel 1998 (10.1093/emboj/17.1.37) — downloaded and verified
  • Frezza 2006 (10.1016/j.cell.2006.06.025) — downloaded and verified
  • Yang 1997 (10.1126/science.275.5303.1129) — not_oa; unverifiable; no-fulltext-access
  • Kagan 2005 (10.1038/nchembio727) — not_oa; unverifiable; no-fulltext-access

Corrections made:

  • Residue numbering convention inverted → Cys15/Cys18/His19/Met81 are UniProt precursor numbers (confirmed via P99999 REST API); mature-form equivalents are Cys14/Cys17/His18/Met80 (opposite of what the draft stated)
  • Met1 processing note corrected: “cleavage of initiator Met after N-terminal acetylation of Gly2” → “removal of Met1; N-terminal acetylation of Gly2 in mature form”
  • Goldstein 2000 release duration: “exits in under a minute” → “within approximately 5 minutes once release initiates” (paper Discussion: “the duration of cytochrome c release remains roughly 5 min in all cases”; Fig 3c mean bars all ~5–7 min)
  • Liu 1996 mechanism attribution: removed “displacing an auto-inhibitory interaction” (not in Liu 1996 — structural mechanism established post-1996); clarified cytochrome c was identified as APAF-2; noted ATP substitution established by later work, not Liu 1996
  • Frezza 2006 “necessary and sufficient / independent of BAX/BAK”: overstated → corrected to: OPA1 controls cristae pool mobilisation; BAX/BAK OMM permeabilisation is still required; OPA1 does not interfere with BAX/BAK activation (paper explicitly states this)
  • Kluck 1997 footnote cell line: “FL5.12” → “CEM cells” (FL5.12 is Yang 1997’s system)

Unverifiable claims (2 sources):

  • Yang 1997: BCL-2/BCL-xL block cytochrome c release upstream of caspase activation — plausible per Kluck 1997 which confirms same finding, but Yang 1997 itself not read
  • Kagan 2005: cardiolipin oxidation peroxidase loop details and quantitative estimates — not independently verified

Downstream propagation needed:

  • studies/liu-1996-cytochrome-c-apoptosis.md — if this study page exists, check whether it correctly describes dATP (not ATP) specificity and does not attribute the WD40/auto-inhibitory mechanism to Liu 1996
  • Any page citing [[cytochrome-c]] with the “under a minute” release claim should be updated to “~5 minutes”
  • molecules/proteins/opa1.md (if it exists) — check whether the Frezza 2006 cristae claims propagated correctly

Final verified state: verified: true (partial scope — Yang 1997 and Kagan 2005 unverifiable; identity fields not re-queried)

[2026-05-04] verify | molecules/proteins/apaf-1

Pages verified: 1

  • molecules/proteins/apaf-1.md — corrections: 7 (see below); set verified: true (partial scope)

Sources checked:

  • Zou 1997 (10.1016/s0092-8674(00)80501-2) — local PDF, verified
  • Acehan 2002 (10.1016/s1097-2765(02)00442-2) — PDF downloaded during verification, verified
  • Cecconi 1998 (10.1016/s0092-8674(00)81732-8) — local PDF, verified
  • Yoshida 1998 (10.1016/s0092-8674(00)81733-x) — local PDF, verified
  • Soengas 2001 (10.1038/35051606) — local PDF, verified
  • Yuan 1992 (10.1242/dev.116.2.309) — not_oa, unverified
  • Honarpour 2000 (10.1006/dbio.1999.9585) — not_oa, unverified; DOI confirmed correct via Crossref

Key corrections made:

  1. Mass “~1.4 MDa” → “~1 MDa” (Acehan 2002 states “calculated mass of 1 MDa”)
  2. Acehan 2002 particle count added: ~3,400 particles; resolution confirmed as 27 Å (not “~27 Å”)
  3. Cytochrome c location on apoptosome: “peripheral… spokes” → “bridge region between two WD40 β-propellers in the Y domain” per Acehan 2002 modeling
  4. Reconstitution (Zou 1997): clarified that caspase-9 was “Apaf-3” (crude fraction) in Zou 1997, not isolated caspase-9
  5. Lethality (Cecconi 1998): replaced vague “perinatal lethality in majority” with precise embryonic lethality by e16.5; expanded phenotype detail from source
  6. Yoshida 1998: expanded to include Fas-independence finding and cytochrome c release finding
  7. Soengas 2001: corrected “CpG methylation of the APAF-1 promoter” → methylation excludes core promoter CpG island; mechanism involves enhancer/regulatory region outside core promoter; corrected LOH frequency (>40% of 24 tumor pairs = 10/24) vs cell-line negative rate (10/19 = 53%)

Pages unverifiable (closed-access):

  • Yuan 1992 — tagged in verified-scope; CED-4 ortholog claim consistent with Zou 1997 discussion
  • Honarpour 2000 — tagged in verified-scope; DOI.9585 confirmed correct

Downstream pages that may need review:

  • studies/zou-1997-apaf1-cytochrome-c-caspase — if exists, check reconstitution description
  • studies/soengas-2001-apaf1-melanoma — if exists, update methylation mechanism description
  • studies/cecconi-1998-apaf1-knockout — if exists, update lethality timing

[2026-05-04] ingest | round-5d-bcl2-family-signaling

  • added: pathways/bcl-2-family-signaling.md (type: pathway, verified: false)
  • Canonical naming decision: bcl-2-family-signaling declared canonical over bcl2-family-pathway for two reasons: (a) hyphenation matches existing bcl-2.md and bcl-xl.md filenames in molecules/proteins/; (b) the page covers the decision-module signaling network rather than the full apoptosis pathway (which is apoptosis-pathway). The alias bcl2-family-pathway (3 inbound) is included in frontmatter so existing wikilinks resolve transparently.
  • Canonical DB IDs confirmed: KEGG hsa04210 (Apoptosis — broader; best available); Reactome R-HSA-109606 (Intrinsic Pathway for Apoptosis, confirmed via API with diagram); Reactome R-HSA-111453 (BH3-only proteins associate with and inactivate anti-apoptotic BCL-2 members, confirmed as sub-pathway); WikiPathways null — WP254/WP710/WP4839 candidates from seeder brief not confirmed via API (#gap/needs-canonical-id)
  • Primary sources cited (footnotes): Chipuk 2010 “BCL-2 Family Reunion” (10.1016/j.molcel.2010.01.025 — local PDF); Czabotar 2014 (10.1038/nrm3722 — not_oa); Singh/Green 2019 (10.1038/s41580-018-0089-8 — green OA PMC7325303, download failed); Letai 2008 (10.1038/nrc2297 — not_oa); Oltersdorf 2005 ABT-737 (10.1038/nature03579 — local PDF); Souers 2013 venetoclax (10.1038/nm.3048 — not_oa); Roberts 2015 venetoclax CLL (10.1056/nejmoa1513257 — local PDF); Chang 2016 navitoclax HSC (10.1038/nm.4010 — green OA, download failed)
  • SCAP claims (HUVECs BCL-xL dominant; IMR90 triple BCL-2+BCL-xL+BCL-w; preadipocytes not BCL-2 family) inherited from bcl-xl.md + bcl-2.md verified pages; Zhu 2015/2016 DOIs exhibit BUG-2 archive mismatch pattern — not re-verified independently
  • gaps surfaced: contradictory-evidence (PUMA direct-activator vs sensitizer tension between puma.md and Chipuk 2010); unsourced (BH3 selectivity matrix affinities; MCL-1 half-life); needs-canonical-id (WikiPathways); needs-human-replication (Chang 2016 HSC rejuvenation in mice only)
  • implicit stubs created (new wikilinks to non-existent pages): er-stress, hypoxia-signaling, bcl-w, bok
  • ROADMAP updated: Pathways section (after nf-kb), Tier B (bcl-2-family-signaling marked seeded R5d), Tier C (bcl2-family-pathway aliased), Round 5 plan section (pathway aggregator checked off)
  • TODO propagation: resolve PUMA classification in puma.md; propagate BIM-MCL-1 74 nM correction to any pages citing ~1 nM; once WikiPathways ID confirmed, update frontmatter

[2026-05-04] ingest | round-5a-bh3-only

Round 5a closes the BH3-only family sub-cluster of the apoptosis-machinery completion (per ROADMAP.md Round 5). 4 new protein pages drafted in parallel by wiki-seeder agents: bim, bad, bid, noxa. All ship with verified: false + ⚠️ banner pending verification batch.

bim

  • added: molecules/proteins/bim.md (BCL2L11 / BIM, BH3-only direct activator)

Canonical IDs pulled:

  • UniProt O43521 (B2L11_HUMAN) — REST API; 198 aa canonical isoform (BIM-EL), BH3 motif residues 148–162, key PTMs (Ser69/77/87 phosphorylation, TRIM2-mediated ubiquitination)
  • NCBI Gene 10018, HGNC 994 confirmed

Primary sources cited (footnotes):

  • O’Connor 1998 (doi:10.1093/emboj/17.2.384) — original BIM cloning + DLC1 sequestration; archive: bronze OA, pending download
  • Bouillet 1999 (doi:10.1126/science.286.5445.1735) — Bim KO mouse phenotype; archive: not_oa; no-fulltext-access applied
  • Dijkers 2000 (doi:10.1016/s0960-9822(00)00728-4) — FOXO3 transactivation of BCL2L11; archive: bronze OA, pending download; no-fulltext-access applied
  • Chen 2005 (doi:10.1016/j.molcel.2004.12.030) — BH3 affinity panel; archive: bronze OA, pending download; no-fulltext-access applied to specific Kd values
  • Ley 2003 (doi:10.1074/jbc.m301010200) — ERK1/2 phosphorylation of BIM-EL; archive: hybrid OA, pending download
  • Chipuk 2010 (doi:10.1016/j.molcel.2010.10.019) — direct activator vs sensitizer classification; archive: bronze OA, pending download
  • Opferman 2003 (doi:10.1038/nature02067) — BIM–MCL-1 affinity 74 ± 2 nM; archive: local PDF confirmed
  • Vogler 2025 (doi:10.1038/s41392-025-02176-0) — BCL-2 family review; archive: local PDF confirmed

Key decisions / corrections incorporated:

  • BIM classified as direct activator per Chipuk 2010 / bak.md verification; PUMA distinction as sensitizer made explicit
  • BIM–MCL-1 affinity stated as ~74 nM (not ~1 nM); sourced from mcl-1.md verifier (Opferman 2003 FP, confirmed via Vogler 2025)
  • Chen 2005 BCL-2/BCL-xL affinity values cited as “~1–10 nM range” pending PDF download; no-fulltext-access tagged
  • genage-id null + needs-canonical-id (BCL2L11 not found in GenAge-human search)

Implicit stubs reinforced (existing wikilinks gaining inbound count):

Verification priority for wiki-verifier:

  1. HIGH: Chen 2005 (doi:10.1016/j.molcel.2004.12.030) — affinity values for BCL-2/BCL-xL/BCL-w/A1 need direct numerical verification (bronze OA, pending download)
  2. HIGH: Opferman 2003 (doi:10.1038/nature02067) — local PDF available; verify 74 ± 2 nM BIM–MCL-1 FP value in-page
  3. MEDIUM: O’Connor 1998 (DLC1 sequestration); Ley 2003 (ERK phospho-sites); Dijkers 2000 (FOXO3 promoter binding)
  4. LOW: Bouillet 1999 (not_oa; requires institutional access)

noxa

  • added: molecules/proteins/noxa.md — NOXA (PMAIP1), BH3-only MCL-1/A1-selective sensitizer

Canonical IDs confirmed:

  • UniProt Q13794 (NOXA_HUMAN) — confirmed via REST API; 54 aa, 6.0 kDa, BH3 motif residues 29–37, mitochondrial targeting region 41–50
  • NCBI Gene 5366, HGNC 9108 — from task spec; not independently re-verified against NCBI/HGNC web
  • ENSEMBL ENSG00000141682 — from task spec; not independently re-verified

Primary-source DOIs cited and archive status:

  • doi:10.1126/science.288.5468.1053 (Oda 2000) — locally downloaded; 2,078 citations
  • doi:10.1016/j.molcel.2004.12.030 (Chen 2005) — archive pending download (bronze OA); 1,787 citations; no-fulltext-access
  • doi:10.1126/science.1090072 (Villunger 2003) — not_oa; not downloaded; 1,307 citations; no-fulltext-access
  • doi:10.1101/gad.1103603 (Shibue 2003) — archive pending download (diamond OA); 336 citations; no-fulltext-access
  • doi:10.1038/415045a (Tyner 2002) — locally downloaded (previously verified)
  • doi:10.1038/s41418-020-0564-6 (Shahbandi 2020) — download failed; no-fulltext-access
  • doi:10.1128/mcb.10.6.2879 (Hijikata 1990 original cloning) — DOI not in archive or Crossref; pre-DOI-era paper; no-fulltext-access

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

  • bcl-2-family-signaling (4+1 inbound — pre-existing stub, gains one ref)
  • apoptosenes (pre-existing stub in ROADMAP Tier B)
  • A1 / BFL-1 (no wiki page exists; referenced inline without wikilink — correct behavior; stub candidate)

Gaps surfaced:

  • NOXA protein half-life quantification (#gap/unsourced — precise pulse-chase value absent)
  • NOXA in IMR90 SCAP context (#gap/needs-replication — MCL-1 contribution to IMR90 senolysis untested)
  • NOXA in aged human non-tumor tissues (#gap/unsourced — no direct protein-level measurements found)
  • Hijikata 1990 original cloning paper (#gap/no-fulltext-access — DOI unresolvable)
  • p53-independent NOXA induction transcription factors (#gap/unsourced for UPR/ER stress pathway)
  • GenAge entry for PMAIP1 (#gap/needs-canonical-id — not confirmed)

Schema decisions:

  • sens-categories: ["[[apoptosenes]]"] added — NOXA is part of the pro-apoptotic machinery cleared/targeted in SENS ApoptoSENS strategy; consistent with bcl-xl.md and mcl-1.md convention.
  • ENSEMBL included in identity table (body) even though not in CLAUDE.md required frontmatter; consistent with puma.md prototype.
  • Willis 2007 (doi:10.1126/science.1133289) cited in body text but not as numbered footnote — the paper establishes the “indirect sensitizer” model (BH3-only proteins engage pro-survival proteins, not Bax/Bak directly); relevant context but not a primary NOXA-specific source; referenced inline in puma.md and mcl-1.md.

Verification priority for wiki-verifier:

  1. HIGH: Oda 2000 (doi:10.1126/science.288.5468.1053) — locally downloaded; verify p53RE location, murine vs human distinction, and BH3-domain characterization
  2. HIGH: Chen 2005 (doi:10.1016/j.molcel.2004.12.030) — affinity table values are the most load-bearing quantitative claim on this page; bronze OA pending download
  3. MEDIUM: Shibue 2003 (doi:10.1101/gad.1103603) — diamond OA pending; verify Noxa−/− MEF apoptosis resistance data and GI irradiation phenotype
  4. MEDIUM: Villunger 2003 (doi:10.1126/science.1090072) — not_oa; verify DKO additive resistance claim via abstract or institutional access
  5. LOW: Hijikata 1990 — unresolvable DOI; may need PubMed direct PMID search or institutional library request

bid

  • added: molecules/proteins/bid.md — BID (BH3-interacting domain death agonist); type=protein; UniProt P55957; NCBI Gene 637; HGNC 1050
  • 7 primary DOIs cited: Wang 1996 (10.1101/gad.10.22.2859, archive pending), Li 1998 (10.1016/s0092-8674(00)81590-1, local PDF), Luo 1998 (10.1016/s0092-8674(00)81589-5, local PDF), Zha 2000 (10.1126/science.290.5497.1761, local PDF), Yin 1999 (10.1038/23730, not_oa), Kuwana 2002 (10.1016/s0092-8674(02)01036-x, archive pending), Chipuk 2010 (10.1016/j.molcel.2010.01.025, local PDF)
  • gaps surfaced: BH3 affinity Kd values need primary quantitative source; caspase-8 human cleavage site numbering needs verification; cathepsin/calpain cleavage references need primary sources; aging-specific BID regulation in human tissues absent
  • implicit stubs created: none new; all wikilinks point to existing pages

bad

  • added: molecules/proteins/bad.md — BAD (BCL-2-associated agonist of cell death); type=protein; UniProt Q92934; NCBI Gene 572; HGNC 936
  • 4 primary DOIs cited: Yang 1995 (10.1016/0092-8674(95)90411-5, archive pending), Datta 1997 (10.1016/s0092-8674(00)80405-5, local PDF), Zha 1996 (10.1016/s0092-8674(00)81382-3, local PDF), Harada 1999 (10.1016/s1097-2765(00)80469-4, archive pending)
  • source correction surfaced: task brief attributed Ser112 (mouse) to RSK; Harada 1999 (Mol Cell 3:413–422) identifies mitochondria-anchored PKA/AKAP as the primary in-vivo kinase; RSK can phosphorylate this site in vitro but is not the dominant in-vivo kinase per Harada 1999 — tagged contradictory-evidence on the page
  • source correction surfaced: task-brief DOI for Yang 1995 was 10.1016/0092-8674(95)90412-3 (wrong — that DOI maps to “p53 is a direct transcriptional activator of the human bax gene”); correct DOI is 10.1016/0092-8674(95)90411-5 (Cell 80:285–291). TODO: any wiki page citing Yang 1995 BAD with the wrong DOI should be corrected.
  • gaps surfaced: precise Ki/Kd values for BAD BH3 affinity panel (BCL-xL/BCL-2/BCL-w/MCL-1/A1) not confirmed from a single primary-source affinity assay; PP2A dephosphorylation of BAD unsourced (primary study page not created); BAD phosphorylation state in senescent vs non-senescent cells not characterised; Bad-/- mouse lifespan data absent; GenAge entry for BAD not confirmed
  • implicit stubs created (new): 14-3-3 (new wikilink target — Tier D-ish stub); bcl-w (existing Tier D stub gains another inbound)

[2026-05-04] ingest | round-5b-senolytics-apoptosome

Round 5b continues the apoptosis machinery completion cluster (ROADMAP Round 5). This session seeds 4 entities in parallel: 2 senolytic compounds (navitoclax, a1331852) and 2 apoptosome components (apaf-1, cytochrome-c). Together with R5a’s BH3-only family, this closes most of the apoptosis-pathway cluster — only executioner caspases + apoptosis process page remain (R5c).

apaf-1

  • added: molecules/proteins/apaf-1.md — APAF-1 (Apoptotic Protease-Activating Factor 1); core apoptosome scaffold; type=protein
  • canonical IDs confirmed: UniProt O14727 (1,248 aa, 6 isoforms via REST API; CARD 1-90, NB-ARC 104-415, WD40 613-1248), NCBI Gene 317, HGNC 576; genage-id null (not in GenAge as of 2026-05-04)
  • primary sources cited (7 footnotes):
    • Zou 1997 (Cell, 10.1016/s0092-8674(00)80501-2) — local PDF; APAF-1 purification/discovery
    • Yuan/Horvitz 1992 (Development, 10.1242/dev.116.2.309) — not_oa; CED-4 C. elegans ortholog
    • Acehan 2002 (Mol Cell, 10.1016/s1097-2765(02)00442-2) — pending; apoptosome cryo-EM (heptameric wheel)
    • Cecconi 1998 (Cell, 10.1016/s0092-8674(00)81732-8) — local PDF; Apaf1 KO mouse, brain overgrowth
    • Yoshida 1998 (Cell, 10.1016/s0092-8674(00)81733-x) — local PDF; parallel Apaf1 KO, CNS apoptosis
    • Honarpour 2000 (Dev Bio, 10.1006/dbio.1999.9585) — not_oa; Apaf1 KO adult male infertility
    • Soengas 2001 (Nature, 10.1038/35051606) — local PDF; APAF-1 silencing in metastatic melanoma (~40%)
  • DOI correction surfaced: task brief Honarpour 2000 DOI was 10.1006/dbio.1999.9525 (a neural-patterning paper, BUG-2 type mismatch); correct Honarpour Apaf1 paper is 10.1006/dbio.1999.9585. Flagged in footnote; verifier should cross-check.
  • implicit stubs created: apoptosome, xiap, caspase-9, caspase-3, caspase-7, hsp70, bcl-2-family-signaling, apoptosenes
  • gaps surfaced: genage-id null; APAF-1 expression in senescent cells unsourced; HSP70/nucleolin/PHAPI regulatory interactions unsourced; isoform-specific functional distinctions unsourced; atomic-resolution apoptosome structures post-2010 not cited
  • ROADMAP: Tier C apaf-1 marked ✓ seeded R5b; Proteins section entry added

cytochrome-c

  • added: molecules/proteins/cytochrome-c.md (CYCS / cytochrome c somatic, UniProt P99999, NCBI Gene 54205, HGNC 19986)

Canonical IDs confirmed:

  • UniProt P99999 (CYC_HUMAN) — REST API; 105 aa precursor / 104 aa mature; heme c attachment Cys15+Cys18; His19/Met81 axial coordination; pI ~10
  • NCBI Gene 54205, HGNC 19986 — from task spec; matched UniProt gene cross-references
  • Mouse ortholog Cycs confirmed (>99% aa identity)

Primary-source DOIs cited and archive status:

  • doi:10.1016/s0092-8674(00)80085-9 (Liu 1996, Cell) — local PDF available · 5,151 citations · dATP/cytochrome c requirement for cell-free apoptosis
  • doi:10.1126/science.275.5303.1129 (Yang 1997, Science) — not_oa; no-fulltext-access · 4,916 citations · BCL-2 blocks cytochrome c release
  • doi:10.1126/science.275.5303.1132 (Kluck 1997, Science) — local PDF available · 4,717 citations · cytochrome c release as primary apoptosis trigger
  • doi:10.1038/35004029 (Goldstein 2000, Nat Cell Biol) — local PDF available · 1,056 citations · rapid, complete, kinetically invariant release
  • doi:10.1093/emboj/17.1.37 (Bossy-Wetzel 1998, EMBO J) — DOI lookup pending (bronze OA) · 1,227 citations · release precedes ΔΨm collapse
  • doi:10.1038/nchembio727 (Kagan 2005, Nat Chem Biol) — not_oa; no-fulltext-access · 1,225 citations · cardiolipin peroxidase activity amplifies release
  • doi:10.1016/j.cell.2006.06.025 (Frezza 2006, Cell) — DOI lookup pending (bronze OA) · 1,616 citations · OPA1 cristae junction control of cytochrome c pool accessibility

Note on Goldstein 2000 DOI: original task brief cited 10.1038/35008127 (not found in archive); correct DOI resolved via Crossref to 10.1038/35004029 (confirmed in archive, local PDF available). The incorrect DOI may be a transposition error.

Key structural/functional decisions:

  • UniProt residue numbering uses the 105 aa precursor throughout (Cys15/Cys18 heme attachment; His19/Met81 axial coordination); the mature 104 aa form reflects Met1 cleavage post-translation. Task brief used the older mature-104 aa numbering convention (Cys14/Cys17, His18, Met80); both conventions are noted in the page body with the precursor numbering flagged as primary per current UniProt P99999.
  • CYCS is absent from GenAge-human as of 2026-05-04 — tagged needs-canonical-id; not evidence against aging relevance.
  • sens-categories: ["[[apoptosenes]]"] assigned — cytochrome c is pro-apoptotic machinery; consistent with existing convention on bak.md, bcl-xl.md, etc.

Implicit stubs reinforced (gaining one more inbound reference):

  • apaf-1 (Tier C, 3 inbound — gains 1 from cytochrome-c.md)
  • opa1 (new wikilink target — Tier D+ stub; not previously in ROADMAP)
  • oxphos (Tier B/C implicit stub — gains 1 inbound)

Gaps surfaced:

  • Cytochrome c protein abundance in aged human tissue (direct proteomics): unsourced
  • Tyr49/Tyr98 phosphorylation state changes with age: unsourced
  • Apoptosome competence (APAF-1 / dATP availability) in senescent cells: no-mechanism
  • Cardiolipin oxidation loop quantitative contribution in intact aged cells: needs-replication
  • Yang 1997 (doi:10.1126/science.275.5303.1129) not_oa — BCL-2 blocks cytochrome c release claim carries no-fulltext-access
  • Kagan 2005 (doi:10.1038/nchembio727) not_oa — cardiolipin peroxidase claim carries no-fulltext-access

Verification priority for wiki-verifier:

  1. HIGH: doi:10.1016/s0092-8674(00)80085-9 (Liu 1996) — local PDF; verify dATP requirement, cytochrome c identity, and cell-free reconstitution quantitative details
  2. HIGH: doi:10.1038/35004029 (Goldstein 2000) — local PDF; verify “rapid, complete, kinetically invariant” framing and single-cell imaging methodology
  3. HIGH: doi:10.1126/science.275.5303.1132 (Kluck 1997) — local PDF; verify cytochrome c as primary BCL-2-regulated factor
  4. MEDIUM: doi:10.1016/j.cell.2006.06.025 (Frezza 2006) — bronze OA pending download; verify OPA1/cristae remodeling independence from fusion, 85% sequestration figure
  5. MEDIUM: doi:10.1093/emboj/17.1.37 (Bossy-Wetzel 1998) — bronze OA pending download; verify release-before-depolarization ordering
  6. LOW (blocked): doi:10.1126/science.275.5303.1129 (Yang 1997) — not_oa; doi:10.1038/nchembio727 (Kagan 2005) — not_oa

a1331852

  • added: molecules/compounds/a1331852.md — A-1331852 (selective BCL-xL inhibitor, BH3-mimetic, research-grade senolytic); type=compound; verified: false
  • canonical IDs confirmed: PubChem CID 71565985 (via PubChem name search + synonym confirmation); ChEMBL CHEMBL3793424 (PubChem synonym list); CAS 1430844-80-6; MW 658.8 g/mol, formula C38H38N6O3S; InChIKey QCQQONWEDCOTBV-UHFFFAOYSA-N
  • ChEMBL ID discrepancy flagged: task brief suggested CHEMBL3753332 but PubChem synonyms return CHEMBL3793424 — tagged needs-canonical-id; must be resolved at verification against ChEMBL directly
  • author note: task brief referenced “Tao 2014” for the Ki values paper; Crossref and archive confirm first author is Koehler MFT; Tao ZF is a co-author. Cited as Koehler 2014 on the page.

Primary-source DOIs cited and archive status:

  • doi:10.1021/ml500030p (Koehler 2014, ACS Med Chem Lett) — archive pending download (green OA); Ki values BCL-xL ~10 pM / BCL-2 ~6 nM / BCL-w ~95 nM / MCL-1 >440 nM
  • doi:10.1126/scitranslmed.aaa4642 (Leverson 2015, Sci Transl Med) — archive pending download (green OA); cancer xenograft selectivity + thrombocytopenia; 528 citations
  • doi:10.1111/acel.12445 (Zhu 2016, Aging Cell) — local PDF available; primary senolytic data; HUVECs + IMR90 + preadipocyte null; 1,045 citations
  • doi:10.1038/ncomms11190 (Yosef 2016, Nat Commun) — local PDF available; BCL-W + BCL-xL dual SCAP in IMR90; directed senolysis young mice; 927 citations
  • doi:10.1038/s41591-019-0668-z (Khan 2019, Nat Med) — archive pending download (PMC6898785 open access); DT2216 PROTAC context; 601 citations

Constraints propagated from verified pages (bcl-xl, bax, bak verifiers):

  • A-1331852 active in HUVECs + IMR90 (not HUVECs only) — explicit in cell-type table
  • A-1331852 / A-1155463 were primary agents in Zhu 2016; navitoclax was comparator — explicit in body
  • IMR90 SCAP requires triple BCL-2 + BCL-xL + BCL-w (Zhu 2016) — captured in cell-type table
  • Preadipocyte SCAP is ephrins/PI3Kδ/p21/PAI-2, not BCL-2 family — captured in cell-type table

Gaps surfaced:

Implicit stubs created/reinforced:

  • a1155463 — new wikilink target (sister BCL-xL-selective compound; Tier D candidate for later Round 5 or 6)
  • bcl-w — gains another inbound from cell-type SCAP table

ROADMAP updates:

  • molecules/compounds/a1331852 row: [ ] → [x] (drafted, verified: false)
  • Tier A implicit-stub table: a1331852 → a1331852 ✓ seeded R5b

Verification priority for wiki-verifier:

  1. HIGH: doi:10.1111/acel.12445 (Zhu 2016) — local PDF; verify A-1331852 vs A-1155463 roles, HUVEC+IMR90 senolytic data, navitoclax-as-comparator framing, preadipocyte null result
  2. HIGH: doi:10.1038/ncomms11190 (Yosef 2016) — local PDF; verify BCL-W + BCL-xL dual requirement, in vivo model (young mice, oncogene-induced senescence)
  3. MEDIUM: doi:10.1021/ml500030p (Koehler 2014) — pending download (open access); verify Ki values, assay type (FP competitive binding)
  4. MEDIUM: doi:10.1126/scitranslmed.aaa4642 (Leverson 2015) — pending download (open access); verify thrombocytopenia data, cancer selectivity
  5. LOW: doi:10.1038/s41591-019-0668-z (Khan 2019) — pending download (PMC open access); verify DT2216 platelet-sparing mechanism and VHL expression rationale
  • added: molecules/compounds/navitoclax.md (navitoclax / ABT-263; type=compound)

Canonical IDs pulled:

  • PubChem CID 24978538 — confirmed via REST API (C₄₇H₅₅ClF₃N₅O₆S₃, MW 974.6, InChIKey JLYAXFNOILIKPP-KXQOOQHDSA-N)
  • ChEMBL CHEMBL443684 — confirmed via ChEMBL API (pref_name: NAVITOCLAX, MW 974.63); task brief had wrong ID CHEMBL1242177 (that resolves to a different compound, MW 503.99, wrong formula — corrected)
  • DrugBank DB12340 — task brief value; NOT independently verified (DrugBank API returned 403); needs-canonical-id applied

Primary sources cited (footnotes) and archive status:

  • doi:10.1158/0008-5472.CAN-07-5836 (Tse 2008, Cancer Res) — confirmed title + archive pending; no-fulltext-access on Kd values. NOTE: task brief had wrong Tse 2008 DOI (10.1158/0008-5472.can-08-1610 — that maps to a CP-31398/FAP paper). Correct DOI resolved via PubMed (PMID 18451170).
  • doi:10.1111/acel.12445 (Zhu 2016, Aging Cell) — confirmed + 1,045 citations; archive: local PDF — highest priority for verifier
  • doi:10.1038/ncomms11190 (Yosef 2016, Nat Commun) — confirmed + 927 citations; archive: local PDF — high priority
  • doi:10.1038/nm.4010 (Chang 2016, Nat Med) — confirmed + 1,694 citations; archive: download FAILED; no-fulltext-access — most critical verification gap
  • doi:10.1200/JCO.2011.34.7898 (Roberts 2012, JCO Phase I) — confirmed + archive pending; clinical thrombocytopenia source
  • doi:10.1056/NEJMoa1513257 (Roberts 2016, NEJM venetoclax) — confirmed + archive local PDF — venetoclax context
  • doi:10.1038/s41467-020-15838-0 (He 2020, Nat Commun PROTAC) — confirmed + archive pending; PROTAC BCL-xL degrader context

Key corrections vs prior wiki framing:

  1. ChEMBL ID CORRECTED: task brief had CHEMBL1242177 (wrong compound); correct CHEMBL443684 confirmed via API
  2. Tse 2008 DOI CORRECTED: task brief had 10.1158/0008-5472.can-08-1610 (CP-31398 paper, unrelated); correct 10.1158/0008-5472.CAN-07-5836 confirmed via PMID 18451170
  3. Preadipocyte null result (from bak.md verifier): cell-type specificity table explicitly marks navitoclax as NOT senolytic in preadipocytes; SCAP there is ephrin/PI3Kδ/p21
  4. Navitoclax NOT in Zhu 2015 (from bcl-xl.md verifier): primary senolysis sources are Zhu 2016 + Yosef 2016 only
  5. IMR90 triple-SCAP (from bcl-2.md verifier): triple BCL-2 + BCL-xL + BCL-W knockdown required; mechanistically consistent with navitoclax’s broad profile
  6. BIM displacement mechanism (from bim.md): explicit — navitoclax frees BIM from BCL-2/xL/w → BAK/BAX activation

Implicit stubs reinforced:

Gaps surfaced:

  • Chang 2016 fulltext: no-fulltext-access (most important gap; all Chang 2016 in-vivo functional claims unverified)
  • DrugBank DB12340: needs-canonical-id (API blocked)
  • Tse 2008 affinity values: no-fulltext-access (archive pending)
  • Clinical PK in aged non-cancer humans: dose-response-unclear
  • Human senolytic trials for navitoclax: none identified as of 2026-05-04
  • MCL-1-dominant senescent populations: no-mechanism for navitoclax resistance

Verification priority for wiki-verifier:

  1. CRITICAL: doi:10.1038/nm.4010 (Chang 2016) — DOI lookup failed; founding in-vivo aged-mouse senolysis paper; all in-vivo functional claims unverified; recommend alternate download (Europe PMC / PMC OA)
  2. HIGH: doi:10.1111/acel.12445 (Zhu 2016) — local PDF; verify HUVEC/IMR90/preadipocyte specificity table, triple-knockdown BCL-2+BCL-xL+BCL-W in IMR90, MCL-1 Ki ~550 nM
  3. HIGH: doi:10.1038/ncomms11190 (Yosef 2016) — local PDF; verify BCL-W + BCL-xL dual requirement in IMR90, A1331852 vs navitoclax activity
  4. MEDIUM: doi:10.1158/0008-5472.CAN-07-5836 (Tse 2008) — pending; verify BCL-2/BCL-xL/BCL-W Ki <0.5 nM, MCL-1 Ki ~550 nM, oral bioavailability data
  5. MEDIUM: doi:10.1200/JCO.2011.34.7898 (Roberts 2012) — pending; verify thrombocytopenia incidence + dose-response relationship
  6. LOW: doi:10.1056/NEJMoa1513257 (Roberts 2016) — local PDF; venetoclax BCL-2 selectivity context only

[2026-05-04] ingest | round-5c-executioners-apoptosis-process

caspase-3

  • added: molecules/proteins/caspase-3.md
  • updated: ROADMAP.md — executioners line updated to [x] caspase-3 (drafted) seeded R5c
  • canonical IDs confirmed: UniProt P42574, NCBI Gene 836, HGNC 1504
  • DOIs cited (7 footnotes):
    • 10.1016/s0021-9258(18)47344-9 (Fernandes-Alnemri 1994 CPP32) — archive pending
    • 10.1038/376037a0 (Nicholson 1995 apopain) — local PDF available
    • 10.1016/0092-8674(95)90541-3 (Tewari 1995 Yama) — archive pending
    • 10.1016/s0092-8674(00)80197-x (Liu 1997 DFF/ICAD) — local PDF available
    • 10.1038/34112 (Enari 1998 CAD) — not_oa
    • 10.1038/384368a0 (Kuida 1996 KO) — not_oa
    • 10.1073/pnas.0707715105 (Walsh 2008 caspase-3 vs -7) — archive pending
  • BUG-2 DOI mismatches caught and corrected (flagged in page footnotes):
    • Brief DOI 10.1016/0092-8674(95)90426-3 → wrong paper (retinotectal ELF-1/Mek4); corrected to 10.1016/0092-8674(95)90541-3 (Tewari 1995 Yama, PMID 7774019)
    • Brief DOI 10.1073/pnas.0805089105 → wrong paper (plant fatty acid dehydratase PASTICCINO2); corrected to 10.1073/pnas.0707715105 (Walsh 2008, PMID 18723680)
    • Brief DOI 10.1074/jbc.269.49.30761 → not in archive; correct DOI 10.1016/s0021-9258(18)47344-9 confirmed via Crossref/PubMed PMID 7983002
  • implicit stubs created: caspase-9, caspase-8, xiap, smac-diablo
  • gaps surfaced:
  • schema note: none; all fields covered by CLAUDE.md protein schema

caspase-9

  • added: molecules/proteins/caspase-9.md
  • updated: ROADMAP.md — executioners line updated to include [x] caspase-9 (drafted) seeded R5c; Proteins section added entry
  • canonical IDs confirmed: UniProt P55211, NCBI Gene 842, HGNC 1511
  • DOI correction caught during source-validation: brief stated 10.1016/s0092-8674(00)80213-5 as Li 1997 Cell — that DOI resolves to the SREBP paper (Brown & Goldstein); correct DOI is 10.1016/s0092-8674(00)80434-1 (Li et al. 1997, “Cytochrome c and dATP-Dependent Formation of Apaf-1/Caspase-9 Complex…”), confirmed via PubMed PMID 9390557 and DOI lookup (7,208 citations; local PDF available)
  • DOI correction caught: brief stated Srinivasula 1998 DOI as 10.1093/emboj/17.6.1675 — that resolves to Scaffidi 1998 “Two CD95 signaling pathways” (EMBO J, Fas pathway, ~2,997 citations); correct paper is Srinivasula 1998 Mol Cell “Autoactivation of Procaspase-9 by Apaf-1-Mediated Oligomerization” DOI 10.1016/s1097-2765(00)80095-7 (PMID 9651578; archive pending download)
  • DOI correction caught: brief stated Renatus 2001 DOI as 10.1073/pnas.241451798 — not found in archive; correct DOI confirmed via PubMed PMID 11734640 is 10.1073/pnas.231465798 (“Dimer formation drives the activation of the cell death protease caspase 9”; archive pending download)
  • DOIs cited (7 footnotes):
    • 10.1016/s0092-8674(00)80434-1 (Li 1997 — apoptosome reconstitution) — local PDF available (7,208 citations)
    • 10.1016/s1097-2765(00)80095-7 (Srinivasula 1998 Mol Cell — APAF-3/autoactivation) — archive pending (1,125 citations)
    • 10.1073/pnas.96.20.10964 (Salvesen & Dixit 1999 — induced proximity model) — not_oa; no local PDF
    • 10.1073/pnas.231465798 (Renatus 2001 — dimer-activation mechanism) — archive pending (440 citations)
    • 10.1016/s0092-8674(00)81477-4 (Hakem 1998 — Casp9-KO brain overgrowth) — local PDF available (1,283 citations)
    • 10.1016/s0092-8674(00)81476-2 (Kuida 1998 — parallel Casp9-KO) — local PDF available (1,630 citations)
    • 10.1126/science.282.5392.1318 (Cardone 1998 — AKT Ser196 phosphorylation) — not_oa; no local PDF (3,110 citations)
  • implicit stubs created: xiap, smac-diablo, caspase-7 (cross-referenced as caspase-9 substrate)
  • gaps surfaced:
    • unsourced — XIAP elevation in human senescent cells (cited contextually but no primary human cell paper cited on this page)
    • unsourced — 9β isoform ratio changes with age in human tissues
    • unsourced — apoptosome competence measurement in primary aged human cells
    • needs-replication — 9γ isoform function (limited characterisation)
    • needs-human-replication — KO phenotype (brain overgrowth) characterised in mouse only
    • no-mechanism — caspase-9 / NLRP3 crosstalk in aged macrophages
    • no-mechanism — downstream consequences of declining apoptosome competence in aged tissue
  • schema note: genage-id: null — CASP9 does not have a GenAge human entry; field retained as null per convention. No other schema gaps encountered.
  • verification priority: HIGH for Cardone 1998 (AKT Ser196 site — mechanistically important; not_oa), Renatus 2001 (operative-species claim — core structural claim; pending download), Hakem/Kuida 1998 KO papers (both local PDFs available — verify brain overgrowth phenotype specifics and exact genotype details)

caspase-8

  • added: molecules/proteins/caspase-8.md (type=protein; aliases: CASP8, FLICE, MCH5, CAP4)
  • canonical IDs pulled: UniProt Q14790 (CASP8_HUMAN), NCBI Gene 841, HGNC 1509; 479 aa canonical isoform; active-site Cys360/His317 confirmed
  • primary DOIs cited in footnotes:
    • doi:10.1016/s0092-8674(00)81266-0 (Muzio 1996 FLICE discovery; archive: LOCAL PDF available)
    • doi:10.1016/s0092-8674(00)81265-9 (Boldin 1996 MACH discovery; archive: LOCAL PDF available)
    • doi:10.1093/emboj/17.6.1675 (Scaffidi 1998 type I/II cells; archive: pending download)
    • doi:10.1016/s1074-7613(00)80609-3 (Varfolomeev 1998 Casp8-/- KO; archive: pending download)
    • doi:10.1038/nature09852 (Oberst 2011 CASP8-FLIP necroptosis suppression; archive: LOCAL PDF available)
    • doi:10.1038/nature09857 (Kaiser 2011 Casp8/Rip3 double-KO rescue; archive: LOCAL PDF available)
  • key sections: identity, discovery (Muzio/Boldin 1996), domain architecture (DED-DED-p18-p10), DISC assembly/dimerization model, type I vs type II cells, BID cleavage bridge, FLIP threshold switch, necroptosis suppression (Oberst 2011, Kaiser 2011), Casp8-/- KO (Varfolomeev 1998), PTMs (Tyr380/Ser387), non-canonical inflammatory roles, aging relevance (AICD decline, senescent cell DISC, SASP death ligands), limitations/gaps
  • implicit stubs created/reinforced: fadd, cflar, ripk1, ripk3, caspase-3, caspase-9
  • BUG-2 flag: task brief supplied wrong DOIs for Muzio 1996 (10.1016/s0092-8674(00)81877-2 = Ca2+ channel paper) and Boldin 1996 (10.1016/s0092-8674(00)81876-0 = Spo11 meiosis paper). Corrected to confirmed DOIs via PubMed efetch (PMIDs 8681377 and 8681376). Documented on page Limitations section. No archive corrections needed — wrong DOIs were only in the task brief, not previously in the wiki.
  • gaps surfaced:

apoptosis (process)

  • added: processes/apoptosis.md (type=process; aliases: programmed cell death type I, PCD-I)
  • primary DOIs cited in footnotes:
    • doi:10.1038/bjc.1972.33 (Kerr 1972 — discovery/naming; archive: pending download, bronze OA)
    • doi:10.1038/s41418-017-0012-4 (Galluzzi 2018 NCCD — cell death nomenclature; archive: pending download, bronze OA)
    • doi:10.1016/j.cell.2015.05.026 (Bergmann 2015 — cardiomyocyte turnover; archive: LOCAL PDF available)
    • doi:10.1038/ncomms11190 (Yosef 2016 — SCAP/senolytic; archive: LOCAL PDF available)
    • doi:10.1016/j.cell.2013.05.039 (Lopez-Otin 2013 — hallmarks context; archive: LOCAL PDF available)
  • key sections: morphological hallmarks table, discovery/naming, pathway orientation (cross-link to apoptosis-pathway), cell death mode comparison table, dual paradox in aging (excess + insufficient), SCAP table, quantitative tissue rates, aging modulators, measurement methods, limitations/gaps
  • implicit stubs reinforced: anoikis, bcl-2-family-signaling, nlrp3-inflammasome, beclin-1
  • gaps surfaced:
  • schema note: process page carries verified: false for consistency with autophagy.md precedent (multi-source pages — CLAUDE.md schema does not currently require this field on type=process; flagged for schema review).

[2026-05-04] verify | molecules/proteins/caspase-3.md

Pages verified: 1

  • molecules/proteins/caspase-3.md — corrections: 7 substantive; set verified: true (partial scope — Fernandes-Alnemri 1994 + Tewari 1995 download-failed; Enari 1998 + Kuida 1996 not_oa; UniProt/NCBI identity fields not live-checked)

Sources checked:

  • Nicholson 1995 (10.1038/376037a0) — local PDF, verified end-to-end; Nature 376:37–43
  • Liu 1997 (10.1016/s0092-8674(00)80197-x) — local PDF, verified end-to-end; Cell 89:175–184, April 18, 1997
  • Walsh 2008 (10.1073/pnas.0707715105) — downloaded 2026-05-04 via PMC (PMCID PMC2529079); PNAS 105:12815–12819, Sept 2, 2008
  • Fernandes-Alnemri 1994 (10.1016/s0021-9258(18)47344-9) — download failed (old Elsevier); unverifiable
  • Tewari 1995 (10.1016/0092-8674(95)90541-3) — download failed (old Elsevier); unverifiable
  • Enari 1998 (10.1038/34112) — not_oa; unverifiable; claims tagged no-fulltext-access
  • Kuida 1996 (10.1038/384368a0) — not_oa; unverifiable; claims tagged no-fulltext-access

Key corrections (7 substantive):

  1. PARP-1 cleavage site: “DEVD/Asp214” → “between Asp216 and Gly217” (Nicholson 1995)
  2. ICAD cleavage sites: “two DEVD sites” → “DETD (aa 117) and DAVD (aa 224)” (Liu 1997)
  3. Nicholson 1995 cell line: “U937” → “THP-1 human monocytic leukemia cells” (purification source per Nicholson 1995)
  4. Walsh 2008 method: “SILAC-based substrate profiling” → immunodepletion + 2D SDS-PAGE + immunoblotting (SILAC not in Walsh 2008)
  5. Liu 1997 DFF naming: “DFF45 (ICAD) + DFF40 (CAD)” → DFF-45 + DFF-40 (ICAD/CAD naming is from Enari 1998); DFF-40 cDNA cloning incomplete at publication; purification source was HeLa S-100 (not Jurkat)
  6. Kinetics from Nicholson 1995 added: K_m (Ac-DEVD-AMC) = 9.7 ± 1.0 μM; K_i (Ac-DEVD-CHO) < 1 nM
  7. Walsh 2008 substrate quantification added: 12/20 substrates preferentially cleaved by caspase-3; 1/20 (p23) by caspase-7

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

  • processes/apoptosis.md — check PARP cleavage site (Asp214 → Asp216–Gly217)
  • pathways/apoptosis-pathway.md — check ICAD cleavage site description (DEVD → DETD/DAVD)
  • Any page citing Walsh 2008 as SILAC — correct to immunodepletion + 2D-gel method

[2026-05-04] verify | pathways/bcl-2-family-signaling.md

Pages verified: 1 (partial scope — Czabotar 2014, Letai 2008, Souers 2013 not_oa; Chang 2016, Singh 2019 OA-download-failed)

Sources checked:

  • Chipuk 2010 “BCL-2 Family Reunion” (10.1016/j.molcel.2010.01.025) — local PDF, verified end-to-end (10 pages); Mol Cell 37(3):299-310
  • Oltersdorf 2005 ABT-737 (10.1038/nature03579) — local PDF, verified end-to-end (5 pages); Nature 435:677-681
  • Roberts 2016 venetoclax CLL (10.1056/nejmoa1513257) — local PDF, verified end-to-end (9 pages); N Engl J Med 374:311-22
  • KEGG hsa04210 — confirmed via REST API (“Apoptosis - Homo sapiens”; broader than this page)
  • Reactome R-HSA-109606 — confirmed via ContentService API (“Intrinsic Pathway for Apoptosis”)
  • Reactome R-HSA-111453 — confirmed via ContentService API (“BH3-only proteins associate with and inactivate anti-apoptotic BCL-2 members”)
  • Czabotar 2014 (10.1038/nrm3722) — permanently not_oa; tagged no-fulltext-access
  • Letai 2008 (10.1038/nrc2297) — permanently not_oa; tagged no-fulltext-access
  • Souers 2013 (10.1038/nm.3048) — permanently not_oa; tagged no-fulltext-access
  • Chang 2016 (10.1038/nm.4010) — green OA but 0 candidate URLs (permanent download failure); tagged no-fulltext-access
  • Singh 2019 (10.1038/s41580-018-0089-8) — green OA but 0 candidate URLs (permanent download failure); tagged no-fulltext-access

Corrections made:

  • PUMA classification Fig reference: “Fig. 1C” → “Fig. 1B” (Chipuk 2010 Fig 1B is the protein-tier classification diagram; Fig 1C is the binding-profile matrix). Additionally expanded the PUMA note to explain that pan-binding profile ≠ direct-activator classification under Chipuk 2010 framework.
  • Effector BH domain description: “BH1+BH2+BH3 (no BH4)” → added nuance: “originally described as lacking BH4; structure-based alignment revealed a conserved BH4 motif per Chipuk 2010” (Chipuk 2010 p.299 explicitly notes this)
  • Citation key rename: [^roberts2015] → [^roberts2016] throughout (paper published January 28, 2016; footnote body already said “NEJM 2016” but key name was inconsistent)
  • Roberts footnote expanded with verified numbers: n=116 (56+60), 79% ORR (95% CI 71-86), 20% CR (95% CI 13-28), 15-month PFS 66% (dose-escalation), 2-year OS 84%; full journal citation added
  • Oltersdorf 2005 footnote expanded with verified ABT-737 affinities: BCL-xL Ki ~1 nM (competitive FP), BCL-2 IC50 ~103 nM (10% serum), MCL-1 Ki >1 µM, A1 Ki >1 µM; in-vivo details added (25-100 mg/kg i.p., n=9-10/group, 77% H1963 complete regression)
  • Chipuk 2010 footnote expanded with verified volume/page, author list, and specific page-level claim confirmations
  • Chang 2016 footnote updated: “download failed” → “permanent download failure (0 candidate URLs — cannot verify)” + no-fulltext-access
  • Singh 2019 footnote updated similarly
  • Limitations section: added two explicit no-fulltext-access entries (one for Chang 2016/Singh 2019; one for Czabotar 2014/Letai 2008/Souers 2013)
  • Auto-extraction ⚠️ banner replaced with verification comment block
  • Frontmatter: verified: false → true; verified-date: 2026-05-04; verified-by: claude; verified-scope added (full scope description)

Unverifiable claims (sources not accessible):

  • BH3-binding selectivity matrix quantitative affinities (Czabotar 2014, Chen 2005, Opferman 2003 — unsourced remains)
  • Venetoclax Ki values in BH3-mimetics table (BCL-2 Ki ~0.01 nM; BCL-xL Ki ~48 nM; MCL-1 Ki >444 nM) — Souers 2013 not_oa
  • BH3-profiling method details (Letai 2008 not_oa)
  • Chang 2016 quantitative HSC outcomes (% clearance, engraftment improvement, myeloid-bias correction) — permanent download failure

PUMA classification resolution:

  • Chipuk 2010 Fig 1B unambiguously places PUMA in “Sensitizers/de-repressors BH3-only proteins” (alongside BAD, BIK, BMF, HRK, Noxa)
  • PUMA pan-binding to all 5 anti-apoptotic members (same profile as BIM, per Fig 1C) does NOT qualify it as a direct activator under Chipuk 2010 — direct activation requires ability to engage and allosterically activate BAX/BAK, which Chipuk et al. 2005 found “may not be direct” for PUMA
  • Tension between this page (sensitizer) and puma.md (direct activator) persists — requires propagation pass to puma.md

Downstream propagation needed (for main agent):

  • puma.md — classification still uses “direct activator” framing; needs update to reflect Chipuk 2010 Fig 1B classification + gap tag for the academic controversy
  • bak.md — already incorporates Chipuk 2010 revision per upstream notes; verify the Fig reference there is also 1B not 1C
  • bim.md — affinity table method-note (Chen 2005 IC50 vs Opferman 2003 FP Kd) should be propagated to any claims on bcl-2-family-signaling.md that use those values (currently table is qualitative, so no immediate numeric fix needed here)
  • molecules/compounds/venetoclax.md or navitoclax.md — if these pages cite the Roberts 2016 trial as [^roberts2015], the citation key name should be updated there too (not the DOI, just the internal key label for consistency)
  • Any caspase pages (caspase-8.md, caspase-9.md) that cite Chipuk 2010 using the wrong DOI (10.1016/j.molcel.2010.10.019 = Gavathiotis/Walensky 2010 BAX NMR paper, not the Chipuk review) — check and correct if present

[2026-05-04] round-5 summary

Round 5 (Apoptosis machinery completion) — 13 new pages drafted + verified in single session. 4 sub-rounds with parallel seeder + verifier batches:

Verification outcomes: 13/13 verified (most verified-partial due to closed-access papers); zero pages remain on verified: false. Verifier batch surfaced dozens of corrections across the 13 pages, including:

  • Chipuk 2010 DOI mismatch — 10.1016/j.molcel.2010.10.019 actually resolves to Gavathiotis/Walensky 2010 (BAX NMR). The correct Chipuk 2010 “BCL-2 Family Reunion” review DOI is 10.1016/j.molcel.2010.01.025. Corrected on bim.md; verified correct already on bak.md and bcl-2-family-signaling.md.
  • PUMA classification resolved (with academic-controversy framing): Chipuk 2010 Fig 1B explicitly places PUMA in sensitizer/derepressor tier (NOT direct activator). Tension between Chipuk 2005 (direct activator) and Chipuk 2010 (sensitizer) flagged with contradictory-evidence on puma.md.
  • BID caspase-8 cleavage species inversion — Asp59 is the caspase-8 site (conserved human/mouse, LQTD↓G); Asp75 is granzyme B (mouse). Wiki had this backwards in 6 locations.
  • Yosef 2016 used ABT-737 throughout — not navitoclax or A-1331852. All Yosef 2016 → navitoclax/A-1331852 senolytic attributions corrected on multiple pages.
  • Zhu 2016 used navitoclax — A-1331852 senolytic data in a1331852.md was wrongly attributed to Zhu 2016; corrected to navitoclax in source paper.
  • DT2216 PROTAC uses navitoclax pharmacophore (not A-1331852).
  • Caspase-3 PARP cleavage at Asp216-Gly217 (not Asp214); ICAD sites are DETD/aa117 + DAVD/aa224 (not generic DEVD).
  • Caspase-8 Jurkat = type II (not type I); Oberst/Kaiser 2011 volumes are 471 (not 480); active-site mutant is C360A (not C360S in Oberst 2011).
  • APAF-1 apoptosome mass = 1 MDa (not 1.4 MDa); Soengas 2001 melanoma silencing is NOT CpG promoter methylation (bisulfite-excluded).
  • Cytochrome c release ~5 min (not <1 min); residue numbering Cys15/18/His19/Met81 are precursor (not mature).
  • IMR-90 SCAP nuance — Yosef 2016 shows BCL-W + BCL-xL dual sufficient; Zhu 2016 shows triple BCL-2 + BCL-xL + BCL-w necessary. Different assays/conditions; both findings documented (#gap/contradictory-evidence).
  • Multiple BUG-2 DOI mismatches in seeder briefs caught and corrected: 3 on caspase-3, 2 on caspase-8, 3 on caspase-9, 1 on cytochrome-c (Goldstein 2000), 1 on BAD (Yang 1995), 1 on APAF-1 (Honarpour 2000), 1 on navitoclax (Tse 2008), 1 on a1331852 (ChEMBL ID — task brief had wrong ID for potassium polysulfide).

Propagation pass 2026-05-04 (this session):

  • puma.md: PUMA “direct activator” framing supplemented with Chipuk 2010 sensitizer reclassification; contradictory-evidence flagged
  • mcl-1.md: BIM-MCL-1 affinity table now notes method-dependence (Opferman 2003 FP Kd 74 nM vs Chen 2005 IC50 2.4 nM, not directly comparable); PUMA row updated with verified Chen 2005 IC50 = 5.0 nM
  • senolytics.md, apoptosis-pathway.md: spot-checked — already have correct Yosef 2016 ABT-737 framing

Coverage delta after Round 5:

  • Proteins seeded: 1 → 14 verified (+13 in this round)
  • Compounds: navitoclax + a1331852 added
  • Processes: apoptosis added (companion to apoptosis-pathway)
  • Pathways: bcl-2-family-signaling added (canonical; bcl2-family-pathway alias)
  • Tier A implicit-stub queue: 6 entries closed (bim, bad, bid, navitoclax, a1331852 — plus bid moved from 7 to 0)

Open follow-ups for next round:

  • BAD Ser-155/Ser-118 PKA site needs primary citation (Harada 1999 doesn’t cover it)
  • IMR-90 SCAP triple-vs-dual reconciliation between Zhu 2016 and Yosef 2016
  • Puthalakath 1999 should be added as DLC1 sequestration source on bim.md
  • A-1331852 Ki values need primary citation (Koehler 2014 doesn’t report them)
  • Salvesen 1999 (induced proximity) and Cardone 1998 (AKT-Casp9 Ser196) remain not_oa unverifiable
  • Chang 2016 (Nat Med navitoclax in vivo) remains permanent no-fulltext-access

Graph View

Backlinks