R45

In-progress round. Theme: AGE biology + skin-aging completion + schema escalations.

Entries from log.md (chronological within the round):

2026-05-20 — schema R45: type:experiment + matrix-as-blocker-tracker

Origin: user brainstorming session on largest blockers to meaningful aging-rate reduction / tissue rejuvenation, framed by their wet-lab training + interest in an intratissue AGE breaker as a candidate target. Survey of the wiki’s verified causal graph + intervention-tractability landscape (per hallmark-causality-graph + causal-graph-data) identified that the existing causal-graph framework is already structurally a blocker tracker but doesn’t surface tractability + resolving-experiment data as first-class fields. User agreed to formalize this as an “experiments” folder with status-tracked proposal pages.

Schema additions:

1. type: experiment — new page type in CLAUDE.md § Page types and frontmatter. Lives in experiments/ directory. Schema fields:

   • status: — proposed | accepted | in-progress | paused | complete | published | abandoned
   • status-changed:, proposed-date:, last-updated: — ISO dates
   • question: / hypothesis: — one-sentence each
   • resolves-edges: / resolves-nodes: / adds-nodes: — load-bearing connection back to causal-graph-data
   • scale: — small-lab | small-clinic | large-clinic | multi-site | multi-decade
   • duration-estimate:, cost-estimate:, resources-required:
   • preregistered: + preregistration-doi: (OSF / ClinicalTrials.gov)
   • published-as: — wikilink to a type: study page in studies/ when results are written up
   • related-experiments: + related-pages:

   Discipline: experiment pages do NOT carry verified: (proposals are not facts). Citation discipline DOES apply to motivating-fact claims in the body. When complete, results may be promoted to a type: study page in studies/.

2. Matrix-as-blocker-tracker. causal-graph-data extended with two new sections:

   • Edge-blocker tracker — per-edge table: Edge | Upstream tractability | Resolving experiment | Scale. 32 edges populated. 11 of 34 edges flow from low-tractability upstream nodes (Tier 1 / genomic-instability + telomere-attrition cluster) — these are the structural matrix blockers. 17 edges have null resolving experiments — operationally blocked rather than biologically blocked.
   • Missing-node blockers — two damage classes outside the López-Otín 12-hallmark frame surfaced as first-class matrix nodes: ECM crosslinks (GlycoSENS) anchored to glucosepane + advanced-glycation-end-products + age-crosslink-breakers; intracellular indigestible aggregates (LysoSENS) anchored to lipofuscin + 7-ketocholesterol. Each with inbound + outbound edges to existing hallmark nodes. Both low tractability.

3. experiments/ directory created with:

   • experiments/README.md — folder index with Dataview by-status + by-edge queries; lifecycle documentation
   • 5 initial proposal pages drafted (all status: proposed):
     • experiments/midas-cell-type-stratification.md — resolves the disputed [[mitochondrial-dysfunction]] → [[cellular-senescence]] edge in cell-type-stratified primary human cultures (HUVEC + HAEC + iPSC-cardiomyocyte + dermal fibroblast); small-lab scale; the only edge in the current matrix with a populated resolving-experiment: field at this round
     • experiments/age-breaker-lcms-replication.md — modern LC-MS replication of Yang 2003 on aged human cadaver tissue across 5–6 candidate AGE-breakers; resolves whether any current AGE-breaker class cleaves glucosepane in real tissue; small-lab scale + LC-MS access
     • experiments/faod-directed-evolution-glucosepane.md — engineer FAOD via yeast-surface-display directed evolution to fill the genuinely-empty Tier-3 (mature-crosslink cleavage) defense slot in biology; small-lab scale; gated on synthetic-chemistry collaboration for glucosepane fluorogenic substrate
     • experiments/senescent-cell-pet-tracer.md — atlas-mining → surface-marker shortlist → radiolabeled ligand → PET; resolves the biomarker-stratification bottleneck for senolytic clinical trials (Farr 2024 quercetin precedent); large-clinic scale overall but Phase 0 (atlas mining + bench IHC validation) is small-lab tractable as a year-1 standalone deliverable
     • experiments/glucosepane-monoclonal-antibody.md — hybridoma campaign with counter-screen vs pentosidine / CML / CEL / MG-H1; enables population-scale + intervention-screen glucosepane quantification; paired with the LC-MS replication and the directed-evolution work as a three-pronged ECM-crosslink program

4. index.md updated with new Experiments section + Dataview status query, placed after Studies, before Open hypotheses.

Cross-cutting framing surfaced during the brainstorming:

The matrix’s leverage ranking suggests three classes of blocker:

• Node blockers — high-out-degree + low-tractability nodes (genomic instability, telomere attrition) — drive cascade but no human intervention
• Edge blockers — disputed / weak edges where direction matters for intervention sequencing (mito↔senescence, AIC→stem-cell-exhaustion via GDF11)
• Missing-node blockers — damage classes outside the hallmark frame that feed into hallmark nodes (GlycoSENS, LysoSENS)

This taxonomy is captured in causal-graph-data § Edge-blocker tracker + Missing-node blockers. Future seeder/verifier passes on this layer should follow standard wiki discipline (verifier reads primary sources for any quantitative-claim updates; resolving-experiment links should be kept current as proposals advance through the status lifecycle).

→ direct user request (ad-hoc R45-stream addition; not part of a numbered seeding round)

[2026-05-22] seed | interventions/dietary/probiotics — ad-hoc, low-gas-strain focus

User clinical-question driven: post-Tuesday-5/19 high-fiber-load adaptation gas (“hot” flatulence) prompted a question about probiotic evidence specifically for low-gas-producing strains and cross-feeding strategies. Ad-hoc seed (not on ROADMAP per workflow convention); not added to ROADMAP queue.

• added: interventions/dietary/probiotics.md — type:intervention; mode:dietary; clinical-stage:implemented; human-evidence-level:limited; safety-profile:well-established; translation-gap:biomarker-only; clinical-trials-active:401 (ClinicalTrials.gov v2 API); literature-checked-through:2026-05-22; verified:false. Sections: scope + class boundaries, 2020 Zheng taxonomic reclassification, mechanism classes, aging-specific evidence (cross-linked to Hutchinson 2021 / Kim 2021 / Asaoka 2022 on existing pages — not duplicated), load-bearing low-gas section (fermentation-type taxonomy with gas-output table for heterofermentative vs homofermentative Lactobacillales vs Bifidobacterium bifid-shunt, H2-cross-feeding strategies with gap flag for absence of marketed hydrogenotrophic probiotic, strain-by-strain RCT evidence table for LP299v / B. infantis 35624 / B. lactis HN019 / LGG / multistrain / S. boulardii, in vitro gas data gap, fiber-adaptation-phase scenario), Goodoory 2023 (82-RCT meta-analysis) as current evidence standard explicitly downgrading older positive single-trial signals, AGA 2020 against-IBS posture, safety, regulatory context.
• propagated: hallmarks/dysbiosis.md — “Probiotics” and “Prebiotics” therapeutic-table entries wikilinked to intervention pages; probiotics row updated with AGA 2020 + Goodoory 2023 framing.
• propagated: interventions/dietary/prebiotics.md — “Related intervention classes” table Probiotics row now links to the new intervention page (was only linking to genus pages).
• propagated: microbiome/bifidobacterium.md — see-also section adds probiotics above the prebiotics/postbiotics/fmt cluster.
• propagated: microbiome/lactobacillus.md — see-also section adds probiotics above prebiotics/postbiotics/fmt cluster.
• updated: frameworks/interventions-by-hallmark.md § Dysbiosis class-level synthesis — removed the “lack compound pages” coverage-gap framing (probiotics now seeded); retained needs-tagging for urolithin-a → dysbiosis.

Recency search findings worth surfacing:

• Goodoory 2023 (82-RCT meta-analysis, Gastroenterology) downgrades LP299v + B. infantis 35624 evidence to low or very low certainty — contradicts the older positive single-trial signals (Whorwell 2006, Ducrotté 2012). Both perspectives are presented; the wiki does not default to the older view.
• No marketed hydrogenotrophic probiotic exists — flagged #gap/no-approved-hydrogenotrophic-probiotic. Blautia hydrogenotrophica and related candidates are mechanistically promising but no commercial product has come to market.
• Anwar 2025, Zeng 2025 umbrella MA, Hidayat 2025 LGG MA all integrated as part of the 2024-2025 evidence refresh.

Suggested verifier priority (per seeder return summary):

1. Whorwell 2006 (locally available) — dose-specificity (10^8 only) + bloating endpoint
2. Hutchinson 2021 (locally available) — n=17 RCT count + microbiota finding
3. Goodoory 2023 (hybrid OA, DOI lookup failed) — priority download for LP299v “low certainty” grade
4. Ducrotté 2012 (hybrid OA) — n=214 + 78.1%/8.1% response figures
5. Su 2020 AGA guideline (closed-access) — verify “against IBS” recommendation language
6. DOI confirmation for gap-flagged footnotes (Hidayat 2025, Nordström 2021, Winham 2011, Yuan 2017, Cochrane 2025)

Files touched: interventions/dietary/probiotics.md (created), hallmarks/dysbiosis.md, interventions/dietary/prebiotics.md, microbiome/bifidobacterium.md, microbiome/lactobacillus.md, frameworks/interventions-by-hallmark.md, log.md.

---

2026-05-22 — verify: interventions/dietary/probiotics.md

Page: interventions/dietary/probiotics.md (seeded 2026-05-22 in response to user clinical question on low-gas probiotic strains during high-fiber Mediterranean diet adaptation)

Sources verified against PDFs:

• Whorwell 2006 (doi:10.1111/j.1572-0241.2006.00734.x) — full PDF read; claims confirmed
• Hutchinson 2021 (doi:10.3390/microorganisms9061344) — full PDF read; n=17 RCTs, 11 assessed microbiota (10/11 significant), heterogeneity precluded meta-analysis — confirmed
• Ducrotté 2012 (doi:10.3748/wjg.v18.i30.4012) — full PDF read; brief-article format confirmed; dose 10^10 CFU/capsule added
• Goodoory 2023 (doi:10.1053/j.gastro.2023.07.018) — full PDF read; LP299v bloating-endpoint treatment corrected (not separately analyzed; NS under “All Lactobacillus strains”)
• Hill 2014 (doi:10.1038/nrgastro.2014.66) — full PDF read; definition wording confirmed exact match
• Zeng 2025 (doi:10.1186/s40001-025-02788-w) — full PDF read; OR analysis NS (OR 1.60) disclosed alongside confirmed RR 0.74

Abstract-only verified (closed-access):

• Su 2020 AGA (doi:10.1053/j.gastro.2020.05.059) — not_oa; claims verified via abstract

Gap DOIs confirmed via PubMed efetch:

• Hidayat 2025: doi:10.1039/d5fo01780g (confirmed)
• Nordström 2021: doi:10.3920/BM2020.0191 (confirmed)
• Winham 2011: doi:10.1186/1475-2891-10-128 (confirmed; title confirmed as “Perceptions of flatulence from bean consumption”)
• Yuan 2017: doi:10.1080/03007995.2017.1292230 (confirmed)
• Cochrane CDiff: corrected to doi:10.1002/14651858.CD006095.pub5 (was missing.pub5 suffix)

Corrections made: 8 substantive corrections + 5 DOI confirmations — see verifier summary in main conversation.

Supersession check: no post-2023 B. infantis 35624 or LP299v meta-analyses/RCTs found that supersede existing framing; 103 elder-probiotic SR/MAs found post-2022 but none directly targeting healthy elderly in a way that supersedes Hutchinson 2021 (spot-checked top 3); hydrogenotrophic probiotic: 3 results, none marketed as probiotic.

clinical-trials-active: 401 — re-confirmed via ClinicalTrials.gov v2 API 2026-05-22.

literature-checked-through: 2026-05-22 — already set by seeder; confirmed current.

Final state: verified: true (partial scope — Anwar 2025 full PDF not verified; Su 2020 abstract-only)