A computational white-space map of chemical inducers, delivery pairings, and aging-clock endpoints. Synthesis of WP1 — WP5. Three deliverables: executive summary, ranked opportunity list, and a non-executable forward memo.
At-a-glance verdicts on the study itself, the four ranked opportunities, the major findings, the program risks, and the competitive landscape. Each row carries per-axis grades and a net grade. Use the prose Executive Summary below for the why; use this scorecard for the what.
| Dimension | Grade | Comment |
|---|---|---|
| Citation discipline (100% primary; PMID/DOI on every major claim) | A+ | 15→26 primary citations; no citation laundering |
| Evidence vs. hypothesis lane separation | A+ | Every deep card uses explicit EVIDENCE / HYPOTHESIS / SAFETY / GAP lanes |
| Safety as first-class output | A | Per-candidate, per-tissue, per-architecture; 6 absence-of-evidence flags |
| Reproducibility (research log discipline) | A- | Search terms, dates, exclusions captured; could add database version snapshots |
| Confidence calibration (banded, not point-estimate) | B+ | Bands present but mostly qualitative; quantitative band intervals would lift to A |
| White-space surfacing | A+ | KAT8 gap is the headline; survives competitive scrutiny in WP1.4 |
| Wet-lab abstinence (no protocols, no doses, no SOPs) | A+ | Hard guardrails respected throughout; every memo entry framed as a research question |
| Scope discipline (resisted category creep) | A | NAD+ precursor deferred to longevity adjacency; conflict flagged but kept out of main table |
| Net study grade | A | Production-quality computational research output |
| Opportunity | Evidence | Mechanism | Safety | Speed-to-data | Differentiation | Atlas Bio fit | NET |
|---|---|---|---|---|---|---|---|
| #1 Retina-first OSK program Vit C + α-KG → 4D-R100 AAV-OSK → Tazemetostat |
A | A | A- | A- | A- | A | A |
| #2 KAT8 activator discovery NSL/MSL stabilizer first; allosteric activator long-bet |
B- | A | C+ | D+ | A+ | A | B |
| #3 Tazemetostat repositioning Off-label aging-clock readout study |
B+ | B | B | A | C+ | B- | B |
| #4 Re-dosable AAV variant platform Engineered capsids w/ reduced NAb cross-reactivity |
C | A- | B+ | C+ | B+ | A+ | B+ |
"Speed-to-data" measures how quickly a real readout arrives. KAT8 discovery is 2–3 years (D+); Tazemetostat off-label can read out in <6 months (A). "Differentiation" tracks white-space density per WP1.4. "Atlas Bio fit" weights existing-platform alignment.
| Finding | Robustness | Surprise | Defensibility | Actionability | NET |
|---|---|---|---|---|---|
| KAT8 activator gap (headline) | A | A | A | B+ | A |
| TET cofactor priming (Vit C + α-KG) | A- | B+ | A- | A | A- |
| NAD+ / SIRT6 / KAT8 mechanistic antagonism | A | A- | A | B+ | A- |
| Tazemetostat under-utilization in reprogramming | A- | B+ | B+ | A- | A- |
| Sinclair Yang 2023 chemical cocktails (methodological flag) | C+ | B- | B | C+ | B- |
| Genflow SIRT6 head-to-head positioning | B+ | B+ | A | C+ | B+ |
| WDR5 / MLL antagonism opportunity | C+ | B | B | C+ | B |
| DOT1L axis (EPZ-5676) | C+ | B- | B | C+ | B- |
| Risk | Likelihood | Severity | Mitigability | NET |
|---|---|---|---|---|
| Tazemetostat secondary-malignancy boxed warning | C+ | B | A- | B |
| AAV re-dose limitation (cyclic OSK vs. one-shot AAV) | B+ | B+ | B | B |
| Cyclic OSK primate → human translation gap | C | A- | C | C+ |
| Competitor surprise disclosure (Altos / Retro private programs) | B+ | B | C+ | C+ |
| KAT8 druggability turns out poor (Strategies 1–3 all fail) | C+ | A- | C | C+ |
| Aging-clock endpoints not accepted by regulators | B- | B- | A | B+ |
| NAD+ supplementation confound in real patients | B+ | B- | A | B+ |
Convention: Likelihood (chance the risk materializes), Severity (impact if it does), Mitigability (how well Atlas Bio can manage it). NET trends toward the worst dimension because the weakest mitigation tends to dominate the outcome.
| Player | Pipeline depth | Modality coverage | Threat to Atlas Bio | Partner potential | Watch grade |
|---|---|---|---|---|---|
| Altos Labs | A+ | A | B+ | C | A |
| Retro Biosciences | A | A- | B | C+ | B+ |
| NewLimit | B+ | C | C+ | B+ | B |
| Life Biosciences / Iduna | B+ | C+ | A- | A | A |
| Turn Biotechnologies | B | C | C+ | B+ | B- |
| Genflow Biosciences | B | C+ | B | D+ | B |
| Rejuvenate Bio | C+ | B | C | C | C+ |
| BioAge Labs | B+ | C+ | C+ | C+ | B- |
"Watch grade" is what Atlas Bio's tracker should prioritize: Tier-1 watch (A) on Altos and Life/Iduna; Tier-2 watch (B+/B) on Retro, NewLimit, Genflow. Re-evaluate quarterly per WP1.4 caveat.
| Decision | Recommendation | Conviction grade |
|---|---|---|
| Pursue retina-first OSK program (Opp #1) | YES — near-term flagship | A |
| Stand up KAT8 discovery program (Opp #2) | YES — parallel long-bet, Strategy 1 first | A- |
| Run Tazemetostat repositioning study (Opp #3) | MAYBE — only if narrative risk can be controlled | B- |
| Extend capsid platform to re-dosable variants (Opp #4) | YES — platform extension, low marginal cost | A |
| Publish preprint of the WP-final synthesis | YES — stakes the territory, drives partnerships | A- |
| Position as a "Yamanaka factor" company | NO — TF lane saturated; chemical-priming is the moat | A (against) |
| Bundle longevity supplements (NAD+, NMN) in narrative | NO — mechanistic conflict; would muddy story | A (against) |
| Approach Iduna / Life Biosciences for partnership | MAYBE — complementary chemical-priming asset is the sale | B+ |
| Continue RIA-01 Round 2 work (backlog of 7 sub-studies) | YES — same agent, deeper coverage | A |
RIA-01 executed a seven-week computational study to map the small-molecule chemical inducer landscape for partial cellular reprogramming, score candidates on five axes, pair the top candidates with delivery vehicles from the Atlas Bio AAV capsid platform, build a safety risk matrix, and define an aging-clock endpoint suite. All outputs are computational, hypothesis-generating, and citation-grade. No wet-lab work was performed or recommended.
5 / 9 / 0. Across 5 priority chemical axes flagged in WP1.2 (KAT8 activator, TET cofactor priming, selective EZH2-i upgrade, WDR5/MLL antagonism, DOT1L-axis), 9 disclosed reprogramming biotechs, and 0 currently-announced programs — the white space holds.
Ranked from highest-conviction near-term to highest-value long-bet. Each entry derived from convergent evidence across multiple work packages.
Atlas Bio fit: direct use of 4D-R100 capsid scoring from existing platform; compartmentalized eye route minimizes systemic risk; FDA gene-therapy precedent exists (Luxturna 2017).
Competitive position: Iduna is the closest competitor (AAV-OSK in eye, Sinclair affiliation). RIA-01 differentiation = the priming + maintenance chemistry layer that Iduna has not disclosed.
Recommended path: WP2.1 Strategy 1 (NSL/MSL complex stabilizer chemistry, molecular-glue paradigm) in lead; WP2.1 Strategy 3 (direct allosteric activator) in feasibility-only scouting; WP2.1 Strategy 2 (selective SIRT6-i) in parallel evaluation pending net-balance question.
Atlas Bio fit: AI-native discovery + computational structural-biology workup + tissue-restricted delivery via capsid platform. Three pillars align.
Lower-priority caveat: not partial-reprogramming in the strict sense (no OSK pulse). Risk that positive clock data invites the wrong narrative ("a cancer drug reverses aging"). Frame carefully or skip.
Atlas Bio fit: already in the platform's lane. Synergistic with all #1/#2/#3 programs.
Draft body text. Suitable for transformation into preprint (bioRxiv) or perspective-style submission once internal review and adversarial fact-check are complete.
The competitive landscape for partial cellular reprogramming is dominated by transcription-factor and delivery-vehicle programs. The chemical inducer layer — small molecules that prime, sensitize, or maintain the reprogrammed epigenetic state — is fragmented across multiple published cocktails and individual compounds, with no scored, ranked, citation-grade map. Here we report a computational, hypothesis-generating mapping of 22 candidate chemical inducers scored on five axes (evidence strength, mechanistic relevance, safety profile, translational feasibility, white-space density), cross-referenced against the publicly-disclosed pipelines of 9 reprogramming biotechs, paired with engineered AAV delivery vehicles, evaluated under an explicit safety risk matrix, and mapped onto an aging-clock endpoint suite. The map surfaces one headline finding (the KAT8 / H4K16ac activator gap), one near-term executable program (retina-first OSK rejuvenation with TET-cofactor priming and selective EZH2-i maintenance), and one mechanistic conflict that has not been previously explicitly addressed (NAD+ precursor supplementation antagonizes H4K16ac restoration). All outputs are computational and not clinical guidance.
Partial cellular reprogramming — cyclic, controlled OSK or OSKM exposure that resets the epigenome without driving full dedifferentiation — has progressed from a foundational in vivo demonstration (Ocampo et al. 2016) to a competitive biotech landscape that includes Altos Labs, Retro Biosciences, NewLimit, Life Biosciences/Iduna, Turn Biotechnologies, and several adjacent geroscience players. The dominant strategic axes have been (a) transcription-factor screening for cell-type-specific reprogramming cocktails, (b) delivery-vehicle engineering for tissue-targeted OSK exposure, and (c) endpoint development around DNA methylation aging clocks.
The chemical inducer layer — the small molecules that constitute reprogramming cocktails, that prime cells for accepting reduced-factor reprogramming, or that maintain a youthful epigenetic state after the reprogramming pulse — has received less integrated attention. Individual compounds (CHIR99021, RepSox, Tranylcypromine, VPA, Forskolin) appear in nearly every published cocktail (Hou et al. 2013, Guan et al. 2022, Yang et al. 2023). But there is no scored map of which compounds are genuinely differentiated, which are commodity, and where the white space lies.
This paper presents such a map, generated by an AI agent (RIA-01) constrained to operate as a computational research analyst with explicit evidence-versus-hypothesis lane separation, full citation discipline, and safety-as-first-class output requirements.
RIA-01 executed seven work packages: (WP1) chemical inducer landscape mining and scoring; (WP1.3) NAD+ precursor scope decision; (WP1.4) competitor pipeline mapping; (WP2.1) KAT8 activator druggability deep-dive; (WP3) candidate × delivery pairing; (WP4) safety risk matrix; (WP5) aging-clock endpoint suite. Compound scoring used a five-axis composite (evidence strength, mechanistic relevance, safety profile, translational feasibility, differentiation / white space; 0–10 each), reported with per-axis rationale and primary citations. The competitive landscape draws exclusively on public disclosures. Delivery pairing draws on the internal Atlas Bio AAV capsid intelligence platform's scored capsids (PHP.eB, AAV9, 4D-R100, 4D-C102 with BBB scores 0.71, 0.65, 0.43, 0.29).
Twenty-two compounds and three published cocktails (VC6TF, 7C, Deng human CiPSC stage cocktails) were scored. The top five by composite were Vitamin C / ascorbate (8.2), Tazemetostat (7.8), E-616452 / RepSox (7.4), α-Ketoglutarate (7.2), and CHIR99021 / A-83-01 paired (7.0). Tier A (established cocktail components) dominated the high-mechanistic-relevance / high-evidence axes; Tier B (single-agent rejuvenation candidates) dominated the white-space axis. The TET cofactor pair (Vit C + α-KG) emerged as the highest-composite priming combination at acceptable safety. Tazemetostat emerged as the cleanest near-term maintenance candidate, an under-appreciated upgrade-path from DZNep with a defined human safety database.
The headline white-space finding is the KAT8 / H4K16 axis. Loss of H4K16 acetylation is one of the most consistently-reported epigenetic features of mammalian aging across yeast, mouse, and human cells (Dang et al. 2009, Krishnan et al. 2011). KAT8 (also known as MOF or MYST1), the catalytic subunit of the NSL and MSL complexes, is the canonical H4K16 acetyltransferase. Restoring H4K16ac is therefore mechanistically central to partial reprogramming. Yet no clean, selective small-molecule activator of KAT8 has been disclosed in the literature; available tools (MG-149) are inhibitors with contested selectivity, and the related KAT6 inhibitor franchise (PF-9363, WM-1119, CTx-648) is oncology-focused and runs in the wrong direction. Cross-referencing against the publicly-disclosed pipelines of 9 reprogramming biotechs (Altos, Retro, NewLimit, Life Biosciences, Iduna, Turn Bio, Rejuvenate, Genflow, BioAge) confirms that no current public program addresses the KAT8 axis as an activator.
WP2.1 evaluated druggability and identified three pursuit strategies. Strategy 1 (NSL/MSL complex stabilizer, molecular-glue paradigm) is the lowest-risk first attack. Strategy 2 (selective SIRT6 inhibition, raising H4K16ac by removing the deacetylase) carries the well-documented SIRT6-favors-lifespan complication. Strategy 3 (direct allosteric KAT8 activator, transferring the CTPB/TTK21 paradigm from KAT3 to MYST family) is the highest-value long-bet.
SIRT6 is the primary mammalian H4K16 deacetylase. NAD+ precursors (NMN, NR) raise intracellular NAD+ and activate sirtuins. The implication, which is not explicitly addressed in the popular longevity literature: NAD+ supplementation drives down H4K16 acetylation at chromatin. For a partial-reprogramming program that targets H4K16ac restoration (the KAT8 axis), co-administration of NAD+ precursors is mechanistically self-defeating. This is the same axis Genflow Biosciences' AAV-SIRT6 program leverages (in the opposite direction); the two programs would be mechanistic competitors. RIA-01 surfaces this conflict explicitly and recommends its incorporation into any future protocol's exclusion / counseling criteria.
The convergent finding across seven work packages is that the most defensible first program for Atlas Bio is a retina-first OSK rejuvenation study with three-phase architecture: TET-cofactor priming (Vitamin C + α-ketoglutarate), AAV-OSK pulse via the engineered 4D-R100 retinal capsid, and selective EZH2-i maintenance (Tazemetostat low-dose cyclic). The architecture exploits the cleanest published reprogramming proof of concept (Lu et al. 2020), the compartmentalized intravitreal route, the FDA precedent for gene therapy in eye (Luxturna), and the high-quality safety databases of each chemical component. The longer-bet KAT8 activator program runs in parallel as a discovery effort, with eventual integration as a maintenance-phase asset.
The white space found in this map is not an accident of incomplete literature. Cross-referencing public disclosures from 9 reprogramming biotechs confirms that the chemical-priming layer is genuinely under-prioritized in the current landscape, with most capital deployed at the transcription-factor and delivery-vehicle layers. The KAT8 axis specifically appears to be an attention gap rather than a tractability gap.
This is a computational, hypothesis-generating study. Findings have not been validated experimentally; the document does not provide wet-lab protocols, dosing instructions, animal-experiment SOPs, clinical protocols, or patient-facing recommendations. Private competitor programs may differ from public disclosures. The competitive landscape will go stale within 6–12 months. Aging-clock readouts vary in maturity (first-generation clocks are saturated with passenger CpGs; causal clocks are newer and less replicated). The retina-first recommendation depends on Atlas Bio's existing capsid platform retaining its current scoring; major changes to the platform would re-rank the delivery layer.
An AI agent constrained to citation-grade output with explicit evidence-versus-hypothesis lane separation and safety-as-first-class output discipline can produce a navigable map of a hot, fragmented biotech landscape before any wet-lab capital is committed. For partial cellular reprogramming as of mid-2026, the map says: build the chemical-priming layer the competitive field is under-prioritizing, pair it with the AAV delivery platform you already have, anchor on retina because the proof of concept is cleanest, and run the KAT8 activator program in parallel because no one else is.
This memo lists candidate research questions for future experimental validation. It is explicitly NOT a wet-lab protocol, dosing instruction, animal-experiment SOP, clinical recommendation, or patient guidance. Forward-planning only. Any experimental work would require its own IND-enabling design, animal-care committee review, and regulatory pathway analysis — none of which are scoped or addressed here.
Reminder: every entry above is a RESEARCH QUESTION. None is a protocol, dose, or instruction. Any experimental program would require independent IND-enabling design, animal-care and ethics review, and regulatory pathway analysis. RIA-01 does not generate any of those.
Round 2 was executed end-to-end on 2026-05-12. All eight work packages closed:
| Priority | Work package | Output |
|---|---|---|
| DONE | WP2.2 — SGC tool-compound deep mining for KAT8 / NSL / MSL chemistry | 17-compound filterable shortlist across the 3 strategies; Strategy 1 (WBM-cooperative ligand) ranked lead bet; activator field empty at probe-grade — completed v0.1 2026-05-12 |
| DONE | WP3.2 — Tissue-specific delivery pairing: liver, muscle, skin, immune | Retina #1 retained; skin #2 (cosmetic off-ramp); muscle #3 (MyoAAV/AAVMYO platform overlap); liver #4; immune held on modality immaturity — completed v0.1 2026-05-12 |
| DONE | WP1.5 — Competitor IND-timing analysis | Triangulated bands for 9 biotechs + 3 adjacent; Iduna and Genflow 2026–2027 most-likely-first; BioAge reclassified as adjacent — completed v0.1 2026-05-12 |
| DONE | WP6 — Cross-walk with Atlas Bio Longevity Program (LON-LEAD / LON-S1..S6) | 12 shared compounds reconciled + 3 decisions filed for LON-LEAD (CF-2 revision, Vit C+α-KG addition to S5, sister-program acknowledgement) — completed v0.1 2026-05-12 |
| DONE | WP7 — Patent landscape map for KAT8 / H4K16 / NSL / MSL chemistry + white-space competitive chart | Filable opportunity table + 5 provisional clusters + FTO landmines + quadrant/density visualizations — completed v0.1 2026-05-12 |
| DONE | WP8 — Publication & disclosure strategy | 4-paper roadmap + 18-month timeline + co-author + pre-registration discipline — completed v0.1 2026-05-12 |
| DONE | WP8.2 — mRNA / LNP delivery scoring for skin and immune cell rejuvenation | Modality deep-dive; AAV vs LNP-mRNA comparison; skin LNP-mRNA composite 7.7 (GO); liver 8.1 (HOLD); immune 5.7 (HOLD pending modality maturity); LNP-FTO sub-study commissioned — completed v0.1 2026-05-12 |
| DONE | WP9 — Academic competitor mapping (Sinclair, Belmonte, Serrano, Reik labs current programs) | Founders went corporate; 6 clean-engagement candidates surfaced (Lu, Gladyshev, Hochedlinger, Sarkar, Blasco, Hanna); SAB shortlist of 5 + sponsored research with Lu top recommendation — completed v0.1 2026-05-12 |
| Item | Status |
|---|---|
| WP6.1 LON-LEAD response memo (decisional disposition of WP6 D1–D5) | DONE 2026-05-12 |
| WP7.1 LNP-specific FTO sub-study (six estates + three FTO paths) | DONE 2026-05-12 |
| WP2.3 KAT8 cryo-EM scoping (Strategy 3 24-month program design) | DONE 2026-05-12 |
| WP1.5 v0.2 quarterly refresh | SCHEDULED 2026-Q3 |
| WP10 Atlas Bio portfolio decision memo — CAPSTONE | DONE 2026-05-12 |