Check the evidence substrate.
Adjudicate 30 experiment/control units. Check negative-study coverage, data rights and overlap with published model training. Narrow the question if an independent holdout is infeasible.
AGING RESEARCH / 15 SEPTEMBER 2026
Longer life.
Stronger evidence.
Map the mechanisms. Test whether the effects transfer. Invest in the next result that could change our understanding.
The mission is durable human rejuvenation. The evidence is narrower.
Biomarker change, restored function, longer animal lifespan and sustained human survival are separate outcomes. This review establishes no way to defeat aging entirely. It is a targeted research map, not an exhaustive systematic review or a treatment guide.
THE RESEARCH MAP
Different mechanisms.
Different levels of proof.
Each budget funds a specific computational test. Probabilities are uncalibrated judgments about that test, conditional on usable data and staffing.
DNA sequence is one input. Cell identity, regulatory state, tissue context and accumulated damage also determine what an organism can repair. The Hallmarks framework helps organize these mechanisms; it is not a complete causal blueprint.
Compare all 22 paths: evidence, test cost, odds and market
| Research path | Evidence anchor | Next-test cost | P(test passes) | Annual US tool TAM |
|---|---|---|---|---|
| 01 · Does an intervention transfer? ↘First test | Executed data audit | $18,3244–6 weeks | 30%15–45% judgment | $16.0M0.115% cost / TAM |
| 02 · Partial cellular reprogramming ↘High-upside test | Mouse function; human phase 1 | $31,5846–8 weeks | 20%8–35% judgment | $36.0M0.088% cost / TAM |
| 03 · Senescent-cell removal and control ↘Exploratory | Mouse survival; mixed human trial | $24,8645–7 weeks | 25%10–40% judgment | $24.0M0.104% cost / TAM |
| 04 · Nutrient sensing and mTOR ↘Benchmark anchor | Multisite mouse survival; human trial | $18,3244–6 weeks | 40%20–60% judgment | $12.8M0.143% cost / TAM |
| 05 · Dietary restriction and metabolic resilience ↘Exploratory | Mouse survival; human biomarkers | $24,8045–7 weeks | 30%15–45% judgment | $9.6M0.258% cost / TAM |
| 06 · Metabolic drug repurposing ↘Exploratory | Disease-specific human outcomes | $21,4444–6 weeks | 25%10–40% judgment | $25.6M0.084% cost / TAM |
| 07 · Autophagy, lysosomes and proteostasis ↘Exploratory | Genetic mouse lifespan experiment | $21,6244–6 weeks | 25%10–40% judgment | $16.0M0.135% cost / TAM |
| 08 · Mitochondrial quality and mitophagy ↘Exploratory | Human randomized biomarker studies | $21,5644–6 weeks | 25%10–40% judgment | $12.8M0.168% cost / TAM |
| 09 · NAD and cellular energy pathways ↘Exploratory | Human target engagement; limited function | $14,6043–4 weeks | 45%25–65% judgment | $4.0M0.365% cost / TAM |
| 10 · DNA maintenance and cancer restraint ↘Exploratory | Comparative mammalian evidence | $31,4646–8 weeks | 20%8–35% judgment | $24.0M0.131% cost / TAM |
| 11 · Telomere maintenance ↘Exploratory | Mouse lifespan experiment | $25,5845–7 weeks | 20%8–35% judgment | $9.6M0.266% cost / TAM |
| 12 · Stem-cell renewal and tissue regeneration ↘Exploratory | Mouse immune-function experiment | $28,2245–7 weeks | 25%10–40% judgment | $28.8M0.098% cost / TAM |
| 13 · Immune renewal and chronic inflammation ↘Priority after benchmark | Mouse survival; human associations | $28,8245–7 weeks | 30%15–45% judgment | $28.8M0.100% cost / TAM |
| 14 · Extracellular matrix, fibrosis and crosslinks ↘Exploratory | Transferred gene: mouse function and survival | $22,3444–6 weeks | 20%8–35% judgment | $16.0M0.140% cost / TAM |
| 15 · Circulating and endocrine signals ↘Exploratory | Mouse multi-tissue and survival evidence | $25,0445–7 weeks | 20%8–35% judgment | $16.0M0.157% cost / TAM |
| 16 · Microbiome and host interactions ↘Exploratory | Mouse cognition and immune effects | $21,6844–6 weeks | 20%8–35% judgment | $11.2M0.194% cost / TAM |
| 17 · Organ replacement and preservation ↘Exploratory | Human case report; animal preservation | $26,1845–7 weeks | 20%8–35% judgment | $12.0M0.218% cost / TAM |
| 18 · Brain and neurovascular preservation ↘Exploratory | Disease-specific human trial; mouse function | $31,5846–8 weeks | 15%5–30% judgment | $36.0M0.088% cost / TAM |
| 19 · Comparative longevity genomics ↘High-upside test | Cross-species association plus perturbation example | $31,5246–8 weeks | 20%8–35% judgment | $36.0M0.088% cost / TAM |
| 20 · Aging measurements that predict outcomes ↘Priority after benchmark | Human associations; intervention-response models | $28,4645–7 weeks | 25%10–40% judgment | $25.6M0.111% cost / TAM |
| 21 · Combinations and sequential repair ↘Exploratory | Mouse lifespan combination study | $25,7645–7 weeks | 15%5–30% judgment | $19.2M0.134% cost / TAM |
| 22 · Proven function-preserving comparators ↘Exploratory | Randomized human functional outcomes | $12,6242–4 weeks | 50%30–70% judgment | $4.0M0.316% cost / TAM |
22 paths
01Evidence infrastructure / First testDoes an intervention transfer?Executed data audit
What the evidence supports
Our DrugAge audit finds 23 compound names with mouse observations in at least two publications; 20 also cover both sexes. Publications do not establish independent replication.
Human translation
The audit contains no human intervention outcomes and estimates no human efficacy.
The difficult part
Positive-literature selection, dependent controls, chemical aliases and study leakage can inflate apparent confidence.
DrugAge bulk release and documentation ↗ · DrugAge database publication ↗ · NIA Interventions Testing Program ↗ · Prior DrugAge sex-specific mouse analysis ↗
The next public-data test
Build a publication-linked intervention-outcome benchmark with a separately searched negative-study set and an untouched publication-date holdout.
Pass condition
Adjudicate at least 150 independent experiment-control units with ≥95% source precision; a locked model reduces Brier score by ≥10% versus the strongest simple baseline on at least 50 unseen units, with a study-cluster bootstrap 95% interval excluding zero improvement.
Stop or narrow the claim
If independence or negative-study coverage cannot be established, publish a coverage dataset only. If calibration fails, stop prediction claims.
Fully costed next test$18,324
$11,911–$29,318 sensitivity range · 4–6 weeks. Tools/data cash allowance: $564; contributed labor still has economic cost.
P(this test passes)30%
15–45% judgment range, low confidence. The ingestion is working; independence and negative-study recovery are unresolved. This forecast is for the composite benchmark gate, not intervention efficacy.
Annual R&D-tool TAM scenario$16.0M
320 assumed buyers × $50,000/year. Auditable preclinical evidence and replication planning. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
100 scientific hours × $100 + 32 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $250 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $25,000–$100,000 annual price imply $4.0M–$64.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
02Repair cell state / High-upside testPartial cellular reprogrammingMouse function; human phase 1
What the evidence supports
OSK restored visual function in mouse experiments. This is organ-specific functional evidence, not whole-body lifespan reversal.
Human translation
ER-100 is registered as recruiting phase 1 in optic neuropathies: estimated 18 participants, safety-focused primary outcomes, no posted results in the checked record. Last registry update: May 19, 2026.
The difficult part
Preserving cell identity, controlling delivery and excluding proliferative or cancer risk are central barriers. A younger clock can coexist with damaged function.
Reprogramming and restored vision in mice ↗ · ER-100 phase 1 optic-neuropathy trial ↗ · Cross-species transcriptomic aging and mortality models ↗
The next public-data test
Reanalyze public reprogramming expression data across tissues; separate youthful-state recovery from loss of mature cell identity.
Pass condition
On two untouched studies, improve prediction of independently measured functional recovery by ≥10% over age-only and cell-identity baselines, with a positive study-level uncertainty interval; identity-loss controls must be correctly flagged.
Stop or narrow the claim
If functional labels or two independent studies are unavailable, stop at dataset feasibility. Clock reversal alone cannot pass.
Fully costed next test$31,584
$20,530–$50,534 sensitivity range · 6–8 weeks. Tools/data cash allowance: $1,344; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. Mechanistic and clinical activity support research interest; sparse matched function labels and strong existing models make this test difficult.
Annual R&D-tool TAM scenario$36.0M
480 assumed buyers × $75,000/year. Reprogramming-response and identity-preservation assay analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
180 scientific hours × $100 + 48 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $900 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 240–960 assumed buyers and $37,500–$150,000 annual price imply $9.0M–$144.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
03Remove accumulated damage / ExploratorySenescent-cell removal and controlMouse survival; mixed human trial
What the evidence supports
Senolytic studies report improved function and survival in old mice; targeted CAR T approaches show durable metabolic effects in mice.
Human translation
A 60-participant randomized bone study did not meet its primary resorption-marker endpoint. Baseline-burden subgroup signals were exploratory.
The difficult part
Senescence is heterogeneous and can help wound healing and tumor suppression. Marker abundance is not a universal target or response predictor.
Senolytics: randomized bone-metabolism trial ↗ · Senolytics: function and survival in old mice ↗ · Senolytic CAR T cells in mice ↗
The next public-data test
Test whether baseline senescent-cell burden predicts measured response in public intervention datasets, withholding entire studies and tissue types.
Pass condition
≥10% lower response-prediction error than age plus cell-composition controls on two external studies, with a positive study-cluster interval and prospectively fixed burden markers.
Stop or narrow the claim
Stop if only post-treatment markers or exploratory human subgroups supply labels. No claim that all senescent cells should be eliminated.
Fully costed next test$24,864
$16,162–$39,782 sensitivity range · 5–7 weeks. Tools/data cash allowance: $864; contributed labor still has economic cost.
P(this test passes)25%
10–40% judgment range, low confidence. Biological effects exist, but target heterogeneity and small human studies limit transfer.
Annual R&D-tool TAM scenario$24.0M
480 assumed buyers × $50,000/year. Senescence-response stratification for R&D. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
140 scientific hours × $100 + 40 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $500 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 240–960 assumed buyers and $25,000–$100,000 annual price imply $6.0M–$96.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
04Slow damage accumulation / Benchmark anchorNutrient sensing and mTORMultisite mouse survival; human trial
What the evidence supports
Rapamycin is among the interventions with lifespan evidence in the NIA ITP, a three-site program using genetically heterogeneous mice.
Human translation
PEARL reported no significant overall change in visceral adiposity, its registered primary endpoint; some secondary sex-specific outcomes changed. It did not test human lifespan.
The difficult part
Effects depend on context, and immune, metabolic and tolerability tradeoffs matter. Mouse survival cannot supply human odds.
NIA Interventions Testing Program ↗ · ITP positive and negative publications ↗ · PEARL: human rapamycin trial results ↗ · PEARL registered outcomes ↗
The next public-data test
Audit available ITP survival data with site, sex, censoring and shared-control structure preserved; test whether effects transfer to an unseen site or cohort.
Pass condition
Reproduce published survival direction, then predict held-out site/sex effects with ≥10% lower error than the pooled intervention-only baseline and calibrated uncertainty.
Stop or narrow the claim
If unseen-site performance does not improve, retain a reproducibility report without a response-prediction claim.
Fully costed next test$18,324
$11,911–$29,318 sensitivity range · 4–6 weeks. Tools/data cash allowance: $564; contributed labor still has economic cost.
P(this test passes)40%
20–60% judgment range, low confidence. Replicated mouse data offer a relatively tractable analysis; predicting heterogeneity beyond pooled effects remains unproven.
Annual R&D-tool TAM scenario$12.8M
320 assumed buyers × $40,000/year. Preclinical trial analysis and response modeling. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
100 scientific hours × $100 + 32 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $250 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $20,000–$80,000 annual price imply $3.2M–$51.2M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
05Slow damage accumulation / ExploratoryDietary restriction and metabolic resilienceMouse survival; human biomarkers
What the evidence supports
A large genetically diverse female-mouse study found lifespan benefits with substantial individual variation. Sex and genetic context limit generalization.
Human translation
CALERIE methylation analysis found a small DunedinPACE change; different clocks did not agree uniformly. This is not evidence of added years of human life.
The difficult part
Weight loss, lower biological-age scores and resilience are different outcomes. Frailty, adherence and nutritional adequacy complicate translation.
Dietary restriction in genetically diverse female mice ↗ · CALERIE randomized-trial methylation analysis ↗
The next public-data test
Use the public diversity-outbred study to test baseline predictors of survival response, with all post-treatment information excluded from prediction.
Pass condition
On untouched families or genotype groups, ≥10% lower integrated survival Brier score than age, baseline weight and diet assignment, with a positive clustered interval.
Stop or narrow the claim
No responder claim if individual treatment effects are inferred from observed survival alone or benefits disappear after baseline-only restriction.
Fully costed next test$24,804
$16,123–$39,686 sensitivity range · 5–7 weeks. Tools/data cash allowance: $804; contributed labor still has economic cost.
P(this test passes)30%
15–45% judgment range, low confidence. Rich measurements help, but one female-only cohort cannot establish broad personalized efficacy.
Annual R&D-tool TAM scenario$9.6M
320 assumed buyers × $30,000/year. Experimental resilience and nutrition-response analytics. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
140 scientific hours × $100 + 40 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $450 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $15,000–$60,000 annual price imply $2.4M–$38.4M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
06Slow damage accumulation / ExploratoryMetabolic drug repurposingDisease-specific human outcomes
What the evidence supports
Canagliflozin extended lifespan in male, but not female, ITP mice. The result illustrates why sex cannot be discarded during aggregation.
Human translation
Empagliflozin reduced cardiovascular events and mortality in a high-risk type 2 diabetes trial. A disease-specific result is not general human aging reversal.
The difficult part
Indication, baseline disease and treatment selection create confounding; benefits in ill populations may not transfer to healthy adults.
Canagliflozin: sex-dependent mouse survival ↗ · Empagliflozin: cardiovascular outcomes in diabetes ↗ · ITP positive and negative publications ↗
The next public-data test
Build an endpoint-aligned comparison of animal and human metabolic-intervention results, keeping randomized and observational evidence separate.
Pass condition
≥95% endpoint/population precision in a blinded 100-assertion audit, plus improved held-out trial-outcome classification over indication-only baseline with a positive study-level interval.
Stop or narrow the claim
If only observational comparisons support transfer, stop causal claims; a useful evidence table is a narrower outcome.
Fully costed next test$21,444
$13,939–$34,310 sensitivity range · 4–6 weeks. Tools/data cash allowance: $564; contributed labor still has economic cost.
P(this test passes)25%
10–40% judgment range, low confidence. Human disease trials provide anchors, but cross-indication transfer is a much harder target.
Annual R&D-tool TAM scenario$25.6M
640 assumed buyers × $40,000/year. Endpoint and indication evidence services. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
120 scientific hours × $100 + 36 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $250 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 320–1280 assumed buyers and $20,000–$80,000 annual price imply $6.4M–$102.4M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
07Remove accumulated damage / ExploratoryAutophagy, lysosomes and proteostasisGenetic mouse lifespan experiment
What the evidence supports
A beclin-1 mutation that increased basal autophagy extended mouse lifespan and healthspan. Lifelong genetic change differs from late-life treatment.
Human translation
The cited experiment establishes no human lifespan benefit from broadly increasing autophagy.
The difficult part
Protein clearance is tissue-dependent; static marker levels do not measure flux, and excessive recycling can damage cells.
Beclin-1 autophagy and mouse longevity ↗ · Cross-species transcriptomic aging and mortality models ↗
The next public-data test
Compare public perturbation datasets using functional clearance labels; test whether the model separates increased flux from blocked degradation.
Pass condition
≥10% lower error on externally measured clearance outcomes than expression-only pathway scores on two untouched studies, with a positive clustered interval.
Stop or narrow the claim
Stop if available labels are only marker abundance or if flux direction cannot be distinguished.
Fully costed next test$21,624
$14,056–$34,598 sensitivity range · 4–6 weeks. Tools/data cash allowance: $744; contributed labor still has economic cost.
P(this test passes)25%
10–40% judgment range, low confidence. The causal animal foothold is real; public functional-flux labels are the main feasibility risk.
Annual R&D-tool TAM scenario$16.0M
320 assumed buyers × $50,000/year. Proteostasis assay interpretation. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
120 scientific hours × $100 + 36 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $400 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $25,000–$100,000 annual price imply $4.0M–$64.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
08Repair cell function / ExploratoryMitochondrial quality and mitophagyHuman randomized biomarker studies
What the evidence supports
A 66-person urolithin A trial found secondary muscle-endurance signals, but no significant between-group benefit on its walk-distance or ATP primary endpoints.
Human translation
A 2025 randomized 50-person study found immune-cell and metabolic changes over four weeks. It did not establish fewer infections or longer survival.
The difficult part
Improved mitochondrial markers may not restore meaningful function. Tissue context and baseline impairment affect response.
Urolithin A: randomized muscle-function trial ↗ · Urolithin A: randomized immune-biomarker trial ↗
The next public-data test
Test whether mitochondrial pathway changes predict functional improvement beyond baseline fitness and age in public intervention studies.
Pass condition
≥10% lower held-out functional-outcome error than baseline fitness plus age on two studies; primary and secondary endpoints remain separately scored.
Stop or narrow the claim
If only surrogate biomarkers are available, stop the functional prediction claim.
Fully costed next test$21,564
$14,017–$34,502 sensitivity range · 4–6 weeks. Tools/data cash allowance: $684; contributed labor still has economic cost.
P(this test passes)25%
10–40% judgment range, low confidence. Human measurements exist, but discordant endpoints and limited open participant data constrain the test.
Annual R&D-tool TAM scenario$12.8M
320 assumed buyers × $40,000/year. Mitochondrial response and assay analytics. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
120 scientific hours × $100 + 36 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $350 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $20,000–$80,000 annual price imply $3.2M–$51.2M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
09Repair cell function / ExploratoryNAD and cellular energy pathwaysHuman target engagement; limited function
What the evidence supports
An NMN randomized study increased blood NAD-related metabolites, while its primary stepping test showed no between-group difference.
Human translation
Biochemical target engagement is more established here than durable clinical benefit. Secondary walking and sleep results need independent confirmation.
The difficult part
Raising a metabolite is not proof that it is the limiting cause of decline; formulation, tissue uptake and endpoint choice differ.
The next public-data test
Audit preregistered endpoints against published results across public NAD-precursor trials, with separate target-engagement and functional-outcome labels.
Pass condition
≥95% precision on a blinded 100-assertion endpoint audit and retrieval of at least 10 usable randomized comparisons; no missing primary outcome reclassified as success.
Stop or narrow the claim
Stop a response-prediction program if comparable function data are insufficient; publish the endpoint audit, including nulls.
Fully costed next test$14,604
$9,493–$23,366 sensitivity range · 3–4 weeks. Tools/data cash allowance: $444; contributed labor still has economic cost.
P(this test passes)45%
25–65% judgment range, low confidence. This is an evidence-audit gate with limited biological inference, so it is more feasible than a therapeutic prediction.
Annual R&D-tool TAM scenario$4.0M
160 assumed buyers × $25,000/year. Trial endpoint auditing for R&D teams. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
70 scientific hours × $100 + 32 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $150 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 80–320 assumed buyers and $12,500–$50,000 annual price imply $1.0M–$16.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
10Preserve information / ExploratoryDNA maintenance and cancer restraintComparative mammalian evidence
What the evidence supports
Somatic mutation accumulation rates scale inversely with lifespan across studied mammals. This association motivates repair research without proving that changing one repair pathway extends human life.
Human translation
No cited study demonstrates safe organism-wide genome repair that reverses human aging.
The difficult part
Repair fidelity, cancer surveillance, mosaic mutations and delivery across trillions of cells are coupled problems.
The next public-data test
Reanalyze public comparative mutation and longevity data with phylogeny, tissue, body mass and study effects retained.
Pass condition
A candidate mechanism improves held-out-clade lifespan-related prediction by ≥10% over body mass and phylogeny, with a positive clade-bootstrap interval and independent perturbation support.
Stop or narrow the claim
If the association collapses under clade holdout or relies on one tissue, stop general repair claims.
Fully costed next test$31,464
$20,452–$50,342 sensitivity range · 6–8 weeks. Tools/data cash allowance: $1,224; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. Comparative data offer leverage, but small species counts and confounding make causal transfer uncertain.
Annual R&D-tool TAM scenario$24.0M
320 assumed buyers × $75,000/year. Comparative repair and target-evidence analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
180 scientific hours × $100 + 48 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $800 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $37,500–$150,000 annual price imply $6.0M–$96.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
11Preserve information / ExploratoryTelomere maintenanceMouse lifespan experiment
What the evidence supports
Telomerase gene therapy extended lifespan in an adult/old mouse experiment without increased cancer observed in that study.
Human translation
That result does not establish cancer safety or longevity efficacy for generally healthy humans.
The difficult part
Telomere dysfunction and malignant growth are competing concerns; human and mouse telomere biology differ.
Telomerase gene therapy in adult and old mice ↗ · Somatic mutations across mammalian lifespans ↗
The next public-data test
Integrate public telomere perturbation, damage and proliferation data to identify contexts where repair and proliferation signatures separate.
Pass condition
Predict held-out functional rescue while discriminating proliferative-risk controls at prespecified ≥90% sensitivity across two independent datasets.
Stop or narrow the claim
Stop if the only success measure is longer telomeres or if functional rescue cannot be separated from proliferation.
Fully costed next test$25,584
$16,630–$40,934 sensitivity range · 5–7 weeks. Tools/data cash allowance: $864; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. The mouse result supports investigation; safety proxies are incomplete and functional labels scarce.
Annual R&D-tool TAM scenario$9.6M
160 assumed buyers × $60,000/year. Telomere and proliferation-context analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
140 scientific hours × $100 + 44 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $500 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 80–320 assumed buyers and $30,000–$120,000 annual price imply $2.4M–$38.4M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
12Replace damaged components / ExploratoryStem-cell renewal and tissue regenerationMouse immune-function experiment
What the evidence supports
Depleting myeloid-biased stem cells in aged mice shifted immune production toward a younger pattern and improved immune responses.
Human translation
This preclinical result does not establish systemic rejuvenation from commercial stem-cell treatments.
The difficult part
An aged niche, clonal expansion, engraftment and malignant potential can defeat otherwise youthful cells.
Depletion of biased stem cells in aged mice ↗ · Transient restoration of immune cues in aged mice ↗
The next public-data test
Use public single-cell aging atlases to predict functional lineage balance across donors, separating cell composition from within-cell changes.
Pass condition
≥10% lower donor-held-out lineage/function error than cell-composition and age baselines, plus external validation in a perturbation dataset.
Stop or narrow the claim
Stop if donor leakage or changing cell mixture accounts for the gain; a younger average expression profile is insufficient.
Fully costed next test$28,224
$18,346–$45,158 sensitivity range · 5–7 weeks. Tools/data cash allowance: $1,104; contributed labor still has economic cost.
P(this test passes)25%
10–40% judgment range, low confidence. Cell-level public data are rich, but integration with functional lineage outcomes remains difficult.
Annual R&D-tool TAM scenario$28.8M
480 assumed buyers × $60,000/year. Regenerative-cell and niche quality analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
160 scientific hours × $100 + 44 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $700 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 240–960 assumed buyers and $30,000–$120,000 annual price imply $7.2M–$115.2M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
13Coordinate repair / Priority after benchmarkImmune renewal and chronic inflammationMouse survival; human associations
What the evidence supports
IL-11 inhibition extended mouse lifespan in a published study. A newer mRNA study restored immune cues and improved aged-mouse vaccine and antitumor responses.
Human translation
A 2026 thymic-imaging study links thymic health to human outcomes in two cohorts; it is observational, not proof that altering the thymus improves survival.
The difficult part
Suppressing inflammation can impair host defense. Immune rejuvenation must preserve infection control and self-tolerance.
IL-11 inhibition and mouse survival ↗ · Transient restoration of immune cues in aged mice ↗ · Thymic imaging and human health outcomes ↗ · Depletion of biased stem cells in aged mice ↗
The next public-data test
Test a tissue-aware immune-resilience signature against held-out functional challenge outcomes, retaining infection and inflammatory responses separately.
Pass condition
≥10% lower external-study error than age and immune-cell counts for functional response, with no degradation in prespecified host-defense risk classification.
Stop or narrow the claim
Stop if the signal predicts inflammation markers only or cannot distinguish protection from immune suppression.
Fully costed next test$28,824
$18,736–$46,118 sensitivity range · 5–7 weeks. Tools/data cash allowance: $984; contributed labor still has economic cost.
P(this test passes)30%
15–45% judgment range, low confidence. Functional perturbation studies create a useful target; human causal validation is still missing.
Annual R&D-tool TAM scenario$28.8M
480 assumed buyers × $60,000/year. Immune-resilience and trial-stratification analytics. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
160 scientific hours × $100 + 48 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $600 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 240–960 assumed buyers and $30,000–$120,000 annual price imply $7.2M–$115.2M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
14Remove accumulated damage / ExploratoryExtracellular matrix, fibrosis and crosslinksTransferred gene: mouse function and survival
What the evidence supports
Expression of naked mole-rat Has2 in mice increased hyaluronan, improved health measures and extended lifespan. This is one transferable-mechanism example.
Human translation
The cited work does not demonstrate a human crosslink-removal or whole-body matrix-rejuvenation treatment.
The difficult part
Matrix damage differs by organ. Removing scar or crosslinks without preserving tissue strength requires functional validation.
Naked mole-rat Has2 transferred to mice ↗ · Cross-species transcriptomic aging and mortality models ↗
The next public-data test
Compare aging and anti-fibrotic perturbation datasets using matrix mechanics or organ-function labels where available.
Pass condition
≥10% lower held-out functional/mechanical outcome error than fibrosis-marker expression alone on two independent studies.
Stop or narrow the claim
If only transcript levels are available, stop claims about restored tissue mechanics.
Fully costed next test$22,344
$14,524–$35,750 sensitivity range · 4–6 weeks. Tools/data cash allowance: $744; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. Cross-species transfer is demonstrated for one mechanism, but the relevant public mechanical measurements are sparse.
Annual R&D-tool TAM scenario$16.0M
320 assumed buyers × $50,000/year. Matrix and fibrosis response analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
120 scientific hours × $100 + 40 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $400 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $25,000–$100,000 annual price imply $4.0M–$64.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
15Coordinate repair / ExploratoryCirculating and endocrine signalsMouse multi-tissue and survival evidence
What the evidence supports
Prolonged heterochronic parabiosis produced sustained molecular changes and extended survival in old mice in a published study.
Human translation
This does not establish a benefit from human young-plasma products, because shared circulation includes many physiological changes beyond added factors.
The difficult part
Young factors, removal of harmful factors, shared organ function and surgery effects must be separated.
Youthful circulation and mouse lifespan ↗ · Parabiosis at single-cell resolution ↗
The next public-data test
Reanalyze public single-cell parabiosis data and independent factor-perturbation studies to test mechanism-specific transfer.
Pass condition
A frozen signature predicts an independent measured functional outcome beyond cell composition and generic injury response in two held-out studies.
Stop or narrow the claim
Stop if surgery, tissue mix or age labels explain the prediction; parabiosis results cannot be equated to plasma infusion.
Fully costed next test$25,044
$16,279–$40,070 sensitivity range · 5–7 weeks. Tools/data cash allowance: $1,044; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. Rich expression data are available; causal attribution to individual circulating signals remains a major gap.
Annual R&D-tool TAM scenario$16.0M
320 assumed buyers × $50,000/year. Systemic signaling and response-evidence services. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
140 scientific hours × $100 + 40 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $650 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $25,000–$100,000 annual price imply $4.0M–$64.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
16Coordinate repair / ExploratoryMicrobiome and host interactionsMouse cognition and immune effects
What the evidence supports
Young-donor microbiota altered immune/metabolic features and improved selected behavioral outcomes in aged mice.
Human translation
The cited study supplies no evidence that microbiota transfer reverses general human aging.
The difficult part
Diet, drugs, geography, cages and donor-recipient compatibility strongly affect composition and function.
The next public-data test
Test whether microbial functional pathways transfer across aging cohorts more reliably than taxonomic abundance, with cohort and host covariates withheld.
Pass condition
≥10% lower external-cohort functional-outcome error than diet, age and taxonomy baselines across at least two independent cohorts.
Stop or narrow the claim
Stop if the model identifies cohort, cage or diet rather than function; no causal claim from cross-sectional composition.
Fully costed next test$21,684
$14,095–$34,694 sensitivity range · 4–6 weeks. Tools/data cash allowance: $804; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. Data volume is large but functional endpoints and confounder control are inconsistent.
Annual R&D-tool TAM scenario$11.2M
320 assumed buyers × $35,000/year. Host-microbiome functional evidence analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
120 scientific hours × $100 + 36 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $450 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $17,500–$70,000 annual price imply $2.8M–$44.8M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
17Replace damaged components / ExploratoryOrgan replacement and preservationHuman case report; animal preservation
What the evidence supports
A gene-edited porcine kidney supported a living human recipient in a reported case. Rat kidney vitrification and nanowarming enabled life-sustaining transplantation.
Human translation
These are organ-specific advances with major immunological and durability constraints; a single case is not a general survival estimate.
The difficult part
Organ supply, vascular integration, rejection, infection and preservation damage remain bottlenecks; replacing peripheral organs does not preserve brain function.
Porcine kidney transplantation in a living recipient ↗ · Rat kidney cryopreservation and transplantation ↗
The next public-data test
Build an outcome-linked preservation and transplant evidence dataset with organ, species, preservation time and follow-up explicitly separated.
Pass condition
≥95% audited endpoint precision and ≥10% lower held-out-study prediction error for prespecified graft function than preservation-time/species controls.
Stop or narrow the claim
If compatible outcome definitions or independent cohorts are insufficient, stop predictive claims and retain the evidence resource.
Fully costed next test$26,184
$17,020–$41,894 sensitivity range · 5–7 weeks. Tools/data cash allowance: $744; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. Tangible function is measurable, but public samples are small and protocols differ substantially.
Annual R&D-tool TAM scenario$12.0M
160 assumed buyers × $75,000/year. Organ preservation and graft-evidence analytics. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
140 scientific hours × $100 + 48 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $400 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 80–320 assumed buyers and $37,500–$150,000 annual price imply $3.0M–$48.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
18Preserve function and identity / ExploratoryBrain and neurovascular preservationDisease-specific human trial; mouse function
What the evidence supports
Young CSF/FGF17 experiments improved memory-related outcomes in aged mice. Lecanemab slowed decline in early Alzheimer disease rather than restoring a youthful brain.
Human translation
The Alzheimer trial provides disease-specific clinical evidence with adverse-event tradeoffs; it does not establish broad cognitive rejuvenation.
The difficult part
Memory, circuitry, vascular integrity and protein pathology can fail separately. Peripheral rejuvenation cannot substitute for durable cognition.
Young CSF and memory in aged mice ↗ · Lecanemab in early Alzheimer disease ↗
The next public-data test
Test whether pathway-specific molecular changes predict independent cognitive or neuronal-function outcomes beyond age and disease severity.
Pass condition
≥10% lower donor/study-held-out functional error than age, disease stage and generic inflammation baselines, with a positive external-study interval.
Stop or narrow the claim
Stop if only molecular-age changes are available or the model predicts disease diagnosis rather than functional recovery.
Fully costed next test$31,584
$20,530–$50,534 sensitivity range · 6–8 weeks. Tools/data cash allowance: $1,344; contributed labor still has economic cost.
P(this test passes)15%
5–30% judgment range, low confidence. Functional translation is unusually hard and open longitudinal brain data are limited.
Annual R&D-tool TAM scenario$36.0M
480 assumed buyers × $75,000/year. Neurodegeneration target and outcome evidence. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
180 scientific hours × $100 + 48 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $900 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 240–960 assumed buyers and $37,500–$150,000 annual price imply $9.0M–$144.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
19Discover transferable mechanisms / High-upside testComparative longevity genomicsCross-species association plus perturbation example
What the evidence supports
Mammalian mutation-rate patterns and the Has2 transfer experiment support studying unusually long-lived species. They do not imply one universal longevity gene.
Human translation
Human benefit from a comparative target requires tissue-specific functional and safety validation.
The difficult part
Ancestry, body size, cancer adaptation, sampling and tissue state confound naive genome-to-lifespan correlations.
Somatic mutations across mammalian lifespans ↗ · Naked mole-rat Has2 transferred to mice ↗ · Cross-species transcriptomic aging and mortality models ↗
The next public-data test
Extend DNALYZER’s genome–phenome approach with longevity, repair, cell-state and perturbation evidence; withhold entire clades.
Pass condition
An unpublished mechanism survives phylogenetic controls, improves held-out-clade prediction ≥10% over body-mass/ancestry baselines, and has independent functional support.
Stop or narrow the claim
Stop if no genuinely independent transitions or clades remain, or if known-target recovery is mistaken for novelty.
Fully costed next test$31,524
$20,491–$50,438 sensitivity range · 6–8 weeks. Tools/data cash allowance: $1,284; contributed labor still has economic cost.
P(this test passes)20%
8–35% judgment range, low confidence. The strategy fits DNALYZER, but prior anatomy failures and strong comparative confounding lower the initial prior.
Annual R&D-tool TAM scenario$36.0M
480 assumed buyers × $75,000/year. Comparative longevity target-evidence platform. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
180 scientific hours × $100 + 48 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $850 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 240–960 assumed buyers and $37,500–$150,000 annual price imply $9.0M–$144.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
20Evidence infrastructure / Priority after benchmarkAging measurements that predict outcomesHuman associations; intervention-response models
What the evidence supports
2026 tAge work integrates over 11,000 transcriptomes across four mammals and predicts multiple aging and mortality-related outcomes. It is substantial prior art.
Human translation
A biomarker associated with mortality is not automatically a validated surrogate for treatment-induced survival benefit. CALERIE also illustrates clock disagreement.
The difficult part
Training contamination, cell mixture, regression to the mean and optimizing the measured clock can manufacture apparent rejuvenation.
Cross-species transcriptomic aging and mortality models ↗ · CALERIE randomized-trial methylation analysis ↗ · Network-based aging drug prioritization ↗
The next public-data test
Benchmark fixed published models on study- and intervention-held-out data, scoring functional outcomes rather than chronological-age accuracy alone.
Pass condition
≥10% lower held-out functional/mortality prediction error than the strongest licensed applicable published model, calibrated intervals, and documented training-data overlap audit.
Stop or narrow the claim
Stop a new-model claim if it cannot beat existing models or if licensed weights/data cannot be used in the intended setting.
Fully costed next test$28,464
$18,502–$45,542 sensitivity range · 5–7 weeks. Tools/data cash allowance: $1,344; contributed labor still has economic cost.
P(this test passes)25%
10–40% judgment range, low confidence. Open tools reduce implementation cost; current strong baselines reduce the odds of meaningful incremental gain.
Annual R&D-tool TAM scenario$25.6M
640 assumed buyers × $40,000/year. Independent biomarker validation and trial analytics. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
160 scientific hours × $100 + 44 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $900 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 320–1280 assumed buyers and $20,000–$80,000 annual price imply $6.4M–$102.4M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
21Integrate interventions / ExploratoryCombinations and sequential repairMouse lifespan combination study
What the evidence supports
A 2025 study found combined rapamycin and trametinib extended mouse survival more than either alone. Some control/single-treatment survival curves had been published earlier.
Human translation
The study establishes neither human combination safety nor indefinite lifespan extension.
The difficult part
Overlapping mechanisms, toxicity, antagonism and shared controls invalidate simple addition of individual benefits.
Rapamycin plus trametinib in mice ↗ · ITP positive and negative publications ↗
The next public-data test
Audit public combination studies against matched single-agent arms; test interaction prediction with entire combinations withheld.
Pass condition
For at least 10 independent factorial experiment units, predict held-out interaction direction better than an additive baseline with a positive study-level interval.
Stop or narrow the claim
Stop if compatible factorial data are too sparse or reused controls are counted as independent evidence. Do not add claimed years of life.
Fully costed next test$25,764
$16,747–$41,222 sensitivity range · 5–7 weeks. Tools/data cash allowance: $1,044; contributed labor still has economic cost.
P(this test passes)15%
5–30% judgment range, low confidence. Combination benefits provide a foothold, but comparable factorial data are scarce and interaction prediction is difficult.
Annual R&D-tool TAM scenario$19.2M
320 assumed buyers × $60,000/year. Combination experiment design and evidence analysis. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
140 scientific hours × $100 + 44 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $650 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 160–640 assumed buyers and $30,000–$120,000 annual price imply $4.8M–$76.8M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
22Human outcome benchmark / ExploratoryProven function-preserving comparatorsRandomized human functional outcomes
What the evidence supports
The LIFE randomized trial found that structured physical activity reduced major mobility disability in sedentary older adults at risk.
Human translation
This is direct human functional evidence. It does not demonstrate reversal of every aging process or unlimited lifespan.
The difficult part
Experimental aging programs must show incremental value over useful existing care and function-preserving approaches.
LIFE: randomized physical-activity and mobility trial ↗ · Empagliflozin: cardiovascular outcomes in diabetes ↗ · Lecanemab in early Alzheimer disease ↗
The next public-data test
Create an endpoint-aligned benchmark comparing experimental claims with randomized human functional-outcome evidence.
Pass condition
≥95% source/endpoint precision in 100 blinded assertions and zero substitution of biomarker change for clinical benefit in the final adjudicated benchmark.
Stop or narrow the claim
Stop comparisons across incompatible populations or endpoints; avoid treating an aging-score shift as equivalent to preserved mobility.
Fully costed next test$12,624
$8,206–$20,198 sensitivity range · 2–4 weeks. Tools/data cash allowance: $384; contributed labor still has economic cost.
P(this test passes)50%
30–70% judgment range, low confidence. This is a bounded evidence-quality task with established sources, not a forecast that a new treatment works.
Annual R&D-tool TAM scenario$4.0M
160 assumed buyers × $25,000/year. Clinical endpoint comparators for research teams. Buyers and prices are unvalidated.
Inspect budget arithmetic and market sensitivity
60 scientific hours × $100 + 28 review hours × $150 + 20M input tokens × $3/M + 4M output tokens × $15/M + $100 compute/storage + $100 other data allowance, then 20% contingency. Token prices are hypothetical defaults, not a named provider quote. Historical spend is not metered.
Market sensitivity: 80–320 assumed buyers and $12,500–$50,000 annual price imply $1.0M–$16.0M annual TAM. This is a scenario, not a confidence interval. All paths overlap in customers; their TAMs must not be added. Commercial probability and follow-on biological-development costs remain unestimated.
No paths match. Clear the search or choose all roles.
EXECUTED / PUBLIC-DATA AUDIT
Rows are not
independent evidence.
DrugAge bulk release retrieved September 15, 2026. A frozen retrospective plan, followed by independent Python and SQL aggregation.
3,423experimental rows1,046exact compound names
Those rows reduce to 1,592 compound × species × publication units. Even these units can share controls or cohorts.
Source: HAGR DrugAge bulk release ↗ · Barardo et al., 2017 ↗
Within the 134 names with mouse evidence
“Both sexes” means recorded male and female observations somewhere in the evidence, not matched conditions or replicated benefit.
What this tells us
Only 20 of the 1,046 recorded names meet the combined mouse-publication and sex-coverage filter. This is an evidence-coverage result, not a 1.9% success rate. DrugAge selects longevity-positive literature, and publication counts do not establish independent laboratories or successful replication.
Our next opportunity is to test whether study identity, context and negative evidence improve prediction beyond current baselines. This audit has not discovered a new anti-aging intervention.
Methods, checks and limitations
Exact compound names were trimmed, without chemical synonym resolution. Species capitalization variants were checked separately: 38 exact labels, 35 after case folding. Mouse matching ignores capitalization. There are 664 PubMed identifiers, seven duplicate full rows and no missing reference identifiers. Twenty Python/SQL aggregate checks agree.
207 names occur in more than one species. That is a separate coverage filter, not an earlier stage of the mouse sequence above. Among 373 mouse rows, start age is missing in two and treatment duration in three. These fields are far more often absent in the database overall.
A secondary description found nine of the 23 multi-publication mouse names had positive within-publication medians in every reference; 14 had at least one nonpositive median. These are mixtures of conditions, not pooled treatment effects, significance tests or a ranking of interventions.
The ZIP’s accompanying release notes report 3,406 entries and 1,045 drugs; the actual CSV has 3,423 and 1,046. Counts here use the CSV. Internal consistency checks do not verify source experiments. No causal, human-efficacy or new-method claim follows from this analysis.
Public summaries and aggregate JSON support scrutiny but are insufficient for full computational reproduction. Implementation, raw snapshots, detailed protocols and candidate-level analysis remain private.
ZIP SHA-256: d12f59717de60a207748f53bdbcdb484ed07e30d8608eb8e4a101e2a08d7fa80
CSV SHA-256: 7ed9771440fa4e1e30be0d3c8e92d919254b572ab40c81e2440ba78c885401d4
RESEARCH ECONOMICS
Price the test.
Challenge the assumptions.
The annual market below is an illustrative research-tool opportunity. It is not the value of curing aging, expected revenue, investor return or a probability-weighted TAM.
Enable JavaScript to calculate edited assumptions. All default budgets appear in the path details above.
Buyer denominator: 3,200 US biotechnology R&D companies in the cited NSF/Census 2023 table, excluding companies below ten employees; the source has an imputation flag. Qualification and annual contract prices are assumptions. Source table ↗ · Full economics methodology ↗
The cost of a cure remains unscoped.
These estimates fund analysis of existing data. They exclude subsequent laboratory, animal, clinical, manufacturing and regulatory work. A positive result earns a separate experiment design, power calculation and supplier quote. Commercial probability is not estimated for these aging paths; there are no validated buyers or investment-return claims.
THE NEXT 90 DAYS / PROPOSED
Earn the next experiment.
Run the transfer benchmark.
Freeze outcomes, splits, baselines and forecasts. Test whether provenance and biological context improve calibration on unseen studies. Publish failures as well as gains.
Choose one functional domain.
Advance immune-response modeling or reprogramming with cell-identity constraints only when usable functional labels and independent datasets justify it.
What would count as a breakthrough?
A useful computational result must beat strong published baselines on unseen studies. A biological discovery additionally needs a previously unestablished mechanism or intervention effect, independent functional replication, and relevant safety evidence. A human rejuvenation claim requires durable human functional or clinical outcomes. Each is a separate gate.
DNALYZER’s potential contribution is linking genome, cell state, intervention and function with uncertainty that survives independent testing. The current evidence does not establish that contribution yet.
Build on existing work
We used HAGR’s bulk dataset and the existing DNALYZER economics model. Biolearn, pyaging, lifelines and Scanpy are inspected foundations for later biomarker, survival and single-cell analyses; they were not run in the coverage audit.
Software licenses were checked; model weights and datasets can have additional restrictions. The 2026 tAge and network/pAGE studies are baselines and prior art, not DNALYZER discoveries.
The team this needs
Computational biology, reproducible scientific engineering, biostatistics and geroscience expertise. Independent reviewers should be rewarded for finding disconfirming evidence. Multiple AIs repeating the same source are not independent validation.
This is a proposed program. No new lab work, paid compute project, customer contract or funding commitment has been initiated by this page.
Coverage gaps and longer-horizon ideas
This map is a targeted review of major paths, not every trial, compound, company or patent. Comprehensive negative-study retrieval, independent expert extraction, formal meta-analysis, IP review, supplier quotes and buyer validation remain open.
Circadian interventions, extracellular vesicles, direct protein repair and specialized delivery approaches remain research-intake topics. Whole-body cryopreservation/revival, autonomous molecular repair and mind emulation have no scoped test or defensible success estimate in this program. Rat kidney preservation is a real, much narrower precedent.
The five selected human trial registry records were checked directly. ER-100’s recruiting status is a registry statement, not proof that any site is currently accepting a participant; its record was last updated May 19, 2026. Published trial results and posted registry results are different records.
SOURCE REGISTER
Inspect the evidence.
42 linked sources, including primary studies, trial registries, public datasets and research economics references. Selection emphasized functional or survival outcomes, null primary outcomes, human translation and 2025–2026 developments. Sources checked September 15, 2026.
- Twelve hallmarks: a research taxonomy ↗Review / taxonomy · 2023
- DrugAge bulk release and documentation ↗Public dataset · 2024
- DrugAge database publication ↗Data resource paper · 2017
- HAGR attribution and reuse terms ↗Provider terms · 2026
- NIA Interventions Testing Program ↗Research program / data · 2026
- ITP positive and negative publications ↗Primary publication index · 2026
- PEARL: human rapamycin trial results ↗Primary study · 2025
- PEARL registered outcomes ↗Trial registry · 2024
- Reprogramming and restored vision in mice ↗Primary study · 2020
- ER-100 phase 1 optic-neuropathy trial ↗Trial registry · 2026
- Senolytics: randomized bone-metabolism trial ↗Primary study · 2024
- Senolytics: function and survival in old mice ↗Primary study · 2018
- Senolytic CAR T cells in mice ↗Primary study · 2024
- Dietary restriction in genetically diverse female mice ↗Primary study · 2024
- CALERIE randomized-trial methylation analysis ↗Primary study · 2023
- Canagliflozin: sex-dependent mouse survival ↗Primary study · 2020
- Empagliflozin: cardiovascular outcomes in diabetes ↗Primary study · 2015
- Beclin-1 autophagy and mouse longevity ↗Primary study · 2018
- Urolithin A: randomized muscle-function trial ↗Primary study · 2022
- Urolithin A: randomized immune-biomarker trial ↗Primary study · 2025
- NMN: randomized function and NAD trial ↗Primary study · 2024
- Somatic mutations across mammalian lifespans ↗Primary study · 2022
- Telomerase gene therapy in adult and old mice ↗Primary study · 2012
- Depletion of biased stem cells in aged mice ↗Primary study · 2024
- IL-11 inhibition and mouse survival ↗Primary study · 2024
- Transient restoration of immune cues in aged mice ↗Primary study · 2025
- Thymic imaging and human health outcomes ↗Primary study · 2026
- Naked mole-rat Has2 transferred to mice ↗Primary study · 2023
- Youthful circulation and mouse lifespan ↗Primary study · 2023
- Parabiosis at single-cell resolution ↗Primary study · 2022
- Young microbiota and cognition in aged mice ↗Primary study · 2021
- Porcine kidney transplantation in a living recipient ↗Primary study · 2025
- Rat kidney cryopreservation and transplantation ↗Primary study · 2023
- Young CSF and memory in aged mice ↗Primary study · 2022
- Lecanemab in early Alzheimer disease ↗Primary study · 2023
- Cross-species transcriptomic aging and mortality models ↗Primary study · 2026
- Rapamycin plus trametinib in mice ↗Primary study · 2025
- LIFE: randomized physical-activity and mobility trial ↗Primary study · 2014
- Network-based aging drug prioritization ↗Primary study · 2026
- Prior DrugAge sex-specific mouse analysis ↗Primary study · 2025
- NSF biotechnology R&D company reference pool ↗Market denominator · 2025
- BLS medical scientist wage context ↗Labor-cost context · 2026
DrugAge attribution: Human Ageing Genomic Resources; Barardo et al., Aging Cell (2017), doi:10.1111/acel.12585. Aggregate analysis by DNALYZER. HAGR content is available under its attribution terms. This page is an early research note, not peer-reviewed research.