Aging Clocks Catalog · Start here
What can an aging clock tell me?
Useful evidence begins with what a clock was trained to estimate—and the decision you need it to inform.
Start with the decision—not the youngest number. A clock is a model applied to a measurement. Its result can describe age resemblance, a health-related phenotype, risk or pace. Similar-looking numbers do not necessarily answer the same question. [M03; M05; SC6]
What are you trying to decide?
Choose an endpoint or research model
Match the outcome to your specimen, assay, population and implementation. Then ask what it adds to age and an appropriate ordinary comparator.
Understand what a test buys
Separate the assay, algorithm, report and advice service. Compare evidence and total burden without treating four unlike offerings as a ranked shortlist.
Read a changed or conflicting result
Check units, specimen, version and uncertainty before calling a difference biological change. A report does not replace assessment of a clinical concern.
See three decisions worked through
Follow a purchase comparison, a real human trial with six computational clocks, and a study of added risk information.
One label, different quantities
An age-trained model learns a relationship with calendar age; it does not measure every component of aging.
Score-years are not years of remaining life.
A reference-calibrated pace score is not a biological birthday.
Not every aging-related measurement is an age in years.
What the evidence supports now
Research use can be valuable without a clinical claim. Selected models have external prognostic evidence; controlled interventions show responsive and nonresponsive measures. A carefully chosen exploratory endpoint can help a study even when it is not a validated surrogate for health benefit. [SC4; M25; M26]
A lower result does not prove rejuvenation. The rentosertib case applied six clocks to measured serum from a human idiopathic pulmonary fibrosis (IPF) trial. Its selected ancillary cohort and shared disease-responsive proteins limit what the pattern can establish. It supports further biomarker investigation—not six independent demonstrations of geroprotection. [SC1; SC2]
More information is not automatically a better decision. A mortality score can rank outcomes better in a study without providing calibrated individual risk or proving that using the test improves care. The worked risk example explains that distinction. [M11]
“No additional test” is a legitimate conclusion. The four-offering documentary audit did not establish that buying a testing-and-advice package improves long-term outcomes relative to a realistic no-test alternative. That is a boundary of the inspected evidence, not proof that such benefit is impossible. See the offering-specific evidence and unanswered questions.
How to use this catalog
Follow the decision path, then inspect the source and its limits beside the claim. Author-reported results, seller statements, implementation documentation and hypothetical arithmetic are labeled separately. The review is selective, not a systematic census; it does not identify one true age or an overall best clock.
For the standards behind a claim, read How the evidence is judged. For existing implementations and why a public calculator is not supplied, read Tools and access.
Sources and reading limits
Source labels distinguish primary research, seller documents and implementation notes. The access descriptions below refer to the evidence review dated above, not a new replication.
M01. Horvath (2013), DNA methylation age of human tissues and cell types Primary abstract and relevant model/tissue passages inspected; not every supplement or later variant.
M03. Levine et al. (2018), An epigenetic biomarker of aging for lifespan and healthspan Clinical selection, units/coefficient table, methylation stage and validation passages inspected; no raw-data reanalysis.
M05. Belsky et al. (2022), DunedinPACE Primary longitudinal target, normalization, technical/cross-platform reliability and relevant validation text inspected; model not executed.
SC6. Kuo et al. (2024), Proteomic aging clock (PAC) Primary cohort, selection, Gompertz target and train/test passages; full coefficients and ancillary-trial artifact not verified.
T01. BioAge R package README and relevant source files inspected, not executed. DESCRIPTION 0.1.0 declares GPL-3; package paper was identified, not independently read.
SC4. Argentieri et al. (2024), Proteomic aging clock predicts mortality and disease risk Primary training, external validation and covariate passages inspected through publisher-provided text delivered on ResearchGate. No model execution.
M25. Waziry et al. (2023), CALERIE DNA methylation analysis Relevant primary methods/results, analysis population and null outcomes inspected through publisher-supplied full text delivered on ResearchGate; not a longevity-outcome trial.
M26. FDA–NIH BEST: Validated Surrogate Endpoint Official definitions and evidentiary discussion; used as a framework, not a regulatory-status verdict for any aging test.
SC1. Zhavoronkov et al. (2026), Proteomic clocks in a phase 2a trial Published 7 September 2026. Primary results, methods, availability, conflicts and supplement descriptions inspected; numerical supplements not retrieved. Developer-led; Insilico Medicine interests disclosed.
SC2. Xu et al. (2025), A generative AI-discovered TNIK inhibitor for IPF: randomized phase 2a trial Primary registration, disposition, endpoints, safety and protocol-access text inspected; protocol, statistical analysis plan and registry history not obtained. Sponsor/developer interests apply.
M11. Deelen et al. (2019), A metabolic profile of all-cause mortality risk Primary results, conventional comparator, FINRISK evaluation and scaling limitation inspected; reported models not rerun.
Next: Explore the three-axis clock map or compare what commercial tests actually measure.