Organ Proteomic Clocks
Organ Proteomic Clocks estimate how old individual organs look from proteins circulating in blood, but prediction is much better established than clinical actionability.
Also known as: organ-specific proteomic aging clocks, organ-age clocks, plasma proteomic organ-age models
One blood draw can produce separate age estimates for the brain, heart, kidney, liver, immune system, and other organs. That sounds like direct measurement. It isn’t. An organ proteomic clock is a statistical model. It recognizes an age-associated protein pattern in a reference population, then reports how far an individual’s pattern differs from the expected value.
What It Is
An organ proteomic clock estimates an organ’s biological age from plasma proteins enriched in that organ. The model begins with proteins that are expressed much more strongly in one tissue than in others. It then learns how the plasma concentrations of those proteins vary with chronological age.
The output is usually an age gap. After the model accounts for chronological age, a heart-age gap of plus six years means the heart-associated protein pattern resembles that of someone six years older. It doesn’t mean the heart itself has aged exactly six extra years. Nor does it identify which structure or pathway produced the difference.
The 2023 Stanford study that established this approach modeled 11 organs in 5,676 adults across five cohorts. It found that organs within the same person can follow different trajectories. An older-looking kidney can coexist with a younger-looking heart or immune system. Accelerated age in a given organ was associated with later disease involving that organ and with mortality (Oh et al., 2023).
A 2025 multi-cohort study tested a related framework in 43,616 UK Biobank participants, 3,977 China Kadoorie Biobank participants, and 800 participants from the Nurses’ Health Study. It modeled ten organ systems from 418 organ-enriched proteins. Brain age showed the strongest associations with dementia and all-cause mortality in that analysis (Wang et al., 2025).
Why It Matters
Most biological-age reports compress many systems into one number. Organ clocks make the heterogeneity visible. A whole-body estimate can look ordinary even when one organ-associated signal is far from its age expectation. That creates a more specific research question: does the age gap predict the diseases linked to that organ?
Large cohort studies have linked older-looking organ profiles with incident disease and death after accounting for many conventional risk factors. Organ age may therefore add a risk-prediction signal, although the size and usefulness of that addition depend on the organ, model, and population.
Risk prediction isn’t the same as clinical utility. A test has clinical utility when its result changes a decision and improves an outcome. No randomized trial has shown that acting on a proteomic organ-age gap reduces disease incidence, preserves function, or extends healthy life. A result may restate risk already visible through symptoms, family history, imaging, kidney function, lipids, blood pressure, glucose status, cognition, or fitness.
This distinction matters as organ-age reports move from cohort research into commercial testing. A model can predict risk across several cohorts and still lack a defined place in routine care. A purchasable number doesn’t inherit a treatment pathway merely because it has an organ’s name attached.
An organ-age report is a risk-prediction product, not a scan of how old an organ literally is. No trial has shown that changing a protocol because of the reported age gap improves health outcomes. The result can sharpen a question; it can’t yet tell a clinician what intervention will answer it.
How It Is Measured
The current research lineage uses high-multiplex plasma-proteomics platforms, chiefly Olink and SomaScan. These assays measure hundreds or thousands of proteins from a blood sample. Researchers map proteins to tissue-expression data, select the organ-enriched subset, and train a model to predict chronological age. The residual between predicted and chronological age becomes the organ-age gap.
The platforms don’t measure the same proteins in the same way. Olink uses antibody pairs with DNA-based readout. SomaScan uses modified DNA aptamers that bind target proteins. A model trained on one platform isn’t automatically portable to the other. Pre-analytical handling, assay version, population ancestry, chronic disease burden, and model calibration can all affect the estimate.
The 2025 study also tested smaller panels containing roughly 10 to 20 proteins per organ. Those compact models retained about 88 percent of the larger models’ predictive performance. Smaller panels could lower assay costs, but this result doesn’t establish that they are ready to guide care.
An interpretable report needs more than an age gap. It should name the assay platform, model version, reference population, organ definition, technical repeatability, and validated outcome. Without those details, two reports labeled “brain age” may not measure the same construct.
| Claim | Current support | What remains missing |
|---|---|---|
| The model predicts age-associated risk | Replicated large human cohorts | Prospective use in a defined clinical pathway |
| Different organs can show different age gaps | Repeated cross-sectional and longitudinal associations | A gold-standard measure of each organ’s biological age |
| A high organ-age gap can trigger a better intervention | Not established | Trials showing result-guided care improves outcomes |
| Repeat testing can track treatment response | Early and mostly observational | Test-retest standards and validated within-person change thresholds |
How It Plays Out
A 57-year-old receives a report showing an older brain profile and ordinary estimates for the heart, kidney, liver, and immune system. The number is not a dementia diagnosis. It is a prompt to review established context: cognitive symptoms, blood pressure, sleep apnea risk, hearing, medications, vascular risk, family history, and whether standard evaluation is warranted. If none of that changes, the report has added risk information without adding a new decision.
Another reader sees a liver-age gap improve after six months that included weight loss, less alcohol, better sleep, a new exercise plan, and medication changes. The movement can’t identify which change mattered. It may reflect a real shift in circulating proteins, assay variation, regression to the mean, or a combination. Assigning the improvement to the newest supplement would exceed the evidence.
A researcher has a cleaner use. In a prevention trial, organ-age gaps can sit beside clinical outcomes as exploratory endpoints. The model may reveal that an intervention shifts kidney-associated proteins without shifting the heart or brain profile. That finding can guide later hypotheses while harder outcomes determine whether the intervention helped.
A longevity clinic includes organ ages in an annual deep screen. The important feature isn’t the number of organs reported. It is the interpretation plan. A report that names the model, compares the result with established risks, and states when no action follows is more credible than a dashboard that pairs every older-looking organ with a product or procedure.
Evidence
Evidence tier: Observational (human, large). The prediction evidence merits attention. It can’t support a prescription by itself.
Oh and colleagues developed organ-specific models from plasma proteomic data across five cohorts. Accelerated organ age was associated with disease involving the same organ, and extreme aging in one or more organs tracked higher mortality risk. The study showed that the circulating proteome contains separable organ-linked age signals (Oh et al., 2023).
The later multi-population validation expanded the sample and tested generalization across British, Chinese, and US cohorts. In that study, a one-standard-deviation older brain-age estimate was associated with about 1.88 times the risk of dementia and 1.44 times the risk of all-cause mortality. The associations persisted beyond conventional clinical and genetic predictors. They remain associations, not proof that lowering a brain-age estimate lowers either risk (Wang et al., 2025).
The evidence weakens as the claim moves closer to action. Cross-sectional links between organ ages, lifestyle factors, and medication use suggest that the profiles may be responsive. They don’t establish within-person responsiveness, causal direction, or a minimum change that exceeds assay and biological variation. The field still lacks a standardized organ-age unit, a reference assay, and trials of result-guided care.
Caveats and Open Questions
“Organ-specific” is a model property, not perfect anatomical isolation. Proteins enriched in one organ can still circulate from several tissues, and disease in one system can alter proteins associated with another. Kidney function, inflammation, cancer, medications, acute illness, and body composition can change the plasma proteome broadly.
Population calibration also matters. A model developed in one age, ancestry, or disease distribution may miscalibrate another. Cross-cohort validation helps, but it doesn’t remove the need to show performance in the population receiving the test.
Repeat measurement is the largest consumer-facing gap. The model can predict long-term risk even if short-term movement is noisy. Until studies establish test-retest reliability and meaningful within-person thresholds, a quarterly organ-age change shouldn’t be treated as an intervention scorecard.
Consequences
Benefits. Organ Proteomic Clocks give biological age a finer grain. They can identify organ-linked risk signals that a single composite age might hide, help researchers separate systemic from organ-specific change, and support more precise hypotheses about disease development.
Liabilities. The organ label can create false concreteness. A brain-age gap feels more diagnostic than an opaque composite score, even though it remains model output. That invites Single-Biomarker Tunnel Vision, especially when one alarming result outranks established clinical findings.
Repeat panels can also feed Biomarker Treadmill. If every movement triggers a new supplement, scan, or procedure, measurement has outrun its decision rule. For now, the defensible use is narrow: understand the model, ask whether the result changes a standard clinical question, and let established risks and function keep their priority.
Related Articles
Sources
- Oh, Hamilton S., et al. “Organ Aging Signatures in the Plasma Proteome Track Health and Disease.” Nature 624 (2023): 164–172. https://doi.org/10.1038/s41586-023-06802-1
- Wang et al. “A Plasma Proteome Measure of Organ Aging.” Nature Aging (2024). https://doi.org/10.1038/s43587-024-00568-5
- Wang et al. “Organ-Specific Proteomic Aging Clocks Predict Disease and Longevity across Diverse Populations.” Nature Aging (2025). https://doi.org/10.1038/s43587-025-01016-8
Medical and Legal Boundary
This entry is a reference, not medical advice. It describes research-grade risk models and emerging laboratory tests. It does not diagnose disease, prescribe an intervention, or replace a clinician’s judgment for a specific person.
No organ-specific proteomic aging clock is FDA-cleared or approved to diagnose an age-related disease or direct a longevity intervention. An abnormal report should be interpreted with a qualified clinician in the context of symptoms, established risk factors, standard laboratory testing, imaging when indicated, medications, and family history.