Biological Age Tests: What Epigenetic Clocks Can and Cannot Tell You

• An epigenetic clock is an algorithm that estimates biological age from DNA methylation levels at selected sites in the genome.
• Biological age describes the condition of cells, tissues and organ systems, while chronological age only counts the years since birth.
• Second- and third-generation clocks predict age-related diseases and mortality better than first-generation clocks, which were trained to match chronological age.
• Meals, acute stress, sample type and laboratory method can all shift a result, so a single test says little about one person's health status.
• In the CALERIE randomized trial, two years of calorie restriction slowed the measured pace of aging by 2 to 3%, a real but modest effect.
• Aegis Capital, a HealthTech & Longevity VC, backs early-stage startups in Central and Eastern Europe that build tools for earlier detection of age-related diseases.

What Is an Epigenetic Clock?

An epigenetic clock is a statistical model that converts DNA methylation levels at hundreds of genomic sites into an estimate of age, known as DNA methylation age or DNAm age. Methylation attaches small chemical groups to DNA without changing the genetic code, and it helps control which genes a cell switches on or off. Some of these marks change so consistently across a lifetime that a model trained on them works as an aging clock. The difference between the predicted age and chronological age is called epigenetic age acceleration.

A positive difference points to increased age acceleration, and a negative one points to slower aging than the calendar alone would suggest. Researchers treat these tools as biological clocks because they track change inside tissues, not the passage of time. Studies link health outcomes to that difference, not to the predicted age on its own, so the gap is the first figure worth reading in any report.

How DNA Methylation Patterns Record Cellular Aging

Every tissue carries DNA methylation patterns that drift with age, with some sites gaining methyl groups and others losing them. Steve Horvath, who published the first multi-tissue clock in Genome Biology in 2013, proposed that DNA methylation age captures the cumulative effect of an epigenetic maintenance system, the cellular machinery that keeps the epigenome in order. As that system accumulates errors over the decades, the methylation profile of an aging cell moves further from its early-life state.

Methylation marks respond to lifestyle factors as well as to time, and they can be read from routine blood samples. These properties make them a practical record of cellular health that researchers can compare across thousands of people. A clock reads a molecular footprint of aging and does not measure cellular damage directly, a distinction that shapes how any result should be interpreted.

Horvath's Multi-Tissue Epigenetic Clock

Horvath built his multi-tissue epigenetic clock from about 8,000 samples covering 51 healthy tissues and cell types, drawn from 82 DNA methylation datasets. DNAm age in this model is a weighted average of methylation at 353 clock sites, with the weights learned from the training data sets. In independent test data, the paper reported an age correlation of 0.96 and a median error of 3.6 years, meaning predicted age fell within 3.6 years of chronological age for half of the samples. The clock works across most somatic tissues and yields a highly heritable measure of age acceleration, which made it the reference point for a decade of aging research.

Did you know: Horvath's clock returns an age close to zero for embryonic and induced pluripotent stem cells, and the 2013 paper found it applicable to chimpanzee tissues as well. That performance across a broad spectrum of human tissues and a closely related species indicated that the clock reads a shared biological process, not a quirk of one dataset.

Biological Age vs. Chronological Age

Chronological age counts the years since birth and rises at the same rate for everyone. Biological age, sometimes described as functional age, estimates how well tissues and organ systems work compared with population norms for each age. Three terms appear side by side in most test reports:

• Chronological age is the number of years lived, and no intervention can change it.
• Biological age is an estimate of physiological condition built from molecular data or routine clinical biomarkers.
• Age acceleration is the difference between the two, and it is the figure that studies link to disease status and mortality.

Why People of the Same Chronological Age Differ

The Dunedin Study in New Zealand showed how wide the gap can be. In a 2015 paper in PNAS, researchers assessed 954 study members who were all 38 years old and found biological ages ranging from under 30 to nearly 60. Members who were aging faster were less physically able, showed signs of cognitive decline and brain aging, and looked older to other people.

The same team estimated that roughly half of the difference in biological age seen at 38 had built up over the previous 12 years. Divergence in aging therefore starts in early adulthood, long before most people think about chronic diseases or age-related diseases.

Why Biological Age Matters for Healthy Lifespan

Epigenetic age is linked to survival: a 2015 study in Genome Biology found that DNA methylation age of blood predicts all-cause mortality in later life. A 2025 comparison of 14 clocks in 18,859 people, published in Nature Communications, found the strongest mortality association for GrimAge version 2, with a hazard ratio of 1.54 per standard deviation of age acceleration. Adding that clock to a model with standard risk factors raised the area under the curve from 0.851 to 0.865, a measurable but small gain.

The pattern also holds at the far end of the lifespan. A 2015 study in Aging analyzed peripheral blood mononuclear cells from Italian semi-supercentenarians, people aged 105 to 109, and found them 8.6 years younger than expected based on chronological age, while their offspring were 5.1 years younger than age-matched controls. A 2019 analysis by El Khoury and colleagues showed that the Horvath and Hannum clocks systematically underestimate age in older people, so figures at extreme ages call for cautious reading.

How Do Tests Measure Biological Age?

Commercial and research tests rely on three main types of age predictors:

• DNA methylation clocks read methylation at hundreds of genomic sites from blood or cheek swabs and remain the most thoroughly studied approach.
• Blood-chemistry calculators combine routine laboratory results with chronological age into a single score.
• Glycan and protein panels measure sugar structures on antibodies or large sets of plasma proteins linked to aging.

Each method captures a different layer of the body, and results from different methods do not always agree. According to the Stanford Report from September 2026, epigenetic clocks, blood protein panels and wearable metrics often produce conflicting results because each one measures a different aspect of aging.

Three Generations of DNA Methylation Clocks

The authors of the 2025 Nature Communications comparison group DNA methylation age measures into three generations, each trained on a different target.

Generation

Example clocks

Trained to predict

Best suited for

First

Horvath, Hannum

Chronological age

Age estimation and research on the mechanisms of cellular aging

Second

PhenoAge, GrimAge

Mortality risk and markers of general health

Assessing disease and mortality risk

Third

DunedinPACE

Pace of decline across multiple organ systems in people of the same chronological age

Tracking change over time, including responses to interventions

In the same study, second- and third-generation clocks significantly outperformed first-generation clocks in predicting disease, and the authors noted that first-generation clocks may remain more suitable for studying the mechanisms of cellular aging.

DunedinPACE uses 173 sites selected for high test-retest reliability and works from a single blood sample, according to a 2026 review in Biogerontology. The same review reports an intraclass correlation coefficient of 0.96 for the clock. For questions about health risk, second- and third-generation results are the relevant part of a report.

Blood-Chemistry and Glycan Age Predictors

Phenotypic Age, developed by Morgan Levine and colleagues in 2018, combines chronological age with nine standard blood markers: albumin, creatinine, glucose, C-reactive protein, lymphocyte percentage, mean cell volume, red cell distribution width, alkaline phosphatase and white blood cell count. The markers were selected for their ability to predict mortality, and together they reflect inflammation, metabolic function, kidney function and blood cell health. Free online calculators apply the published formula to results from an ordinary laboratory panel to predict biological age at low cost.

Glycan panels follow a different logic. A 2014 study in The Journals of Gerontology analyzed immunoglobulin G glycosylation in 5,117 people from four European populations and found that three glycans explained up to 58% of the variance in age. The remaining variation in those glycans correlated strongly with physiological markers of biological age, and the age-related shifts promote inflammation. Glycan-based age estimation remains less precise than methylation clocks, with an error of 9.7 years for chronological age in that study.

Blood Samples vs. Buccal Cells

The tissue sampled changes what a test can show. Steve Horvath noted at the Precision Medicine World Conference 2026 that the evidence linking clocks to disease and mortality comes from blood, and that without such evidence it is impossible to confirm whether a saliva-based clock predicts anything meaningful. A 2016 study in PNAS found a significant correlation between earlier menopause and increased epigenetic age acceleration in blood, but age at menopause showed no relationship with epigenetic age in buccal cells or saliva.

The same study found a lower epigenetic age in buccal cells of women who used menopausal hormone therapy, and genetic variants tied to age at menopause were also associated with epigenetic age acceleration. One person can therefore receive different readings from different tissues, and the sample type belongs in any comparison between two tests.

Organ Age and Brain Age

Organs within one body do not age in step, and a single systemic aging score can hide large differences between them. Research using plasma proteins that originate from specific organs now estimates organ health separately for the heart, brain, kidneys, liver and other tissues.

What Organ-Specific Clocks Reveal

A 2023 study in Nature estimated the age of 11 organs in 5,676 adults across five independent cohorts. Nearly 20% of participants showed strongly accelerated aging in one organ, and 1.7% were multi-organ agers. People with accelerated heart aging had a 250% higher risk of heart failure, and accelerated organ aging overall carried a 20 to 50% higher mortality risk.

Brain Age and Alzheimer's Disease

Brain age has drawn particular attention. A 2025 study in Nature Medicine estimated organ ages for 44,498 UK Biobank participants and found that an especially aged brain carried a hazard ratio of 3.1 for Alzheimer's disease, similar to carrying one copy of APOE4, while a youthful brain carried a hazard ratio of 0.26. Organ-level estimates move biological age testing closer to specific clinical questions about the brain and nervous system.

Example: Inoko Vision, a company backed by Aegis Capital, is developing NeuroFET, a non-invasive optical device that tracks eye movements to assess the neurological state of the brain. The technology uses eye movement as an objective biomarker for early screening and monitoring of neurodegenerative diseases such as Alzheimer's and Parkinson's.

What Can Biological Age Tests Reveal?

Used across groups of people or tracked over time, biological age tests can show three kinds of information:

• Links with risk factors: a multi-cohort study in eBioMedicine from December 2025 found that smoking, higher BMI, elevated glucose and poor blood pressure, features that overlap with metabolic syndrome, accelerate aging measured by DunedinPACE, while physical activity and a healthier diet slow it.
• Links with disease: in the Nature Communications comparison, the strongest associations involved respiratory and liver conditions such as primary lung cancer and cirrhosis, and DunedinPACE was linked to diabetes with a hazard ratio of 1.44.
• Response to lifestyle interventions: pace-of-aging clocks can register changes within a two-year trial, as the CALERIE study shows.

What the CALERIE Trial Showed

CALERIE randomized 220 adults without obesity to a calorie restriction target of 25% or to their usual diet for two years, and a 2023 analysis in Nature Aging measured the effect on DNA methylation clocks. Participants reduced their intake by about 12% on average, according to the US National Institute on Aging. The intervention slowed DunedinPACE by 2 to 3%, while PhenoAge and GrimAge showed no significant change.

The authors note that, in an independent study of older adults, a 3% slower DunedinPACE is associated with a 15% lower risk of death. The effect is modest and far smaller than the multi-year reversals often promised in consumer marketing. Pace-of-aging measures appear more responsive to interventions than clocks that estimate accumulated age, which makes them the better choice for tracking change over time.

Limitations of Epigenetic Clocks

Researchers who use epigenetic clocks in their own work wrote in The Conversation in July 2026 that the tools are highly effective for studying aging across populations but were not designed to make claims about the health of individuals. Consumer tests based on these clocks cost from about US$30 to more than US$1,000. A Yale University study published in Aging Cell in 2026 found that most clocks reproduce results well when the same sample is measured twice, but reliability across repeated collections from the same person was substantially lower, as meals, stress and environmental exposures shifted results within short intervals.

Laboratory method matters as well. Horvath pointed out in 2026 that methylation arrays and sequencing correlate but measure methylation in fundamentally different ways, and the lack of standardized methods across the industry means that results for the same sample can vary significantly. The Yale team also cites earlier work in which first- and second-generation clocks produced deviations of nearly a decade between technical replicates.

Can Epigenetic Test Results Diagnose Disease?

No DNA methylation clock is a diagnostic test for a specific disease. The authors of the Nature Communications comparison state that epigenetic clocks do not provide insight into disease mechanisms, and adding a clock to classic risk factors improved classification accuracy by more than 1% in only 32 of 176 significant clock-disease associations. A 2025 commentary in the AMA Journal of Ethics describes such tests as potentially meaningful to patients but not clinically actionable, and warns that misread results can cause psychological harm.

A result showing increased age acceleration justifies a review of known risk factors with a physician. It does not establish that any disease is present.

What Cancer Tissue Reveals About the Clock

Cancer tissue shows why clock readings need context. Horvath's original 2013 paper reported that most cancer types appeared strongly age-accelerated, but a 2015 erratum corrected a coding error and retracted that claim: only 6 of 20 cancer types or affected tissues showed positive age acceleration, and others appeared younger than expected. The corrected analysis kept a significant relationship between mutation burden and the clock, with tumors carrying a high number of somatic mutations tending to show low age acceleration.

In breast cancer, estrogen receptor-positive tumors showed increased age acceleration in four independent datasets and progesterone receptor-positive tumors showed the same pattern, while HER2 amplification showed no association. Combinations of various genomic aberrations shifted the estimate further, so tumor biology can push DNAm age in either direction. A biological age test is not a cancer screening tool.

How to Choose a Biological Age Test

Tests differ in the clock they report, the sample they use and the way results are presented. Three questions help compare them:

• Which clocks does the report include? Second- and third-generation clocks carry more information about health risk than first-generation clocks.
• Which sample does the test use? Clinical validation of methylation clocks comes mainly from blood.
• How does the laboratory handle repeat testing? Changes between two results only carry meaning when both tests use the same method.

Free online calculators based on blood chemistry are a low-cost entry point because they use results from a standard laboratory panel. Methylation kits cost more and add a different layer of information, and neither option replaces a medical examination.

Tip: Retest with the same laboratory, the same sample type and the same clock, ideally fasting and at a similar time of day, and judge the trend across several months instead of a single number.

Why Longevity Diagnostics Need Early-Stage Capital

Moving a biological age measure from a research cohort to a clinical tool takes years of validation, standardized laboratory methods and regulatory work before the first sale. The reliability gaps described above call for specialized funding: new clocks need repeated sampling studies, organ-level models need independent cohorts, and diagnostic claims need clinical evidence. Emerging research on organ-specific and pace-of-aging measures is advancing faster than the tools that bring it into clinics.

Aegis Capital is a HealthTech & Longevity VC that invests in early-stage startups from Central and Eastern Europe, and diagnostics focused on automation and early disease detection is one of its core investment areas. The fund has a capitalization of PLN 80 million and offers an initial ticket of up to PLN 3 million, with total funding of up to PLN 8 million per company across follow-on rounds.

FAQ

Is age just a number?

Chronological age is just a number, but biological age does not move in lockstep with it. The Dunedin Study found that people born in the same year differed by about three decades in biological age by the time they turned 38.

Is there an optimal biological age?

No clinical standard defines an optimal biological age. Most tests compare a result with chronological age, and a value below it suggests slower aging relative to the reference population, not a validated health target.

How accurate is a biological age test?

For age prediction, Horvath's multi-tissue clock reached a median error of 3.6 years in test data. Accuracy for health risk depends on the clock generation and the sample type, and repeated samples from the same person can shift with meals or stress.

Can lifestyle interventions lower biological age?

Evidence from randomized trials is limited. In the CALERIE trial, two years of calorie restriction slowed the pace of aging measured by DunedinPACE by 2 to 3%, without significant changes in PhenoAge or GrimAge.

Does menopause affect epigenetic age?

A 2016 study in PNAS found that earlier menopause is associated with increased epigenetic age acceleration in blood. The same study found no such link in saliva or buccal cells, which shows how much the sampled tissue matters.

Can a biological age test predict Alzheimer's disease?

A biological age test does not diagnose Alzheimer's disease in an individual. Research using plasma proteomics, published in Nature Medicine in 2025, found that an especially aged brain carried a 3.1-fold higher hazard of Alzheimer's disease at the population level.

‍