Epigenetic clocks measure biological age by analysing patterns of DNA methylation — chemical tags on genes that change predictably with age and lifestyle. GrimAge is the strongest predictor of mortality risk among existing clocks; DunedinPACE measures the speed of biological aging rather than absolute age. These tools are entering consumer testing but must be interpreted carefully alongside lifestyle context.
Key Takeaways
- DNA methylation patterns across hundreds of genomic sites change in predictable ways across the lifespan, allowing mathematical algorithms to estimate biological age from a blood or saliva sample.1
- The original Horvath clock (2013) uses 353 CpG sites across multiple tissue types and achieves a correlation above 0.9 with chronological age, with a median error of fewer than 5 years.1
- GrimAge — a second-generation clock — integrates DNA methylation proxies for plasma proteins and smoking exposure; it has demonstrated strong independent associations with mortality, frailty, walking speed, and other health outcomes in multiple large cohort studies.2,3
- DunedinPACE measures the rate of biological aging per calendar year rather than estimating a fixed biological age number; a score above 1.0 indicates faster-than-average aging.5
- A two-year randomised controlled trial of caloric restriction found that DunedinPACE slowed significantly in participants who reduced caloric intake, while PhenoAge and GrimAge showed no significant change — suggesting different clocks capture different aspects of the aging process.7
- A small RCT in healthy males found that an 8-week diet, sleep, exercise, and relaxation programme produced a 3.23-year reduction in Horvath biological age compared to controls.8
- Consumer epigenetic tests are available through companies including TruDiagnostic and Elysium Health; results should be interpreted as probabilistic health indicators, not diagnostic scores, and should always be discussed with a qualified healthcare professional.
Chapter 1: What Is DNA Methylation and Why Does It Track Aging?
The human genome contains roughly 3 billion base pairs of DNA. Embedded within that sequence are approximately 28 million sites — known as CpG sites — where a cytosine nucleotide sits directly before a guanine. At many of these positions, a small chemical group called a methyl group can attach to the cytosine, a modification called DNA methylation.
DNA methylation does not alter the genetic code itself. Instead, it acts as a regulatory layer — influencing whether genes are switched on or off without changing the underlying sequence. This is the core subject of epigenetics: changes in gene expression that do not involve alterations to the DNA sequence.
What makes methylation scientifically compelling in the context of aging is that these patterns do not change randomly. Across large populations, specific CpG sites become more or less methylated in consistent, predictable ways as people grow older. This consistency is strong enough that a machine learning algorithm trained on methylation data can estimate a person's age from a blood or saliva sample with meaningful accuracy.1
The distinction between chronological age and biological age is central to understanding why epigenetic clocks attract scientific and consumer interest. Chronological age is simply the number of years elapsed since birth. Biological age reflects the functional state of cells and tissues — and two individuals of identical chronological age can have meaningfully different biological ages depending on genetics, lifestyle, environment, and disease history.
Epigenetic clocks attempt to quantify that biological age. When a person's epigenetic age exceeds their chronological age — a condition referred to as epigenetic age acceleration — research has associated this with higher risks for several age-related outcomes. When the reverse is true, researchers speak of epigenetic age deceleration, which some studies link to more favourable health trajectories.
It is important to note that these are associations, not certainties. An epigenetic age is a probabilistic estimate derived from population-level patterns. It does not determine any individual's future health with precision, and it should not be treated as a clinical diagnostic result.
Chapter 2: The Horvath Clock — The First Generation
In 2013, geneticist Steve Horvath published a landmark study in Genome Biology introducing what became the foundational epigenetic aging tool in the field.1 Analysing methylation data from 8,000 samples across 51 tissue and cell types, Horvath identified 353 CpG sites whose collective methylation patterns correlated with chronological age across all tissues examined.
The clock's accuracy was notable for its time. Applied across diverse tissue types, it achieved a Pearson correlation with chronological age above 0.9 — meaning age explained the majority of the variance in methylation patterns at those 353 sites. The median error was fewer than 5 years, making it considerably more precise than many biological age measures available at the time.1
The clock's pan-tissue design — its ability to function across blood, saliva, brain tissue, breast tissue, and many other sample types — made it an invaluable research instrument. It offered scientists a consistent ruler with which to compare biological age across studies, populations, and tissue sources.
However, the Horvath clock was designed and validated as a research tool, not a consumer health instrument. Its primary function was to estimate chronological age from methylation data; it was not trained to predict health outcomes or mortality. Subsequent research found that while the original Horvath clock correlates with age, its direct association with mortality and disease risk was weaker than that of later-generation clocks.9
The Horvath clock also reflects a mixture of age-related biological processes — some associated with aging-related damage, others simply reflecting developmental or maintenance programmes that change across the lifespan without obvious health consequences. This limits its sensitivity as a health-outcome predictor compared to clocks trained explicitly on mortality or disease endpoints.
Despite these limitations, the Horvath clock remains one of the most widely cited and reproduced tools in aging research, and it set the methodological template for every epigenetic clock that followed. Its development was a critical proof of concept: DNA methylation patterns can be harnessed to track biological time.
Chapter 3: GrimAge — The Mortality Predictor
The second generation of epigenetic clocks was built not merely to estimate age, but to predict health outcomes. GrimAge — developed by Ake Lu, Steve Horvath, and colleagues — represents the most studied of these outcome-trained clocks and is frequently described by researchers as having the strongest associations with mortality risk among existing epigenetic aging measures.
GrimAge's design differs fundamentally from first-generation clocks. Rather than being trained directly on chronological age, it was trained on DNA methylation proxies for seven plasma proteins that change with age, plus a methylation proxy for smoking pack-years. These underlying targets — which include proteins related to tissue remodelling, inflammation, and lung function — represent biological mechanisms more directly relevant to disease and mortality than simple age passage.2
The published evidence for GrimAge's predictive validity is substantial. A 2021 study published in The Journals of Gerontology examined GrimAge alongside seven other epigenetic clocks in a large population sample.2 GrimAge was associated with 8 of 9 clinical outcomes tested — including walking speed, polypharmacy, frailty index, and mortality — and remained a significant predictor in fully adjusted models that controlled for chronological age and other confounders.
A further study published in Geroscience in 2025 used data from the US National Health and Nutrition Examination Survey (NHANES), following 2,105 participants over a median of 17.5 years, during which 998 deaths occurred.3 Among all clocks examined, GrimAge epigenetic age acceleration was the most statistically significant predictor of mortality (p < 0.0001) after adjustment for chronological age, sex, race, and other covariates.
A Finnish twin study published in Clinical Epigenetics provided additional important evidence: GrimAge emerged as a strong mortality predictor even after controlling for genetic influences shared by identical twins.4 The 18-year follow-up design and the twin comparison allowed researchers to partially disentangle genetic contributions from environmental and lifestyle factors.
In practical terms, a GrimAge score higher than a person's chronological age — referred to as a positive epigenetic age acceleration — suggests that the biological processes captured by the clock are advancing faster than expected for that calendar age. The clinical significance of a given numerical gap (for example, 3 years of acceleration versus 8 years) continues to be studied; population-level risk associations exist, but individual-level clinical thresholds have not been established.
GrimAge should not be understood as a death clock in any literal sense. It is a statistical tool trained on mortality outcomes across large populations. It captures a signal about the state of aging-related biological processes — not a predetermined timeline for any individual.
Chapter 4: DunedinPACE — Measuring the Speed of Aging
While GrimAge estimates a biological age number, DunedinPACE takes a different approach: it measures the rate at which a person is aging, expressed as biological change per calendar year. This conceptual distinction is significant for anyone interested in tracking the effects of lifestyle interventions over time.
DunedinPACE was developed from the Dunedin Study, a birth cohort study that has followed 1,037 individuals born in Dunedin, New Zealand between 1972 and 1973 from birth through midlife.5 Researchers tracked 19 organ-system biomarkers — spanning cardiovascular, renal, hepatic, pulmonary, immune, and metabolic function — across multiple assessments at ages 26, 32, 38, and 45. From these longitudinal measurements, they calculated each participant's pace of biological decline over time.
The DunedinPACE algorithm uses 173 CpG methylation sites to produce a single score representing biological change per calendar year. A score of 1.0 indicates that a person's biological systems are aging at the same pace as the population average. A score above 1.0 indicates faster-than-average biological aging; a score below 1.0 indicates slower aging relative to the reference population.5
The predecessor to DunedinPACE — termed DunedinPoAm — was validated against physical function, cognitive decline, facial aging assessed by independent raters, and mortality in a separate publication.6 DunedinPACE refined this methodology and improved its test-retest reliability, making it more suitable for repeated measurement in intervention contexts.
The key published intervention study for DunedinPACE is the CALERIE trial, a randomised controlled trial in which 220 non-obese participants were assigned to two years of 25% caloric restriction or an ad libitum control diet.7 Published in Nature Aging in 2023, the analysis found that caloric restriction slowed DunedinPACE significantly compared to controls — while PhenoAge and GrimAge did not show statistically significant changes. The authors noted that even modest slowing of biological aging pace at a population level could translate to meaningful reductions in age-related burden over time.
This finding illustrates an important practical point: different epigenetic clocks appear to capture partially distinct biological signals. A dietary intervention may shift pace-of-aging measures without producing detectable changes in outcome-optimised clocks like GrimAge, and vice versa. Interpreting any single clock in isolation may therefore present an incomplete picture.
Chapter 5: How to Get Tested and What to Do With Results
Consumer Testing Options
Epigenetic age testing has moved into the consumer space through several providers. TruDiagnostic offers blood-based testing via a healthcare provider or direct-to-consumer kit, reporting both TruAge (a composite biological age) and DunedinPACE. Elysium Health offers the Index test, which reports a biological age estimate. InsideTracker incorporates epigenetic age as an optional add-on to its blood biomarker platform. The methodologies, algorithms, and reference populations used by each provider differ, which means results from one platform may not be directly comparable to results from another.
Most consumer tests use a saliva or dried blood spot sample collected at home and returned by post to a certified laboratory. Turnaround times vary by provider but typically range from two to four weeks. Prices generally fall between €200 and €500 per test depending on the provider and package.
Interpreting Your Results
The most useful frame for interpreting an epigenetic age result is probabilistic rather than diagnostic. A GrimAge score or DunedinPACE value does not determine your health future — it reflects where your biological aging processes appear to be positioned relative to reference populations at a single point in time. Serial measurements taken over months or years, ideally during or after a structured lifestyle change, provide considerably more information than a single test.
Factors known to be associated with faster epigenetic aging in human studies include higher body mass index, smoking, poor sleep, physical inactivity, low dietary quality, and chronic psychological stress.10 Fish intake, moderate alcohol consumption, higher education level, and elevated blood carotenoid concentrations have been associated with slower epigenetic aging in large cohort studies.10
Lifestyle Interventions with Published Human Evidence
The strongest published intervention evidence comes from structured lifestyle programmes. A randomised controlled trial published in Aging in 2021 assigned 43 healthy males aged 50 to 72 to an 8-week programme incorporating dietary guidance (methylation-supportive foods), targeted sleep, exercise, relaxation techniques, and supplemental probiotics and phytonutrients.8 The treatment group measured 3.23 years younger on the Horvath biological age clock compared to controls at the end of the programme (p = 0.018). This was a small trial with a short follow-up, and replication in larger and more diverse populations remains important — but it represents one of the first controlled demonstrations that lifestyle changes can shift epigenetic age measures within weeks.
The CALERIE trial's findings with DunedinPACE, described in the previous chapter, represent the most methodologically rigorous published evidence for a lifestyle intervention shifting an epigenetic aging measure — specifically, caloric restriction slowing pace of aging over two years in a randomised design.7
Realistic Expectations
Epigenetic clocks are research-validated tools being applied in consumer contexts. Their predictive value at the population level is established; their clinical utility for individual decision-making is still developing. Results should be understood as one data point in a broader picture of health, not as a verdict. Any individual considering epigenetic age testing should discuss the results with a qualified healthcare professional, particularly if they have existing health conditions or are taking medication.
The connection between epigenetic health and specific nutrient pathways has generated research interest in compounds studied for their roles in methylation biology — including NMN and NAD+ precursors (for NAD-dependent cellular pathways), and resveratrol (for sirtuin-pathway interactions). However, no consumer supplement currently carries evidence equivalent to the lifestyle intervention trials cited above, and any supplementation decisions should involve professional guidance.
Q&A: Epigenetic Clocks, Testing, and Aging Science
What is an epigenetic clock?
An epigenetic clock is a mathematical algorithm that estimates biological age by measuring DNA methylation — chemical modifications to specific sites in the genome that change predictably with age. The algorithm is trained on large datasets and uses the pattern of methylation across hundreds of sites to calculate an age estimate or an aging rate score.1
How is GrimAge different from the Horvath clock?
The original Horvath clock was trained to estimate chronological age across tissues. GrimAge was trained on plasma protein proxies and smoking pack-years, making it an outcome-optimised clock specifically designed to predict mortality-related biological processes. As a result, GrimAge shows stronger and more consistent associations with mortality and disease outcomes in published studies than first-generation clocks.2,3
What does DunedinPACE actually measure?
DunedinPACE measures the rate of biological aging per calendar year, not a fixed biological age number. A score of 1.0 means aging at average pace. A score of 1.1 means aging approximately 10% faster than average; a score of 0.9 means aging 10% slower. It was developed by tracking 19 organ-system biomarkers in a birth cohort from age 26 to 45, then mapping those changes to DNA methylation patterns.5
Can lifestyle changes actually shift epigenetic age?
Published human studies suggest they can. A randomised controlled trial found that an 8-week structured lifestyle programme produced a statistically significant 3.23-year reduction in Horvath biological age compared to controls in healthy males.8 A separate RCT found that two years of caloric restriction significantly slowed DunedinPACE, the pace-of-aging measure.7 Both trials were relatively small and further replication is needed, but the direction of evidence is encouraging.
Which lifestyle factors are associated with faster epigenetic aging?
In large cohort studies, higher body mass index, smoking, poor dietary quality, low physical activity, and elevated psychological stress have each shown associations with faster epigenetic aging.10 These findings are consistent across multiple research groups and populations, though most are based on observational data rather than controlled interventions.
Is GrimAge the best epigenetic clock available?
GrimAge currently has the most consistent and extensive published evidence for predicting mortality and age-related health outcomes across multiple large, independent studies.2,3,4 Whether it is the "best" depends on the question being asked. For measuring the pace of aging in response to interventions, DunedinPACE may offer greater sensitivity. Different clocks appear to capture partially distinct biological signals.
How many times should I test my epigenetic age?
A single epigenetic age test provides a baseline data point. Serial testing — typically at intervals of 6 to 12 months, ideally aligned with a documented lifestyle change — provides more actionable information by allowing comparison over time. Individual test-to-test variability exists, so interpreting a single result as definitive is not advisable.
Are epigenetic age tests suitable for everyone?
Consumer epigenetic testing is generally accessible to most healthy adults. However, results should always be interpreted alongside other health data and discussed with a qualified healthcare professional, particularly for individuals with existing health conditions, those on medication, or those who may experience anxiety from probabilistic health estimates. These tests are not diagnostic tools and should not be used to guide clinical decisions without professional input.
Frequently Asked Questions
What is an epigenetic clock and how does it work?
An epigenetic clock is an algorithm that estimates biological age from DNA methylation data — chemical modifications at specific sites in the genome that change in predictable patterns with age. A blood or saliva sample is analysed in a laboratory for methylation at hundreds of specific sites, and the resulting pattern is compared to reference data from large populations to generate an age estimate or aging-rate score.1
What is GrimAge and why is it considered the most predictive epigenetic clock?
GrimAge is a second-generation epigenetic clock trained on DNA methylation proxies for plasma proteins and smoking exposure. Unlike first-generation clocks trained to estimate chronological age, GrimAge was optimised to predict mortality-related biological processes. Multiple large studies, including a 17.5-year prospective cohort study (n=2,105) and an 18-year Finnish twin study, have found GrimAge to be the strongest predictor of mortality and adverse health outcomes among existing epigenetic clocks.3,4
What does DunedinPACE measure and how is it different from other clocks?
DunedinPACE measures the pace of biological aging — the rate of biological change per calendar year — rather than estimating a fixed biological age number. A score of 1.0 indicates average aging pace; above 1.0 indicates faster aging; below 1.0 indicates slower aging. It was developed from a longitudinal birth cohort study tracking 19 organ-system biomarkers across nearly two decades, and its test-retest reliability makes it particularly suited to measuring responses to lifestyle interventions over time.5
Can diet and lifestyle changes improve epigenetic age?
Published human research suggests yes, within structured contexts. An 8-week randomised controlled trial in healthy males found a 3.23-year reduction in Horvath biological age versus controls following a programme of dietary change, sleep improvement, exercise, and relaxation practices.8 A two-year RCT found that caloric restriction significantly slowed DunedinPACE relative to controls.7 These are early findings, and replication in larger trials is ongoing.
How do I access an epigenetic age test?
Consumer epigenetic age testing is available from several providers including TruDiagnostic (reporting TruAge and DunedinPACE), Elysium Health (Index), and InsideTracker (as an optional add-on). Tests typically use a blood or saliva sample collected at home. Prices generally range from approximately €200 to €500 depending on the provider and the breadth of the report. Results should be discussed with a healthcare professional.
References
- Horvath S. DNA methylation age of human tissues and cell types. Genome Biology. 2013;14(10):R115. View on PubMed ↗
- McCrory C, Fiorito G, Hernandez B, et al. GrimAge Outperforms Other Epigenetic Clocks in the Prediction of Age-Related Clinical Phenotypes and All-Cause Mortality. The Journals of Gerontology: Series A. 2021;76(5):741–749. View on PubMed ↗
- Mendy A, Mersha TB. Epigenetic Clocks and All-Cause Mortality: A Prospective Cohort Study. Geroscience. 2025. View on PubMed ↗
- Föhr T, Waller K, Viljanen A, et al. GrimAge as a Robust Predictor of Mortality — a Longitudinal Study of Older Finnish Twins. Clinical Epigenetics. 2021;13(1):127. View on PubMed ↗
- Belsky DW, Caspi A, Corcoran DL, et al. DunedinPACE, a DNA Methylation Biomarker of the Pace of Aging. eLife. 2022;11:e73420. View on PubMed ↗
- Belsky DW, Caspi A, Arseneault L, et al. Quantification of the Pace of Biological Aging in Humans through a Blood Test, the DunedinPoAm DNA Methylation Algorithm. eLife. 2020;9:e54870. View on PubMed ↗
- Waziry R, Ryan CP, Corcoran DL, et al. Effect of Long-Term Caloric Restriction on DNA Methylation Measures of Biological Aging in Healthy Adults from the CALERIE Trial. Nature Aging. 2023;3(3):248–257. View on PubMed ↗
- Fitzgerald KN, Hodges R, Hanes D, et al. Potential Reversal of Epigenetic Age Using a Diet and Lifestyle Intervention: A Pilot Randomized Clinical Trial. Aging. 2021;13(7):9419–9432. View on PubMed ↗
- Horvath S, Raj K. DNA Methylation-Based Biomarkers and the Epigenetic Clock Theory of Ageing. Nature Reviews Genetics. 2018;19(6):371–384. View on PubMed ↗
- Quach A, Levine ME, Tanaka T, et al. Epigenetic Clock Analysis of Diet, Exercise, Education, and Lifestyle Factors. Aging. 2017;9(2):419–446. View on PubMed ↗
