Epigenetic clocks explained: Horvath, GrimAge, DunedinPACE

In 2013, Steve Horvath published a paper that changed aging research. Using DNA methylation data from about 8,000 samples across 51 healthy tissues and cell types, he built a multi-tissue algorithm that estimated DNA methylation age with a median error of about 3.6 years. That model became known as the epigenetic clock.

A decade later, the field has produced multiple clock generations, each answering a different question. Horvath-style clocks estimate age from methylation patterns. GrimAge and GrimAge v2 estimate mortality and healthspan risk. DunedinPACE estimates how fast biological aging is moving right now. They all use DNA methylation, but they are not interchangeable.

Understanding these differences is practical, not academic. It determines which test fits your goal, which number deserves attention, and how to interpret a biological age result without overreading it.

If you are deciding whether methylation is the right layer to measure at all, compare an epigenetic clock with blood biomarker age before choosing a specific clock.

Quick answer

Short answer: Epigenetic clocks are algorithms that read DNA methylation patterns to estimate biological age, mortality and healthspan risk, or pace of aging. Horvath is best for cross-tissue age estimation, GrimAge v2 for mortality-risk research, and DunedinPACE for tracking current aging speed and intervention response.

Key facts

• Epigenetic clocks analyze methylation marks at CpG sites; they do not read your DNA sequence.
• Horvath-style first-generation clocks estimate age and age acceleration across tissues.
• GrimAge and GrimAge v2 are later-generation clocks trained on mortality, smoking, and protein-risk signals.
• DunedinPACE is a pace-of-aging score; values above 1.0 indicate faster biological change than one year per calendar year.
• Clock results should be interpreted as risk and trend signals, not as exact personal forecasts.

What you’ll learn:

• How epigenetic clocks work at the molecular level
• The key differences between first, second, and third-generation clocks
• What each clock actually predicts (and does not)
• How to use epigenetic age data to guide longevity decisions

---

What are epigenetic clocks?

Epigenetic clocks are mathematical algorithms that estimate biological age by analyzing DNA methylation — chemical tags (methyl groups) attached to cytosine bases in your DNA. These tags do not change your genetic sequence, but they control which genes are active or silent.

Quick definition: Epigenetic clocks measure biological age through DNA methylation patterns at specific genomic locations (CpG sites). Different clocks predict different outcomes — from chronological age to mortality risk to the pace of aging.

Why DNA methylation changes with age

DNA methylation patterns shift predictably over the lifespan:

• Some sites gain methylation (hypermethylation) — often silencing protective genes
• Other sites lose methylation (hypomethylation) — often activating inflammatory genes
• These changes are driven by cumulative environmental exposures, lifestyle factors, and stochastic cellular processes

By measuring methylation at hundreds of specific sites, algorithms can estimate how “old” your biology is — regardless of your birthday. Because methylation is so central to this process, genetic variants like MTHFR that impair methylation capacity can directly accelerate all epigenetic clocks.

---

Three generations of epigenetic clocks

First generation: age estimators

Purpose: predict chronological age from DNA methylation

Clock	Year	CpG sites	Tissue	Correlation with age
Horvath	2013	353	Multi-tissue	R ≥ 0.96
Hannum	2013	71	Blood	R ≥ 0.96
Zhang	2019	514	Blood/saliva	R ≥ 0.97

Horvath clock — the original and most widely used. Trained on 8,000 samples across 51 tissue types, it works in virtually any human tissue. When your Horvath age exceeds your chronological age, you are “epigenetically older” — associated with increased disease and mortality risk.

Key insight: first-generation clocks are excellent at estimating chronological age, but their ability to predict health outcomes is limited. A person could be epigenetically “on time” while harboring significant disease risk.

Second generation: mortality predictors

Purpose: predict health outcomes and mortality, not just calendar age

Clock	Year	What it predicts	Mortality prediction
PhenoAge (Levine)	2018	Composite of clinical biomarkers plus age	Strong
GrimAge (Lu)	2019	Mortality, smoking pack-years, and plasma-protein surrogates	Strong
GrimAge v2	2022	Updated protein and clinical-risk surrogates	Strong

GrimAge is a mortality-risk and healthspan clock, not a personal death-date predictor. It was trained on DNA-methylation surrogates for plasma proteins linked to disease and mortality plus smoking pack-years; GrimAge v2 adds updated surrogates including hs-CRP and HbA1c.

Deep dive: Read our guide on PhenoAge and KDM biological age calculations for the clinical biomarker approach.

Key insight: Second-generation clocks were designed to capture health-outcome and mortality-risk signal, making them more clinically informative than pure calendar-age estimators.

Third generation: pace of aging

Purpose: measure how fast you are aging right now

Clock	Year	What it measures	Key feature
DunedinPACE	2022	Rate of biological aging	Dynamic, responsive to interventions

DunedinPACE (Pace of Aging Calculated from the Epigenome) represents a paradigm shift. Rather than estimating a static age, it measures your current speed of aging:

• A score of 1.0 means you are aging at a rate of 1 year per calendar year (normal)
• A score of 0.85 means you are aging at 85% the normal rate (slower — good)
• A score of 1.20 means you are aging 20% faster than normal (accelerated — concerning)

DunedinPACE was developed from the Dunedin Study, which tracked 1,037 people born in 1972–73, measuring 19 biomarkers of organ function repeatedly over decades. This longitudinal design captures the trajectory of aging, not just a snapshot.

Key insight: DunedinPACE is the most responsive to lifestyle interventions. Changes in diet, exercise, and stress management can shift your DunedinPACE score within months — making it the best clock for tracking whether your longevity strategy is working.

---

How the clocks compare

How the clocks differ

The clocks are not interchangeable. Large cohort comparisons show that first-generation clocks are strongest for age estimation, while second- and third-generation clocks generally carry more disease-risk and health-outcome signal.

• First-generation clocks cluster around chronological age and age acceleration.
• GrimAge v2 and DunedinPACE produced many of the strongest disease and mortality associations in a 2025 comparison of 14 clocks.
• Health and Retirement Study analyses also show that later-generation clocks are more closely linked to education, income, race/ethnicity, behavior, mortality, and multimorbidity.
• No single clock is universally best; the useful clock depends on whether you are asking about age, risk, or pace.

What each clock is best for

Use case	Best clock	Why
How old am I biologically?	Horvath or PhenoAge	Designed for static age estimation
What is my mortality risk?	GrimAge v2	Strongest mortality predictor
How fast am I aging right now?	DunedinPACE	Measures rate, not state
Is my intervention working?	DunedinPACE	Most responsive to lifestyle changes
Multi-tissue comparison	Horvath	Works across all tissue types
Research on aging mechanisms	Horvath + GrimAge	Most widely validated

Second and third generation clocks and socioeconomic factors

An important finding from the Health and Retirement Study (2024): second and third-generation clocks correlate more strongly with education, income, and race/ethnicity than first-generation clocks. This suggests they are capturing the biological impact of social determinants of health — chronic stress, environmental exposures, nutritional access — in addition to intrinsic aging.

---

What makes epigenetic clocks tick faster?

Factors consistently associated with epigenetic age acceleration across multiple clocks:

Factor	Effect on epigenetic age	Clocks most affected
Smoking	+2 to 7 years	GrimAge especially
Obesity (BMI >30)	+1 to 3 years	PhenoAge, GrimAge
Chronic stress	+0.5 to 5 years	All clocks
Depression	+2 to 5 years	PhenoAge, DunedinPACE
Air pollution	+0.5 to 2 years	Horvath, GrimAge
Alcohol excess	+1 to 4 years	GrimAge
Low socioeconomic status	+1 to 3 years	GrimAge, DunedinPACE

---

What slows epigenetic clocks?

Intervention	Evidence for clock deceleration	Strongest evidence
Exercise (150+ min/week)	1.5–3 years reversal in 6 months	DunedinPACE
Meditation	Dose-dependent reduction in IEAA	Horvath
Caloric restriction	Slows DunedinPACE in CALERIE trial	DunedinPACE
Mediterranean diet	Reduced GrimAge acceleration	GrimAge
Sleep optimization	Improved across multiple clocks	DunedinPACE
Stress reduction	Reduced cortisol-mediated methylation	Horvath, GrimAge
Metformin	TAME trial in progress	TBD

---

How to get your epigenetic age tested

Available testing services

Several commercial services now offer epigenetic clock testing:

• Blood sample collection (typically a small blood draw or dried blood spot)
• Results typically include multiple clock ages and pace of aging
• Costs range from $200–$500 per test
• Repeat testing every 6–12 months to track trends

How to interpret results

1. Compare to chronological age — are you biologically older or younger?
2. Check DunedinPACE — is your pace above or below 1.0?
3. Look at trends — single measurements are noisy; track changes over time
4. Context matters — recent illness, stress, or sleep disruption can temporarily accelerate clocks

---

How SuperAge complements epigenetic clocks

Epigenetic clocks require a blood test and weeks of processing. SuperAge provides the continuous, real-time monitoring that fills the gaps between clock measurements.

Daily biological age tracking

SuperAge’s biological age calculation uses wearable data to estimate your physiological age daily. While not DNA-based, it tracks the same downstream markers (HRV, sleep, fitness, stress) that epigenetic clocks capture in a single snapshot.

Pace of aging equivalent

SuperAge’s pace of aging feature mirrors DunedinPACE’s concept — measuring how fast you are aging in real time, using wearable-derived biomarkers rather than DNA methylation.

Intervention feedback loop

The power of combining epigenetic clocks with SuperAge: test your epigenetic age every 6–12 months to get the definitive molecular measurement, while using SuperAge daily to ensure your lifestyle interventions are on track between tests.

---

Frequently asked questions

Which epigenetic clock should I choose?

For most consumers, a test that includes GrimAge v2 (mortality prediction) and DunedinPACE (pace of aging) provides the most actionable information. GrimAge tells you your long-term risk trajectory; DunedinPACE tells you whether your current lifestyle is accelerating or decelerating aging.

Can you reverse your epigenetic age?

Yes — partially. Exercise, dietary interventions, stress reduction, and sleep optimization have all been shown to reduce epigenetic age acceleration. The CALERIE trial demonstrated that caloric restriction slowed DunedinPACE by 2–3%. However, the degree of reversibility depends on the clock, the intervention, and individual factors.

Are epigenetic clocks accurate for all populations?

Accuracy varies. Most clocks were developed using predominantly European-descent samples. While they work across ancestries for age estimation, their predictive accuracy for health outcomes may differ across populations. Research is actively expanding training datasets to improve equity.

How often should I test?

Every 6–12 months provides a useful trend. More frequent testing adds noise without useful signal. If you are curious how your genetic makeup predicts your likelihood of aging faster, polygenic risk scores offer a complementary view: while epigenetic clocks show your current methylation state, PRS reveals your underlying genetic predisposition to age-related disease. Between tests, track the wearable-derived biomarkers (HRV, sleep, activity, stress) that correlate with epigenetic age changes. For clock-specific intervals — DunedinPACE, GrimAge, and blood-based calculators each move at different speeds — see the biological age retest timing guide.

---

Key takeaways

• Three generations of clocks answer different questions: Horvath estimates age, GrimAge and GrimAge v2 estimate mortality and healthspan risk, and DunedinPACE estimates current aging pace.
• DunedinPACE is useful for intervention tracking: it measures pace rather than a static biological age estimate.
• GrimAge v2 is a mortality-risk clock: it estimates risk, not an individual’s date of death.
• Clocks are not interchangeable: later-generation clocks generally perform better for disease-risk associations than first-generation age estimators.
• Lifestyle still matters: exercise, diet quality, sleep, stress reduction, and calorie restriction can shift some clock measures, especially pace-of-aging scores.

---

Measure your molecular age

Epigenetic clocks have transformed aging from an abstract concept into a measurable, modifiable biological process. Whether you choose to test your DNA methylation age or track your biology daily through wearable data, the message is the same: how you live determines how fast you age.

Ready to track your aging in real time? Download SuperAge and start monitoring the daily biomarkers that predict what your next epigenetic clock test will reveal.

---

References

1. Horvath S. (2013). “DNA methylation age of human tissues and cell types.” Genome Biology. https://doi.org/10.1186/gb-2013-14-10-r115 — Original multi-tissue Horvath epigenetic clock.
2. Lu AT et al. (2019). “DNA methylation GrimAge strongly predicts lifespan and healthspan.” Aging. https://doi.org/10.18632/aging.101684 — GrimAge mortality and healthspan clock.
3. Belsky DW et al. (2022). “DunedinPACE, a DNA methylation biomarker of the pace of aging.” eLife. https://doi.org/10.7554/eLife.73420 — Pace-of-aging methylation measure.
4. Levine ME et al. (2018). “An epigenetic biomarker of aging for lifespan and healthspan.” Aging. https://doi.org/10.18632/aging.101414 — PhenoAge DNAm clock.
5. Lu AT et al. (2022). “DNA methylation GrimAge version 2.” Aging. https://doi.org/10.18632/aging.204434 — Updated GrimAge v2 clock.
6. Crimmins EM et al. (2024). “Generations of epigenetic clocks and their links to socioeconomic status in the Health and Retirement Study.” Epigenomics. https://doi.org/10.1080/17501911.2024.2373682 — Social gradients and clock-generation comparison.
7. Mavrommatis C et al. (2025). “An unbiased comparison of 14 epigenetic clocks in relation to 174 incident disease outcomes.” Nature Communications. https://doi.org/10.1038/s41467-025-66106-y — Large disease-risk comparison.
8. Waziry R et al. (2023). “Effect of long-term caloric restriction on DNA methylation measures of biological aging in healthy adults from the CALERIE trial.” Nature Aging. https://doi.org/10.1038/s43587-022-00357-y — Caloric restriction and DunedinPACE.

---

Last updated: 2026-06-07. This article is regularly reviewed to ensure accuracy.