- Quick Answer
- What Are Epigenetic Clocks?
- The Horvath Clock (First-Generation)
- PhenoAge (Second-Generation)
- GrimAge (Second-Generation)
- DunedinPACE (Third-Generation)
- Horvath vs. PhenoAge vs. GrimAge vs. DunedinPACE
- What Does Biological Age Actually Mean?
- Can Lifestyle Change Your Epigenetic Age?
- How Reliable Are Epigenetic Clocks?
- Epigenetic Clocks in Longevity Clinics
- Are Epigenetic Clocks Proven to Measure Longevity?
- Which Clock Is Most Useful?
- Bottom Line
- Frequently Asked Questions
If you’ve taken a biological age test, the figure on your report was derived from one of a handful of epigenetic clocks, and which biological age clocks were used matters more than most people realize. Horvath, PhenoAge, and GrimAge are not different flavors of the same test. They are distinct instruments, actually, trained to answer various queries and can provide quite diverse outcomes to the same person.
But before you look at your own figures, here’s the bit to understand: DNA methylation clocks (also called DNAm clocks or epigenetic aging clocks) are categorized into generations depending on what they were originally designed to forecast. Others learned to guess your birthdate. Some were trained to forecast how long you have to live. That distinction makes all the difference in how much weight to give a single score.
This guide walks through each major clock, what makes them different, what the research says they can actually predict, and which one is worth paying attention to depending on what you're trying to find out.
|
|
Quick Answer
Before the full breakdown, here's the fast version of what each major clock was built for and which one to trust for what.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
What Are Epigenetic Clocks?
Before comparing individual clocks, it helps to understand the actual mechanism behind a biological age test; this is real, measurable biology, not a black-box algorithm guessing at your health.
- DNA methylation and CpG sites: Epigenetic clocks estimate biological age by analyzing DNA methylation, a process where methyl groups attach to cytosine-guanine sites (CpG sites) across your genome.
- How methylation becomes an age estimate: Machine learning methods, typically penalized regression, select a specific subset of CpG sites and weight their methylation levels to calculate an age estimate.
- Chronological age vs. epigenetic age: Chronological age is simply time since birth. Epigenetic (biological) age reflects your actual cellular integrity, health status, and functional capacity, which can diverge significantly from your birth certificate.
- What age acceleration means: Epigenetic Age Acceleration (EAA) happens when a clock's biological age estimate comes in higher than your actual chronological age, a signal of faster-than-expected biological aging.
- Why different clocks disagree: Different generations of clocks are trained on different targets. Some aim purely to predict chronological age; others are built to predict mortality risk, healthspan, or your current pace of aging.
The Horvath Clock (First-Generation)
Steve Horvath developed the first pan-tissue epigenetic clock in 2013, and it remains the reference point most other clocks get compared against.
- How it was built: Trained on roughly 8,000 non-cancer samples across 51 different healthy tissues and cell types, an unusually broad foundation for its time.
- What it predicts: A first-generation clock trained primarily to predict chronological age, using methylation patterns at 353 specific CpG sites.
- Its real strength: Genuinely high accuracy, a correlation of 0.96 with actual age, and a median error of just 3.6 years, plus adaptability across blood, brain, and internal organs alike.
- Its real limitation: It's a weak predictor of functional decline or direct disease risk, since it excludes many of the methylation patterns tied to physiological deterioration.
- What it's actually good for: A baseline measure of epigenetic age and a benchmark tool for biological age standards, not a primary tool for assessing clinical health risk.
PhenoAge (Second-Generation)
PhenoAge represents the shift from guessing the birthday clocks to clocks built around actual health outcomes a genuinely different design philosophy.
- What it was built to do: Improve prediction of morbidity and mortality, rather than just calendar age.
- How it works: Trained on a composite of chronological age plus 10 clinical biomarkers tied to immune function, inflammation, and metabolic health, including albumin, creatinine, glucose, and CRP.
- What it predicts: A strong predictor of all-cause mortality, several cancers, physical function decline, and Alzheimer's disease.
- Age acceleration matters here too: Research shows a 1-year increase in DNAm PhenoAge is linked to a 4.5% increase in all-cause mortality risk, a genuinely meaningful jump for a single year of acceleration.
- What it's best interpreted as: A window into phenotypic aging, specifically immune aging and metabolic stress.
GrimAge (Second-Generation)
If you had to pick one clock for predicting how long someone has left, current research points squarely at GrimAge.
- What it was built to do: Specifically optimized to predict lifespan and time-to-death, genuinely its stated design goal, not a side benefit.
- How it works: A two-stage approach using DNAm-based surrogate markers for smoking pack-years, plus seven plasma proteins including PAI-1, growth differentiation factor 15, and leptin.
- What it predicts: Robustly associated with cardiovascular disease, coronary heart disease, cancer, and inflammatory markers.
- The acceleration numbers are stark: Someone in the top 95th percentile for GrimAge acceleration faces roughly twice the mortality hazard of an average person.
- Its real strength: It frequently outperforms both first-generation clocks and PhenoAge at predicting lifespan, healthspan, and clinical phenotypes like walking speed and frailty.
DunedinPACE (Third-Generation)
DunedinPACE takes a genuinely different approach from everything above it instead of a static age, it measures speed.
- What makes it different: Rather than a fixed age estimate, DunedinPACE measures your current pace of aging and how fast your body is deteriorating right now.
- How it was built: Developed using 20 years of longitudinal follow-up data from the Dunedin Study, tracking 19 biomarkers of organ system integrity over time.
- How to read the number: Results are expressed as biological years of aging per calendar year. A value of 1 means a typical pace; anything above 1 signals accelerated biological change.
- Why it's uniquely useful: It's highly sensitive to lifestyle changes and short-term interventions like caloric restriction, which makes it especially well-suited for clinical trials and ongoing health monitoring.
- What higher scores mean: Greater risk of chronic disease, functional impairment (poor balance, slower walking speed), and earlier mortality.
Horvath vs. PhenoAge vs. GrimAge vs. DunedinPACE
With four genuinely different tools on the table, a direct side-by-side makes the distinctions much easier to hold onto.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
What Does Biological Age Actually Mean?
With four clocks measuring genuinely different things, it's worth being precise about what biological age is actually referring to in each case.
- What it represents: Biological age reflects the cumulative functional state of your organ systems and molecular integrity, which may diverge significantly from your actual years lived.
- Static age vs. rate of aging: Most clocks (Horvath, PhenoAge, GrimAge) give a static estimate at one point in time. Third-generation clocks, like DunedinPACE, instead measure the pace of aging, essentially how many years you're aging per calendar year.
- Acceleration vs. absolute age: Age acceleration is the gap between your predicted biological age and your actual chronological age, flagging who's aging faster or slower than expected.
- What this actually means clinically: Accelerated biological age is a meaningful marker for frailty, multi-disease risk, and mortality, though its routine use in everyday clinical settings is still being worked out.
Can Lifestyle Change Your Epigenetic Age?
This is genuinely good news: research shows lifestyle factors can meaningfully shift your epigenetic age, especially on second- and third-generation clocks that are built to be sensitive to real-world change.
- Exercise: Regular physical activity consistently links to lower age acceleration, particularly on GrimAge and DunedinPACE. Highly fit individuals have shown up to 1.5–2 years younger biological age than sedentary peers, worth reading alongside why exercise and diet fundamentals matter so much.
- Nutrition: Plant-based diets rich in leafy greens, folic acid, vitamin B12, and flavanols (natural DNA methyltransferase inhibitors) are linked to reduced epigenetic age.
- Weight management: Obesity is one of the strongest predictors of accelerated aging; each 10-unit rise in BMI is linked to roughly 3.3 additional years of liver epigenetic age, making healthy weight management an important factor if you're looking to lower biological age.
- Smoking: A major driver of acceleration. Current smokers show roughly 7.9 years more GrimAge acceleration than never-smokers, driven by profound smoking-related methylation changes.
- Stress: Chronic psychosocial stress and adverse childhood experiences are robustly linked to accelerated biological aging well into adulthood.
- Sleep: Healthy sleep habits are part of interventions that have set back the Horvath clock by as much as 3.23 years in pilot trials.
- Supplements and medications: Omega-3s (1 g/day) have been shown in clinical trials to slow PhenoAge, GrimAge2, and DunedinPACE specifically. Vitamin D3 and metformin also show geroprotective potential in certain populations.
- Caloric restriction: Shown to slow DunedinPACE by 2–3%, though it may not move static estimates like PhenoAge or GrimAge in the same way.
- The honest caveat: It's still not definitively proven whether 'turning back' a clock reflects a true reversal of aging biology or just a shift in the methylation biomarkers themselves. Different clocks respond differently to the same intervention, which is exactly why picking the right clock for your goal matters.
How Reliable Are Epigenetic Clocks?
Despite being the most successful aging biomarkers developed so far, epigenetic clocks still face real reliability and standardization challenges worth knowing about.
- Reproducibility: Technical reliability varies, though newer versions using Principal Component (PC) analysis have meaningfully improved stability over time.
- Tissue matters: Performance is highly tissue-specific; a clock trained on blood may not accurately reflect what's happening in brain or muscle tissue.
- Lab and algorithm variation: Results can shift based on the DNA extraction method, the specific methylation array used (450k vs. EPIC), and batch effects during processing.
- Population differences: Most clocks were trained on predominantly European or Western populations, which may limit accuracy in other ethnic groups due to genetic variation at methylation sites.
- No gold standard yet: There's still no universally accepted benchmark for biological age. Epigenetic clocks are best understood as a molecular readout of aging that needs to be combined with other clinical and functional data, not read in isolation.
Epigenetic Clocks in Longevity Clinics
Longevity and anti-aging clinics are increasingly building these clocks into personalized care; here's what that actually looks like in practice.
- Where and how: Facilities across the US, Japan, and Singapore use certified providers like TruDiagnostic and Zymo Research for DNA methylation testing. Venous blood is most common, though buccal (cheek) and saliva samples are also used for accessibility.
- Interpretation and action: Results get read alongside standard blood panels, imaging, and functional tests like grip strength and VO2max, similar to the integrated approach covered in executive longevity testing packages. A high biological age result typically triggers real intervention, a personalized diet, targeted supplementation, or intensified exercise.
- Monitoring: Testing is often repeated every 6–12 months, giving clinicians an objective way to measure whether interventions are actually working, rather than relying on how a patient feels.
Are Epigenetic Clocks Proven to Measure Longevity?
This is the honest, important caveat that sits underneath everything: prediction and proof aren't the same thing.
- Mortality and disease: GrimAge and PhenoAge are robust predictors of all-cause mortality and time-to-death and are strongly associated with cardiovascular disease, cancer, and Alzheimer's.
- Healthspan and aging rate: DunedinPACE and GrimAge effectively track healthspan markers mobility, physical function, and cognitive performance.
- The causality gap: Association isn't causation. A higher epigenetic age genuinely correlates with earlier death, but it hasn't been proven in humans that lowering your clock score directly translates into a longer life. That distinction matters more than marketing usually admits.
Which Clock Is Most Useful?
The honest answer is that best depends entirely on what you're actually trying to find out; there isn't one universal winner among these four.
- Horvath: Most useful for broad baseline assessments across different tissue and cell types, a solid starting reference point, not a health verdict.
- PhenoAge: Best for evaluating physiological aging, immune decline, and metabolic stress specifically.
- GrimAge: The premier clock for predicting mortality risk and life expectancy, thanks to its smoking and plasma-protein surrogates.
- DunedinPACE: The most sensitive tool for measuring your current pace of aging and evaluating whether a lifestyle change or intervention is actually working short-term.
For an individual asking, 'How am I actually doing?' GrimAge and DunedinPACE currently carry the most meaningful signal. A DunedinPACE score above 1, or a positive GrimAge residual, consistently links to higher future disease and mortality risk genuinely useful information, even without a guarantee of what changing it will do.
If you're comparing where to actually get tested, our guide to longevity clinics in Europe covers several providers offering validated epigenetic clock testing as part of a broader assessment.
Bottom Line
There's no single correct epigenetic clock; there's only the right clock for the question you're actually asking. Want a rough sense of your baseline biological age? Horvath does that well. Trying to understand your mortality or disease risk? PhenoAge and GrimAge are built for exactly that. Testing whether a new diet or training block is actually working? DunedinPACE is the one designed to notice.
What all four have in common is this: they're powerful, genuinely validated research tools that still fall short of individual diagnostic proof. A high score is a signal worth acting on more movement, better sleep, cutting out smoking, and tighter nutrition not a verdict to panic over. Treat the number as a starting point for change, not a final judgment on how much time you have left.
Frequently Asked Questions
What's the difference between Horvath, PhenoAge, and GrimAge?
Horvath predicts your actual chronological age accurately but is weak at forecasting disease. PhenoAge and GrimAge are trained on health risk and mortality instead PhenoAge focuses on immune and metabolic aging, while GrimAge is specifically built to predict lifespan.
Which epigenetic clock is the most accurate?
It depends on what accurate means to you. Horvath is most accurate at guessing chronological age. GrimAge is generally considered the most accurate at predicting mortality and lifespan-related outcomes.
Can you actually lower your epigenetic age?
Yes, exercise, better nutrition, weight management, quitting smoking, and stress reduction have all shown measurable effects on epigenetic age, particularly on second- and third-generation clocks. Whether this change reflects true biological reversal or just a shift in methylation markers is still being studied.
What does DunedinPACE actually measure?
Your current pace of aging is how many biological years you're accumulating per calendar year rather than a static age estimate. A score above 1 means you're aging faster than typical.
Are epigenetic clocks reliable enough to trust?
They're the most successful aging biomarkers developed so far, but they still face real limitations: tissue specificity, lab variation, and limited testing across diverse populations. They're best used alongside other clinical and functional data, not as a standalone verdict.
Do epigenetic clocks prove that lowering your score extends your life?
Not yet. Lower epigenetic age scores correlate with lower mortality risk, but it hasn't been proven that actively lowering your score causes you to live longer. That's an important distinction between association and causation.
Disclaimer
ExtendMy.Life is an independent longevity and aesthetic concierge service. It does not provide medical advice, diagnosis, or treatment. All medical consultations, clinical assessments, diagnostic testing, treatment recommendations, and procedures are provided exclusively by licensed healthcare professionals at independent partner clinics. ExtendMy.Life's role is to recommend suitable providers, coordinate the client journey, and simplify access to trusted clinics.
References
Oblak, L., van der Zaag, J., Higgins-Chen, A.T., Levine, M.E. and Boks, M.P. (2021) 'A systematic review of biological, social and environmental factors associated with epigenetic clock acceleration', Ageing Research Reviews, 69, 101348.
Margiotti, K., Monaco, F., Fabiani, M., Mesoraca, A. and Giorlandino, C. (2023) 'Epigenetic clocks: In aging-related and complex diseases', Cytogenetic and Genome Research, 163(5–6), pp. 247–256.
McCrory, C., Fiorito, G., Hernandez, B. et al. (2021) 'GrimAge outperforms other epigenetic clocks in the prediction of age-related clinical phenotypes and all-cause mortality', The Journals of Gerontology: Series A, 76(5), pp. 741–749.
Lu, A.T., Quach, A., Wilson, J.G. et al. (2019) 'DNA methylation GrimAge strongly predicts lifespan and healthspan', Aging, 11(2), pp. 303–327.
Horvath, S. (2013) 'DNA methylation age of human tissues and cell types', Genome Biology, 14(10), R115.
Belsky, D.W., Caspi, A., Corcoran, D.L. et al. (2022) 'DunedinPACE, a DNA methylation biomarker of the pace of aging', eLife, 11, e73420.
Bischoff-Ferrari, H.A., Gängler, S., Wieczorek, M. et al. (2025) 'Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation clocks of biological aging in older adults from the DO-HEALTH trial', Nature Aging, 5(3), pp. 376–385.
Jokai, M., Torma, F., McGreevy, K.M. et al. (2023) 'DNA methylation clock DNAmFitAge shows regular exercise is associated with slower aging and systemic adaptation', GeroScience, 45(5), pp. 2805–2817.
Apsley, A.T., Etzel, L., Ye, Q. and Shalev, I. (2025) 'From population science to the clinic? Limits of epigenetic clocks as personal biomarkers', Epigenomics, 17(18), pp. 1447–1461.