Epigenetic Clocks: Measuring Biological Age
From Horvath to DunedinPACE — how DNA methylation patterns reveal your biological age, and what the numbers actually mean.
What Is DNA Methylation?
DNA methylation is a chemical modification where a methyl group (-CH3) is added to cytosine bases in your DNA, typically at CpG sites (where cytosine is followed by guanine). These modifications don't change your genetic code — they change which genes are active or silent. Think of methylation as sticky notes on a book: the words don't change, but the notes tell the reader which pages to pay attention to.
Critically, methylation patterns change predictably with age. Some sites gain methylation, others lose it, in a pattern so consistent that algorithms can estimate your age from a DNA sample with remarkable accuracy.
The Major Clocks
Horvath Clock (2013)
353 CpG sites | Pan-tissue
The original. Steve Horvath's landmark clock works across almost all tissue types — blood, brain, liver, kidney. It was trained on ~8,000 samples and predicts chronological age with a median error of 3.6 years. "Horvath age acceleration" (biological age minus chronological age) predicts mortality independently of other risk factors.
Hannum Clock (2013)
71 CpG sites | Blood-specific
Published the same year as Horvath's, Hannum's clock is optimized for blood samples. It uses fewer CpG sites but performs comparably for blood-based age estimation. Less versatile across tissues than Horvath's clock.
PhenoAge / GrimAge (2018–2019)
513 / 1,030 CpG sites | Blood
Second-generation clocks that go beyond chronological age. Instead of predicting birth date, they predict mortality risk. PhenoAge incorporates clinical biomarkers (albumin, creatinine, glucose, CRP, etc.). GrimAge adds smoking history and plasma protein surrogates. GrimAge is currently considered the best predictor of lifespan and healthspan outcomes.
DunedinPACE (2022)
173 CpG sites | Blood | Rate-of-aging
The newest and arguably most useful clock for tracking interventions. Unlike other clocks that estimate cumulative biological age, DunedinPACE measures your current rate of aging— are you aging faster or slower than 1 year per calendar year right now? Developed from the Dunedin birth cohort study tracking ~1,000 people from birth. A score of 1.0 means you're aging at the normal rate; below 1.0 means slower; above means faster.
Why it matters: DunedinPACE is most responsive to lifestyle changes, making it the best clock for evaluating whether your supplement protocol or exercise routine is actually working.
Which Test Should You Get?
See our full Biological Age Testing page for provider comparisons. In brief:
- For comprehensive analysis: TruDiagnostic TruAge COMPLETE — reports all major clocks including DunedinPACE, immune cell composition, and telomere length.
- For tracking interventions: Focus on DunedinPACE score — it's the most responsive to changes. Test every 6–12 months.
- For mortality risk: GrimAge is the strongest predictor, but it's a sobering number — make sure you're ready for that information.
Interpreting Your Results
Age Acceleration
Biological age minus chronological age. If you're 50 but your biological age reads 45, your acceleration is -5 years (good). If it reads 55, your acceleration is +5 years (concerning). A difference of 1–3 years may be within the test's margin of error.
Rate of Aging (DunedinPACE)
Expressed as years of biological aging per calendar year. A score of 0.85 means you're aging at 85% of the normal rate. The mean in population studies is ~1.0, with a standard deviation of ~0.15. Scores below 0.85 or above 1.15 are meaningfully different from average.
What Moves the Clock?
Interventions with the strongest evidence for improving epigenetic age:
| Intervention | Effect | Evidence |
|---|---|---|
| Exercise (consistent) | -3 to -7 years age acceleration | Strong |
| Mediterranean diet | -1.5 to -3 years | Strong |
| Sleep quality (7-9 hrs) | -1 to -3 years | Moderate |
| Smoking cessation | -3 to -5 years (over time) | Strong |
| Weight loss (if overweight) | -1 to -3 years | Moderate |
| Supplements (NMN, etc.) | Unclear | Limited |
Limitations to Keep in Mind
- Test-retest variability: The same sample tested twice can differ by 1–3 years. Don't obsess over small changes between tests.
- Different clocks, different answers: You might be "younger" on Horvath but "older" on GrimAge. Each measures something slightly different.
- Acute confounders: Illness, recent vaccination, extreme stress, and sleep deprivation can temporarily affect results. Test when you're in your normal routine.
- Correlation vs causation: We know epigenetic clocks predict mortality, but we don't yet know for certain that lowering your clock score causes longer life. It might just be a readout of underlying health.
- Commercial incentive: Companies selling biological age tests have an incentive to make the results feel actionable and worth retesting. Keep a healthy skepticism about how much to invest in chasing the number.