The Bottom Line
- Best overall test: TruDiagnostic TruAge COMPLETE — runs DunedinPACE, OMICmAge, and 11 organ-system ages on 1M+ CpG sites from blood
- Best for mortality prediction: Any test running GrimAge v2 — HR 1.54 per standard deviation of acceleration in the largest head-to-head comparison (18,859 people, 174 disease outcomes)
- Best for tracking interventions: DunedinPACE — the only clock that detected caloric restriction's effect in the CALERIE RCT when GrimAge and PhenoAge showed nothing
- Skip: Saliva-based tests — up to 25-year variance between identical samples. Blood is the only serious sample type.
The biological age testing market has exploded. A dozen companies now sell you a number that's supposed to tell you how fast you're aging. Most of them are running clocks developed 10+ years ago on too few CpG sites with proprietary algorithms they won't disclose.
This matters because these tests cost $150-$500 per draw, and the wrong one will give you a number that means nothing — or worse, a number that changes randomly and convinces you your interventions are working when they're not.
We're going to break this down the way it should have been done from the start: what each clock actually measures at the molecular level, which ones predict death (the only validation that matters), and which commercial tests give you access to the real science.
Three Generations of Epigenetic Clocks — And Why It Matters
Not all biological age tests are created equal. The field has gone through three distinct generations, and each one represents a fundamentally different approach to measuring aging. Understanding this is critical because some commercial tests are still selling you first-generation technology at third-generation prices.
Horvath (2013) & Hannum (2013)
These clocks were trained to predict chronological age from DNA methylation patterns. Horvath uses 353 CpG sites across 51 tissue types; Hannum uses 71 CpG sites from blood only. Both achieve R ≥ 0.96 correlation with chronological age — impressive for estimating your birthday, useless for predicting your deathday. They tell you if your methylation patterns look older or younger than your birth certificate. They do not tell you if you're likely to die sooner or get disease. When compared to second- and third-generation clocks for mortality prediction, they consistently underperform.
PhenoAge (2018) & GrimAge (2019)
The paradigm shift. Instead of training on chronological age, these clocks were trained on health outcomes. PhenoAge uses DNA methylation surrogates for nine clinical biomarkers (albumin, creatinine, glucose, CRP, lymphocyte %, mean cell volume, RDW, alkaline phosphatase, white blood cell count). Each year of PhenoAge acceleration correlates with a 4.5% increase in mortality risk. GrimAge goes further — it uses methylation-based surrogates for seven plasma proteins linked to disease plus smoking pack-years. One year of GrimAge acceleration = HR 1.10 for all-cause mortality. GrimAge v2 (2022) added hs-CRP and HbA1c surrogates and remains the single strongest mortality predictor among all epigenetic clocks.
DunedinPACE (2022)
The most important conceptual leap. DunedinPACE doesn't estimate your biological age at all — it measures your speed of aging. Built from 19 organ-system biomarkers tracked longitudinally over 20+ years in 1,037 people from the Dunedin birth cohort (born 1972-73), it outputs a single number: your pace. A score of 1.0 means you're aging at the average rate. Below 1.0 means slower. Above 1.0 means accelerated. This velocity-based approach is why DunedinPACE detects intervention effects within 6-12 months when static clocks need years.
Why This Matters For Your Wallet
If a test is running only first-generation clocks (Horvath, Hannum), you're paying for a sophisticated way to guess your age — something your birth certificate does for free. You want second- or third-generation clocks. Specifically, you want DunedinPACE for tracking interventions and GrimAge v2 for mortality risk assessment.
The Evidence: Which Clock Actually Predicts Death?
Marketing copy is cheap. Mortality data isn't. Here's what the large-scale comparisons actually show.
The 18,859-Person Head-to-Head (Nature Communications, 2025)
The largest unbiased comparison to date: 14 epigenetic clocks tested against 174 incident disease outcomes over 10 years in the Generation Scotland cohort. The results were clear.
GrimAge v2 demonstrated the strongest mortality prediction: HR 1.54 per standard deviation of age acceleration (P = 7.1 × 10-62). That's not a marginal effect — that's each standard deviation of acceleration increasing your death risk by 54%.
DunedinPACE showed the strongest association with diabetes specifically (HR 1.44, P = 9.6 × 10-19) and produced strong disease associations across the board.
First-generation clocks (Horvath, Hannum) were consistently outperformed. The study concluded that second- and third-generation clocks "significantly outperform first-generation clocks" across all disease categories.
The BASE-II Mortality Study (2026)
The Berlin Aging Study II compared 14 biomarkers of aging — not just epigenetic clocks, but also grip strength, gait speed, IL-6, CRP, cognitive tests, muscle mass, and frailty phenotype — in 1,083 participants aged 60-80 over 7.4 years of follow-up.
DunedinPACE emerged as the strongest and most consistent predictor of mortality across all biomarkers tested. Not just stronger than other clocks — stronger than grip strength, inflammation markers, and functional fitness tests.
The minimal prediction model — muscle mass + standing balance + DunedinPACE — achieved a C-index of 0.63, nearly matching the full 14-biomarker model's 0.65. Three measurements captured almost all the predictive power of fourteen.
The CALERIE RCT: The Intervention Sensitivity Test
This is the study that separates real clocks from decorative ones. The CALERIE trial — 220 healthy adults randomized to 25% caloric restriction or normal diet for 2 years — is the gold standard for testing whether a clock can detect an intervention effect.
Results:
- PhenoAge: No significant effect detected
- GrimAge: No significant effect detected
- DunedinPACE: Detected a 2-3% slowing in the pace of aging
That 2-3% translates to a 10-15% reduction in mortality risk — comparable to the effect of smoking cessation. DunedinPACE caught it. The static clocks missed it entirely.
This is why velocity matters. GrimAge is the best at telling you your cumulative mortality risk. DunedinPACE is the best at telling you whether what you're doing is working.
The Two-Clock Strategy
The optimal approach isn't picking one clock — it's using two. GrimAge v2 for your mortality risk baseline (are you in trouble?). DunedinPACE for longitudinal tracking (is your protocol moving the needle?). Any test that gives you both is worth the price. Any test that gives you neither is worth nothing.
Commercial Tests Ranked
Now that you know what the science says, here's how the actual products stack up. We're ranking on three criteria: which clocks they run, sample quality (blood vs saliva), and whether the underlying algorithms are published and peer-reviewed or hidden behind proprietary marketing.
TruDiagnostic TruAge COMPLETE
The most comprehensive molecular aging assessment available to consumers.
TruAge COMPLETE runs three major algorithms: OMICmAge (multi-omic-informed biological age, published in Nature Aging), DunedinPACE (pace of aging), and SYMPHONYAge (11 organ-system ages covering brain, heart, liver, kidneys, lungs, immune, inflammatory, blood, musculoskeletal, hormone, and metabolic systems). You also get telomere length estimates, immune cell profiles, a fitness report, and 75+ longevity biomarkers — all from a finger-prick blood sample.
- Highest CpG coverage on the market (1M+ sites)
- DunedinPACE for intervention tracking — peer-reviewed, validated
- Organ-system breakdown tells you where your biology is weakest
- 99%+ reproducibility (ICC ≥ 0.96)
- Subscribe-and-save at $249 makes longitudinal tracking affordable
- Does not include GrimAge v2 (the strongest mortality predictor)
- Dense reports — expect to spend time interpreting or pay for the consultation
- $499 one-time price is steep if you're only testing once
NOVOS Age
Strong clock selection at a lower price point, with genuine organ-age reporting.
NOVOS runs DunedinPACE for pace-of-aging alongside their own biological age estimate, covering 11 organ-system ages and telomere length from ~850,000 CpG sites. Their dataset includes 20,000+ humans for benchmarking, and their algorithms are backed by 45+ published studies.
- DunedinPACE included — you get the intervention-tracking clock that matters
- Blood-based with solid CpG coverage
- $100-200 cheaper than TruAge for similar core data
- Highly sensitive to lifestyle interventions for repeat testing
- Narrower biomarker scope than TruAge COMPLETE
- No OMICmAge or equivalent multi-omic algorithm
- Limited organ-specific granularity compared to SYMPHONYAge
Elysium Index
Consumer-friendly with a large benchmarking dataset, but falling behind on clock generation.
Elysium uses a proprietary machine-learning model developed with Illumina based on the Levine DNAmPhenoAge clock (second-generation). ~1-year variance between repeat tests. Their large consumer dataset gives solid benchmarking, but the test runs on saliva with far fewer CpG sites than blood-based competitors.
- Streamlined, consumer-friendly reporting
- Large benchmarking dataset for age/sex comparisons
- PhenoAge-derived algorithm — second-generation, not first
- Saliva sample — up to 25-year variance between identical samples
- Only ~100,000 CpG sites — 10x fewer than TruAge
- No DunedinPACE — you can't track intervention effects
- Proprietary algorithm limits independent validation
- 6-week turnaround is the longest in the category
myDNAge
The original Horvath clock in a commercial wrapper. Research-validated but outdated for intervention tracking.
myDNAge runs the original Horvath pan-tissue clock — the most extensively peer-reviewed epigenetic clock in existence, analyzing over 2,000 epigenetic markers. It gives you a solid, reproducible biological age estimate with decades of research validation behind it.
- Most peer-reviewed clock in the field
- Blood sample option (avoid the urine kit)
- Excellent for a single baseline measurement
- First-generation clock — weakest mortality prediction
- Not sensitive to short-term interventions
- $299 for a single clock with no pace-of-aging data
- You're paying 2026 prices for 2013 technology
GlycanAge
Not an epigenetic clock. Measures IgG glycan patterns linked to inflammation and immune aging.
GlycanAge is not an epigenetic test — it measures IgG glycan biomarkers associated with chronic inflammation and immune system aging. It's highly responsive to lifestyle and stress changes, making it a useful complement to epigenetic clocks but not a replacement for them.
- Most responsive to lifestyle changes (weeks, not months)
- Captures inflammatory aging that epigenetic clocks may miss
- Blood-based, solid reproducibility
- Not an epigenetic clock — no DNA methylation data
- Cannot replace DunedinPACE or GrimAge for aging assessment
- $329 for a single inflammation proxy — your longevity blood panel already covers hs-CRP and IL-6
Tally Health TallyAge
Celebrity-backed cheek swab with a proprietary algorithm. Science not disclosed.
Tally Health uses a proprietary "CheekAge" algorithm on ~850,000 CpG sites from a cheek swab. Their dataset includes ~8,000 individuals. The lowest price point in the category at $129, with a lifestyle platform attached.
- Lowest entry price ($129)
- Non-invasive collection (cheek swab)
- Lifestyle recommendation platform included
- Proprietary algorithm — not independently peer-reviewed
- Cheek swab, not blood — same variance concerns as saliva
- No DunedinPACE, no GrimAge, no published validation
- You can't verify what clock generation it actually runs
The Head-to-Head Comparison
| Test | Price | Sample | CpG Sites | Key Clocks | Intervention Sensitive? | Published Validation? |
|---|---|---|---|---|---|---|
| TruAge COMPLETE | $499 / $249 sub | Blood | 1,000,000+ | DunedinPACE, OMICmAge, SYMPHONYAge | Yes (DunedinPACE) | Yes |
| NOVOS Age | $294-$349 | Blood | ~850,000 | DunedinPACE + Proprietary | Yes (DunedinPACE) | Partial |
| Elysium Index | ~$299 | Saliva | ~100,000 | PhenoAge-derived | No | Partial |
| myDNAge | ~$299 | Blood | 2,000+ markers | Horvath (1st Gen) | No | Yes |
| GlycanAge | ~$329 | Blood | N/A (glycans) | IgG glycan analysis | Yes (different method) | Yes |
| Tally Health | $129-$249 | Cheek swab | ~850,000 | CheekAge (proprietary) | Unknown | No |
Blood vs Saliva: This Is Not a Close Call
Some tests use saliva or cheek swabs because they're cheaper and less intimidating than a blood draw. The science does not support this tradeoff.
Blood-based epigenetic tests achieve 99%+ reproducibility — test the same sample twice, get the same result. Saliva-based tests can vary by up to 25 years between identical samples. That's not a rounding error. That's a number that could tell you you're biologically 35 one week and 60 the next.
The reason is cellular composition. Blood gives you a consistent cell population (primarily leukocytes) with well-characterized methylation patterns. Saliva is a mix of buccal epithelial cells, leukocytes, bacteria, and food debris in ratios that change hour to hour. The clocks were validated on blood. Running them on saliva introduces noise that the algorithms weren't designed to handle.
Skip Saliva-Based Tests
If you're spending $200+ on a biological age test, get blood drawn. A 25-year variance window makes the data meaningless for longitudinal tracking. You cannot tell whether your DunedinPACE score changed because your protocol worked or because you ate breakfast before swabbing. TruAge and NOVOS both use finger-prick blood kits that you can do at home — there's no reason to accept saliva's limitations.
How to Use These Tests (The Protocol)
Buying a test is step one. Using it correctly is what actually generates actionable data. Most people get one test, panic or celebrate, and never retest. That's $300 wasted.
The Baseline Phase
Setting Your Baseline Correctly
- Test timing: First thing in the morning, fasted, after a normal night of sleep. No alcohol for 72 hours prior. No intense exercise for 48 hours.
- Why it matters: Acute stressors (poor sleep, intense training, alcohol) can transiently affect DNA methylation patterns. You want your baseline to reflect your habitual biology, not last night's bourbon.
- Run two tests if budget allows: Take your first TruAge COMPLETE, then retest 2-3 weeks later under identical conditions. If the results are within 1-2 years, you have a reliable baseline. If they diverge significantly, the noise floor may be higher than the signal you're trying to detect.
- Record everything: Body composition, training program, sleep metrics, supplement stack, stress level, recent illness. You need to know what was happening when each test was taken to interpret changes.
The Intervention Phase
Once you have a baseline, pick one major intervention and hold everything else constant for 6 months. Then retest.
This is where most people fail. They test, then simultaneously change their diet, start a new supplement stack, adjust their training, begin a peptide protocol, and add rapamycin. Six months later, DunedinPACE improves by 0.03 and they have no idea which change caused it — or whether it's noise.
Intervention Tracking Protocol
- Minimum retest interval: 6 months for DunedinPACE. The CALERIE trial detected effects at 12 months with 25% caloric restriction. Your 200mg NMN is a smaller signal — give it time.
- Optimal retest schedule: Baseline → 6 months → 12 months → annually. The first 12-month window gives you two data points to establish a trend. Annual retesting thereafter captures long-term trajectory.
- Budget-conscious schedule: Baseline → 12 months → 24 months. Retest sooner only if you make a major protocol change (starting TRT, beginning rapamycin, significant body composition change).
- What to track: DunedinPACE is your primary outcome. A move from 1.05 → 0.98 is signal. A move from 1.02 → 1.00 might be noise. Look for directional consistency over 2-3 tests before drawing conclusions.
Interpreting DunedinPACE Scores
| DunedinPACE Score | Interpretation | Action |
|---|---|---|
| < 0.85 | Significantly slower aging (~15th percentile) | Maintain current protocol. Retest annually to confirm stability. |
| 0.85 - 0.95 | Slower than average aging | Good trajectory. Optimize weak organ-system scores if using TruAge. |
| 0.95 - 1.05 | Average pace of aging | Room for improvement. Focus on the biggest leverage points — bloodwork, body composition, sleep. |
| 1.05 - 1.20 | Accelerated aging | Identify root causes. Get a full longevity blood panel. Address metabolic dysfunction, inflammation, hormonal deficits. |
| > 1.20 | Significantly accelerated aging | Medical evaluation warranted. Rule out undiagnosed chronic disease, insulin resistance, chronic infection, or systemic inflammation. |
What These Tests Cannot Tell You
Epigenetic clocks are powerful. They are not oracles. Here's what they miss.
They don't predict your death date. A GrimAge acceleration of +5 years doesn't mean you'll die 5 years early. It means your methylation patterns resemble those of people who, on average, have higher mortality rates. Population-level statistics don't map linearly to individual outcomes.
They don't capture everything that matters. The 2025 Nature Communications study found that only 32 of 176 significant clock-disease associations improved prediction (AUC) by more than 1% when added to traditional risk factors. Your standard bloodwork — ApoB, fasting glucose, hs-CRP, hormone panel — still carries enormous predictive weight. Epigenetic clocks complement this data. They don't replace it.
They are not diagnostic. No biological age test is FDA-approved for diagnosing any condition. A high biological age is a signal to investigate further, not a diagnosis in itself.
Short-term fluctuations are noise. A single DunedinPACE reading is a snapshot. You need at least two data points, ideally three, before concluding that your aging trajectory has genuinely changed. Single readings can be influenced by acute illness, recent vaccination, major stress events, or measurement variability.
The Decision Framework
| Your Goal | Recommended Test | Why |
|---|---|---|
| Comprehensive baseline + ongoing tracking | TruAge COMPLETE ($249/test with subscription) | DunedinPACE + organ ages + largest CpG coverage. Subscribe-and-save makes retesting sustainable. |
| Intervention tracking on a budget | NOVOS Age ($294-$349) | DunedinPACE from blood at a lower price point. The core metric you need for ~$300. |
| One-time curiosity check | NOVOS Age ($294) | No reason to pay $499 for a test you'll only take once. NOVOS gives you DunedinPACE. |
| Research comparison (matching published studies) | myDNAge ($299) | Horvath clock matches most published literature. Useful only for direct comparisons. |
| Inflammation-specific aging | GlycanAge + standard longevity bloodwork | IgG glycans capture inflammatory aging. But your blood panel may already cover this. |
| Lowest possible entry point | NOVOS Age > Tally Health | We'd rather you wait and save for a blood-based test than buy a cheap cheek swab with an unvalidated algorithm. |
Where This Field Is Going
The biological age testing space is moving fast. Several developments are worth watching.
Brain-specific pace of aging: DunedinPACNI, published in Nature Aging in 2025, estimates the longitudinal pace of brain aging from a single MRI scan. This extends the DunedinPACE concept to neurological aging and could become a critical tool for Alzheimer's risk assessment and cognitive longevity tracking.
Population-level norms: A 2026 preprint established normative DunedinPACE distributions across age, sex, and demographic groups. This will sharpen interpretation — knowing you're at the 15th percentile for your age and sex is more actionable than knowing your raw score.
Multi-omic integration: OMICmAge (used in TruAge COMPLETE) is just the beginning. The next generation of clocks will likely integrate proteomics, metabolomics, and genomics alongside methylation data. The clocks are getting better, which means the tests will get better — another reason to establish your baseline now.
Cheaper and faster: Array costs continue to fall. The 850,000+ CpG arrays that cost thousands a few years ago now power $250-$350 consumer tests. Expect sub-$200 blood-based DunedinPACE tests within 2-3 years as the market matures.
References
- Lu AT, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging. 2019;11(2):303-327. PubMed
- Belsky DW, et al. DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife. 2022;11:e73420. eLife
- Waziry R, 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;3:248-257. Nature Aging
- Zagkos L, et al. An unbiased comparison of 14 epigenetic clocks in relation to 174 incident disease outcomes. Nature Communications. 2025;16. Nature Communications
- Mak JKL, et al. Comparing fourteen consensus biomarkers of aging: epigenetic pace of aging as the strongest predictor of mortality in BASE-II. GeroScience. 2026. PubMed
- McCrory C, et al. GrimAge Outperforms Other Epigenetic Clocks in the Prediction of Age-Related Clinical Phenotypes and All-Cause Mortality. J Gerontol A. 2021;76(5):741-749. J Gerontol A
- Beck D, et al. DunedinPACNI estimates the longitudinal Pace of Aging from a single brain image to track health and disease. Nature Aging. 2025. Nature Aging
- Bergsma T, et al. Longitudinal changes in epigenetic clocks predict survival in the InCHIANTI cohort. Nature Aging. 2026. Nature Aging
Get the Longevity Protocol Playbook
Weekly deep dives on the science of aging slower. Protocols, bloodwork interpretation, and the interventions that actually move the needle.