DNA Methylation Testing: How the Clocks Actually Work

DNA methylation testing estimates biological age by reading chemical tags attached to your DNA rather than the DNA sequence itself. An algorithm called a clock converts the pattern of those tags into a number. The method sits behind most at-home epigenetic tests, including the ones we carry, and understanding how the clocks were built is the difference between reading your result well and reading it badly.

This guide goes a level deeper than the biological age overview. If you want the short version of what these tests are for, start there. If you want to know why one report says you are 41 and another says your pace of aging is 0.94, this is the page.

What methylation is, in plain terms

Your genome is the same in every cell and stays essentially fixed for life. What changes is how it is annotated. A methyl group is a small chemical tag that cells attach to DNA at specific positions, most often where a cytosine sits next to a guanine. Those positions are called CpG sites, and there are tens of millions of them across the human genome.

Methylation is how a cell manages which genes are available for use. A liver cell and a skin cell carry identical DNA and behave nothing alike, because different regions are tagged and untagged. Those patterns are not static. They drift over a lifetime, and the drift is consistent enough across people that it can be measured and modeled.

That consistency is the entire basis of epigenetic age testing. Read the methylation state at a defined set of CpG sites, feed it to a model trained on thousands of samples, and the model returns an estimate.

How a clock gets built

A methylation clock is a statistical model, not a biological instrument. Researchers take a large set of samples where something is already known, measure methylation across many CpG sites, and train an algorithm to predict that known thing from the pattern. What the clock is trained to predict determines what it is useful for, and this is where most confusion about these tests comes from.

First generation: trained on calendar age

The earliest widely used clocks, published in 2013 by Steve Horvath and separately by Gregory Hannum, were trained to predict chronological age from methylation. They did it well, which was a genuinely surprising result and launched the field.

The limitation is built into the design. A model trained to reproduce calendar age is, at its best, a very good calendar. If it returns your exact age, it is working perfectly and telling you nothing about your health. The interesting signal was always the error term, the gap between predicted and actual age, which researchers began calling age acceleration.

Second generation: trained on health outcomes

The next wave inverted the target. Instead of predicting birthdays, clocks such as PhenoAge and GrimAge were trained against clinical biomarkers and long-term outcomes recorded in cohort studies. The output is still expressed in years, which keeps it readable, but the model is estimating something closer to physiological condition than time elapsed.

This is the generation worth paying attention to, and it is why test choice matters more than test frequency. An outcome-trained estimate that sits below your calendar age is a different and more meaningful statement than a calendar-trained one doing the same.

Third generation: pace rather than age

The most recent shift stopped reporting an age at all. DunedinPACE, built from the Dunedin longitudinal study in New Zealand, was trained on how fast a large group of people physically changed while being measured repeatedly over decades. It reports a rate: how many biological years you are accumulating per calendar year.

A pace of 1.0 means you are aging at the population average. Below 1.0 is slower, above is faster. For anyone testing to find out whether a routine is working, a pace marker is the most directly useful output of the three generations, because rate responds to change faster than an accumulated age estimate does.

What the reports we carry actually contain

The TruAge Epigenetic Test reports three things rather than one number, and each answers a different question:

• OMICmAge is an outcome-trained age estimate, built with biomarker data informing the model rather than calendar age alone.
• SymphonyAge breaks the estimate into 11 organ systems, so a single number becomes a map of where the estimate is being pulled.
• DunedinPACE reports the pace described above.

The TruHealth Biomarker Test runs on the same collection method but reports epigenetically estimated biomarkers across inflammation, immune function, metabolic health, nutrition, toxin exposure, and oxidative defense. It answers what is going on rather than how old you seem. The combined kit returns both report sets from one finger-prick sample.

The reliability question, honestly

Early clocks drew fair criticism on test-retest precision: run the same sample twice and the two answers could differ by more than users expected. That criticism produced a real fix. A later approach rebuilt clocks using principal components rather than individual CpG sites, which averages across correlated positions instead of leaning on a handful of them, and the newer generation of clocks was designed specifically around that improvement.

The practical consequence for you is simple. Treat any single reading as an estimate with a margin around it. Do not read a two-year move between tests as a two-year change in your body. Compare trends across two or three tests taken under similar conditions, and let the direction carry the meaning rather than the decimal.

What moves a methylation result

Research has measured shifts in these markers alongside changes in sleep, training load, body composition, and smoking status, with pace markers generally responding sooner than accumulated-age estimates. That is the honest summary, and it comes with two caveats worth stating plainly: individual results vary widely, and nothing sold on this site is offered as a way to guarantee a particular number.

What a test genuinely does is close the loop. Pick one change you will actually sustain, whether that is regular heat exposure, a consistent red light routine, or simply sleep, hold it for six months, and let the retest tell you whether your markers moved. Changing six things at once feels productive and teaches you nothing.

Things a methylation test is not

• It is not a diagnosis and does not screen for any disease. A favorable estimate is not a clean bill of health.
• It does not predict how long you will live. These are population-trained statistics applied to one person.
• It is not comparable to another company's test. Different CpG sets, different training data, different answers. Retest with the same test you started with.
• It does not replace bloodwork or a physician. Kove does not provide medical advice; see the medical disclaimer.

How the sample gets taken

Every epigenetic kit we carry uses an at-home finger-prick blood spot on a collection card, no fasting and no blood draw, returned to a certified lab in a prepaid mailer. TruDiagnostic reports 10 to 14 business days from the point the lab receives your sample. Results arrive digitally.

Prices as of September 2026; product pages are the source of truth.

Test	Primary output	Price
TruAge Epigenetic Test	OMICmAge, SymphonyAge across 11 systems, DunedinPACE	$499
TruHealth Biomarker Test	Epigenetic biomarker estimates across six domains	$499
TruAge + TruHealth	Both report sets from one collection	$850

Methylation is not the only way to estimate biological age. Glycan testing reads a different layer entirely, and the two are compared in epigenetic versus glycan testing.

Frequently asked questions

What is a CpG site?

A position in the DNA sequence where a cytosine sits directly next to a guanine. These are the positions where methyl tags attach, and methylation clocks read the tagging state across a defined set of them.

Is DNA methylation testing the same as genetic testing?

No. Genetic testing reads your DNA sequence, which does not change. Methylation testing reads chemical tags sitting on top of that sequence, which do change through life. The two answer completely different questions.

Which methylation clock is the most accurate?

It depends what you mean by accurate. Calendar-trained clocks predict birthdays best and tell you least. Outcome-trained clocks and pace markers are more useful for tracking whether a routine is working, which is why current reports include several rather than one.

How much does a methylation test cost?

The kits we carry are $499 for TruAge or TruHealth individually and $850 for both together, as of September 2026, with free US shipping and the Price Beat guarantee.

Can I use HSA or FSA money?

Biological age testing is among the categories that can qualify with a Letter of Medical Necessity, decided per order by an independent licensed provider through Truemed at checkout. Details on the HSA and FSA page.

How soon after a lifestyle change should I retest?

Six to twelve months. The biology these clocks read moves on a timescale of months, so testing sooner mostly measures noise. Timing is covered in full in how people retest.

The short version

Methylation clocks are statistical models, and what they were trained to predict decides what they are good for. Calendar-trained clocks are impressive and largely uninformative. Outcome-trained estimates and pace markers are the ones worth reading, single results deserve a margin of doubt, and the trend across repeated tests under similar conditions is the only output that reliably means something.

Start with the TruAge Epigenetic Test at $499, or see how measurement anchors everything else in building your first stack.