Evidence summary: myDNAge Blood Biological Age Test

The myDNAge Blood Biological Age Test is sold by Epimorphy, a company that describes itself as spun out from Zymo Research [4]. You prick a finger at home, collect a few drops of blood into a small tube and post it back [3]. The company says results take 4 to 6 weeks and arrive by email as a report [3][4]. The service is for adults aged 21 and over [10]. Zymo Research separately offers the same kind of DNAge analysis to researchers [11].

The company's pages still mention a urine sample in places, but its product list offers only this blood test and a test for dogs, and the old urine product address now leads to the blood test [3][4][7].

A 2021 sample report on the company's website lists an epigenetic age called DNAge; a percentile comparing that result with people of the same age in the company's own database; a second percentile within your sex; an "Aging Rate"; a methylation-based "Epi-Metabolic Index"; and APOE and MTHFR genotypes, which are DNA-sequence results [8][9]. We do not know whether the report customers receive today is laid out the same way.

DNA methylation is a chemical mark on DNA; at some sites it changes with age. It is read at CpG sites, places in the DNA where it can occur. An epigenetic clock is a formula that turns methylation at chosen CpG sites into an age-like number.

The laboratory extracts DNA from the white blood cells in your sample and reads methylation at a chosen panel of sites by targeted bisulfite sequencing, a method that reads only those sites by DNA sequencing [5][7]. The company says the panel covers over 2,000 sites [6][7]. That is the only methylation measurement.

Everything else is calculated:

• DNAge: the output of the company's own predictor [15]. It is built on sites from a well-known clock published by Steve Horvath in 2013, plus extra sites, read by sequencing rather than on a chip, and fitted again to calendar age [6][12].
• Percentiles: your DNAge compared with people of the same age, and of the same sex, in the company's database [8].
• Aging Rate: worked out from your DNAge and your calendar age [8][9] (see Retesting).
• Epi-Metabolic Index: a score from methylation at a set of genes linked to metabolism [8][9]. In the sources we reviewed, its formula and validation are not published.

The 2013 Horvath clock is a different thing. It is a published formula of 353 CpG sites, built on chip data from many tissues, trained to predict calendar age, and freely available [13]. In its test data, half of the estimates were within 3.6 years of calendar age [13]. myDNAge does not use that formula as published. In the sources we reviewed, its own predictor's list of sites, its weights and the data it was last trained on are not published. Published descriptions of the panel also differ: a 2017 company poster and some research studies describe more than 500 sites [12][14][19][20], a 2024 study and the current website more than 2,000 [6][15], and the company says it uses "more than 90%" of the 353 Horvath sites [6]. We do not know which version customers receive today.

The myDNAge Blood Biological Age Test is currently rated Emerging on Compare [1]. The rating concerns the DNAge result from a home finger-prick blood sample, as an epigenetic age estimate trained to match calendar age. The percentiles and the Aging Rate are worked out from that result, so the same limits apply to them. The rating does not cover the Epi-Metabolic Index, the APOE or MTHFR genotypes, the lifestyle suggestions in the report, disease risk or diagnosis, or the use of Zymo's DNAge service in research. This page explains that rating. It does not change it.

We rate myDNAge Emerging for its own epigenetic-age result: there are published, peer-reviewed data on the DNAge predictor, but all of the accuracy data we found come from the company's developers or from studies they co-authored, the predictor itself is not published, and in the sources we reviewed no study has checked the home finger-prick test.

• Accuracy data come from the developers. A 2017 conference poster by Zymo and Epimorphy scientists, which was not peer-reviewed, reports that half of blood estimates were within 1.9 years of calendar age [12]. It does not say how many samples this figure comes from or whether they were kept apart from the samples used to build the predictor [12]. A 2021 forensic study with Zymo co-authors estimated age from 9 samples of finger-prick blood dried on car surfaces, with a close fit to the donors' ages (R² = 0.963, where 1 would be perfect) [14]. A 2024 heart study with Zymo co-authors reported that, in a reference group of 421 healthy volunteers, DNAge was on average 0.6 years above calendar age, with a spread of about 3.6 years in the 400 adults [15]. It does not say whether those volunteers helped build the predictor [15].
• Small and early. The 2024 study calls itself a "proof-of-concept study" and "a pilot study" [15].
• Other researchers used it, but did not check it. Researchers at several universities and hospitals have used Zymo's DNAge service in their own studies [19][20][21]. In each study we checked, the age was worked out with Zymo's own predictor [19][20][21], and two say the samples were analysed by Zymo's service or at its facility [20][21]. These studies used the result but did not test its accuracy against another method.
• No check of the home test. In the sources we reviewed, no study has examined the home finger-prick pathway that customers use, or how much results vary between two blood samples from the same person.

Our methodology rates a commercial test "on the evidence for that product’s own output, not for the scientific method behind it" [2]. It names "Classic epigenetic clocks (Horvath, Hannum) for estimating chronological age" as an example of strong evidence [2], but that applies to the published clock, not to a company's own predictor on a different laboratory method.

The test fits our Emerging tier, which describes "pilot human studies", "No independent replication in humans yet" and evidence that "May be based on small sample sizes or vendor-only datasets" [2]. It does not reach Promising, which asks for "At least one independent replication or large-cohort validation" [2]: in the sources we reviewed, there is none for DNAge or the home test. Promising also lists "Published methodology but may lack full independent reproduction" [2]; the general approach is described, but the predictor itself is not published. Two criteria in our Insufficient tier also fit, at least in part: "Proprietary models without methodological transparency" and "Marketing language exceeds published evidence" [2]. Its first two criteria do not: "Vendor-only claims without published validation" and "No peer-reviewed studies or only vendor-funded studies without independent oversight" [2]. Peer-reviewed data exist, and the two developer co-authored papers were led by university or hospital researchers and funded from outside the company [14][15]. The Insufficient tier's last criterion, "No independent replication or external validation" [2], overlaps with the Emerging description "No independent replication in humans yet", so we do not treat it as decisive. Our confidence in Emerging is moderate.

A peer-reviewed paper that describes the current predictor and tests it on people it was not built on, an independent comparison with chip-based clocks on the same samples, or published data comparing home finger-prick samples with blood from a vein and with repeat samples could change this rating.

• "Most accurate". The home page calls myDNAge "the most accurate biological age test" [4], and the about page says it delivers "the most accurate biological age test on the market" [6]. The only comparison we found is in the company's 2017 poster: in 51 samples measured both ways, its sequencing-based predictor had a smaller error (2.0 years) than an array-based clock (5.9 years) [12]. That is one comparison by the developers, not peer-reviewed. In the sources we reviewed, no independent study has compared the two on the same samples or compared myDNAge with other tests.
• "Unparalleled precision and reliability". The science page uses this phrase [5]. The published repeat-measurement figures come from the developers and re-measure the same sample [11][12][14]; in the sources we reviewed, there are no data on repeat samples from the same person.
• "Significantly improved upon" the Horvath clock. The science page says this [5]. It rests on the developers' poster [12].
• "Trusted globally by scientists". The science page says the test is "not just validated by tests conducted by our own team" and that scientists have "used it as part of their own research" [5]. Using a service in research is not the same as checking its accuracy, and each study we checked used Zymo's own predictor [19][20][21].
• "Precisely determine biological age". The home page says the test can "precisely determine biological age, as opposed to chronological age" [4]. The predictor was fitted to calendar age [12].
• Aging rate as "healthy" or "unhealthy aging". The report page says "Positive values of aging rate indicate unhealthy aging, while negative values of aging rate indicate healthy aging" [8]. In the sources we reviewed, no study shows that the Aging Rate tells one person whether they are aging healthily; the only related finding is a group difference in one small pilot study [15].
• Metabolic risk. The product page says the Epi-Metabolic Index "quantifies your risk for metabolic diseases" [3]. In the sources we reviewed, it has no published validation, and the company's own terms say the service is not intended to "assess risk of any disease or condition" [10].

The company says lifestyle choices "can influence the methylation marks on our DNA" [7], and its report page says the report includes improvement plans with lifestyle tips "that can improve biological age" [8]. In the sources we reviewed, no published study has tested whether following these plans changes DNAge results, and none has tested whether the home test can detect a change after any intervention. One small study that used the DNAge service drew its 10 participants from a pilot exercise trial; it was a feasibility study, and it is not evidence that exercise changes myDNAge results [20]. This page is not a recommendation to make any particular change.

The company invites you to "track your DNAge® changes if you take multiple tests" [3][4]. The published repeat-measurement data come from the developers and measure the same sample more than once: Zymo reports a median standard deviation below 1 year [11], the 2017 poster that repeat results typically fall within about 1.7 years either side of their average [12], and the forensic study an error of ±1.2 years between repeats [14]. Re-measuring one sample does not show how much results vary between two blood samples taken on different days, posted from home. In the sources we reviewed, there are no such data for myDNAge and no threshold for a meaningful change. For comparison, in a study of the published chip-based Horvath clock, measuring the same 36 blood samples twice gave results a median 1.8 years apart, and up to 4.8 years apart [18]. These studies use different measures of spread, so the figures cannot be compared directly.

The Aging Rate does not show how fast you are aging between tests. In the 2021 sample report it is worked out "Based on your DNAge and calendar age" and shown as a ratio near 1 [9]; the sample figure equals the age estimate divided by calendar age. The company's report page describes the aging rate differently, as "the difference between predicted DNAge and the chronological age" [8]. Either way, it restates one age estimate against your calendar age: a gap built up over your whole life so far, not a current speed. For example (a made-up case), an estimate of 55 at age 50 gives a ratio of 1.10; if the estimate stays at 55 when you are 51, the ratio falls to 1.08, although nothing measured has changed.

A myDNAge result is one estimate of epigenetic age from a few drops of finger-prick blood, made by the company's own predictor, which was fitted to calendar age [12]. A result younger or older than your calendar age is a gap at one point in time. The percentiles compare you with the company's own database, whose size and make-up we found no published details of [8]. In research cohorts, first-generation clocks such as the original Horvath clock show only modest links with death from any cause: an independent review found about 8% higher risk per 5 years of age acceleration, with signs that positive results were more likely to be published [17]. In the sources we reviewed, there are no such data for DNAge itself. A 2024 study found higher DNAge results, on average, in adults living after a particular heart operation than in healthy volunteers [15]; that is a group difference, not a way to read one person's result. The company says the service is "not intended to be health information or medical data or to be used to screen, diagnose, treat, prevent or assess risk of any disease or condition", and that it "has not been cleared or approved by U. S. Food and Drug Administration" [10]. The reviewed studies do not show whether using this test to guide decisions improves health.

Every accuracy figure we found for DNAge comes from Zymo or Epimorphy scientists, or from papers with Zymo employees among the authors [12][14][15]. The 2021 forensic study declares two Zymo employees among its authors and was funded by charitable trusts, NASA and the US National Institutes of Health, among others [14]. The 2024 heart study declares three Zymo employees among its authors, one also listed at Epimorphy, and was funded by a children's hospital charity, with the lead author supported by a British Heart Foundation fellowship [15]. The 2017 poster lists Steve Horvath as a co-author [12]. His 2013 paper states that the University of California owns a provisional patent application on the clock, with him as a named inventor [13]. Epimorphy and Zymo say they hold an exclusive licence to the 2013 clock's 353 sites [4][7].

Every published study of DNAge in people that we checked used Zymo's own predictor [14][15][19][20][21]. The research studies we checked used whole-blood samples taken in a research setting [15][19], apart from the forensic study's dried drops [14]. In a separate study of 20 older adults that did not test myDNAge, published clocks measured on a chip gave similar average ages from finger-prick blood and vein blood, both taken by trained staff [16]. That is a group average for different clocks on a different technology. In the sources we reviewed, no study has compared a home myDNAge sample with blood from a vein in the same people.

The company also offers lifestyle improvement plans in its report [8].

For current price and availability, see Compare. If you spot an error, please tell us.