How to Measure Your Functional Age

Oak
aging biomarkers methylation phenoage longitudinal
A lab instrument holding an upright glowing DNA strand dotted with bright beads, scanned by three slender reader arms.

Functional age is an estimate of how old your body performs, expressed in years and derived from measurements that predict mortality and disability better than your birth date does. No single accepted test produces it. In practice you assemble it from three layers: a physical performance battery (gait speed, grip strength, chair stands, balance, and ideally VO2max), a blood-chemistry algorithm such as PhenoAge, and a DNA methylation clock. Each layer measures something different, the three correlate only weakly with each other, and a single measurement of any one of them tells you much less than most testing companies imply. The value comes from repeating them on a schedule and looking at the slopes.

What the three layers measure

Before choosing tests, it helps to understand what each layer is sensing, because they sit at different distances from the biology you care about.

The physical battery measures current capacity: what your neuromuscular and cardiorespiratory systems can do today. It is the most directly interpretable layer and the hardest to fake. Blood-chemistry clocks sit one step back, capturing the inflammatory, metabolic, and hematologic state that precedes functional decline by years. Methylation clocks measure something further upstream still, a pattern of methylation at CpG sites (positions in the genome where a cytosine sits next to a guanine and can carry a chemical tag) that drifts with age across tissues. Horvath and Raj argued that this drift reflects the activity of an epigenetic maintenance system rather than accumulated damage per se 1.

Because the layers sit at different depths, they are not interchangeable. In a three-year follow-up of older women, associations between epigenetic clocks and physical functioning were real but modest, and the clocks did not simply restate what a gait speed test tells you 2. A cross-sectional and longitudinal comparison of five clocks, telomere length, and functional capacity in older adults found similarly limited overlap 3. The practical implication is simple. If you want a number that reflects how well you walk, measure how well you walk. If you want a number that might change before your walking does, measure methylation.

The physical battery: protocol and thresholds

This section covers the five tests we run and the thresholds that make their results interpretable. Consistency is what makes the numbers comparable over time. Run the tests in a fixed order, at the same time of day, and in the same footwear. Order matters because fatigue from chair stands degrades gait speed measured afterward.

Start with usual-pace gait speed over 4 meters, with a 1-meter acceleration zone before the timing start and a deceleration zone after the finish. Time with a stopwatch, starting on first foot contact past the start line. Run three trials and take the fastest. Roughly 0.8 m/s is the commonly used threshold below which mobility limitation becomes likely, and healthy adults in midlife typically walk 1.2 to 1.4 m/s at usual pace.

Next, measure grip strength with a calibrated hydraulic dynamometer such as a Jamar or equivalent. Sit with the elbow at 90 degrees and use the second handle position. Take three trials per hand and record the maximum. The European sarcopenia working group uses cutoffs of 27 kg for men and 16 kg for women as thresholds for low strength. Consumer spring-gauge grippers are not comparable to hydraulic units, so pick one device and stay with it.

The remaining three tests cover lower-body strength, mobility, and balance. The thirty-second chair stand is performed with arms crossed over the chest and a standard 43 cm seat height, counting full stands. The Timed Up and Go asks you to rise from a chair, walk 3 meters and turn, then return and sit. A time of 12 seconds or more is the usual flag for elevated fall risk. Single-leg stance is performed with eyes open and capped at 30 seconds, and an inability to hold 10 seconds is the common cutoff.

If you have access to a metabolic cart, add a ramp CPET (cardiopulmonary exercise test) for VO2max, the maximum rate at which your body can use oxygen. It is the single most informative number in the battery and the only one that moves substantially with training within a few months. A submaximal estimate from a chest-strap heart rate monitor is less accurate but repeatable, and repeatability is what you need for slopes.

PhenoAge from a standard blood panel

PhenoAge is the most useful blood-based estimator available, because it was trained against mortality rather than against chronological age. Levine and colleagues fit a Gompertz mortality model in NHANES III and reduced it to nine clinical markers plus chronological age: albumin, creatinine, glucose, log C-reactive protein, lymphocyte percentage, mean cell volume, red cell distribution width, alkaline phosphatase, and white blood cell count 4. The output is a mortality-equivalent age in years, and the resulting measure predicted mortality and healthspan-related outcomes across multiple cohorts 4.

You can compute PhenoAge yourself from any comprehensive metabolic panel plus a CBC with differential and a high-sensitivity CRP, applying the published coefficients directly. The dominant failure mode is units. The model expects albumin in g/L, creatinine in µmol/L, and glucose in mmol/L. It expects CRP in mg/dL before the natural log, MCV in fL, and RDW and lymphocyte percentage as percentages. ALP goes in as U/L and WBC as 1000 cells/µL. Most US labs report albumin in g/dL, creatinine in mg/dL, and glucose in mg/dL. That is why a calculator sometimes returns an age of 12 or 140.

Two behavioral cautions follow from how the model is built. CRP responds acutely to infection and to hard training in the preceding days, and it enters the model after a log transform, so a mild cold can shift your result by years. Draw fasting and hydrated, and wait at least 72 hours after an unusually hard session. Never interpret a single draw on its own. Values outside reference ranges belong in a conversation with a physician rather than in a spreadsheet.

Methylation clocks, and what to ask your lab for

The methylation layer depends heavily on what your lab gives you, so the first request matters. Ask for raw IDAT files. These are the two-file-per-sample binary output of the Illumina array, one for the red channel and one for the green, and without them you cannot recompute anything or apply a clock published after your sample was run. Current arrays are the EPIC v2, covering roughly 935,000 CpG probes.

A workable pipeline looks like this. Read the IDATs with sesame or minfi in R and normalize with noob plus dye-bias correction. Then drop probes with a detection p-value above 0.01 and compute beta values, which express the fraction of methylation at each site. Estimate blood cell proportions by Houseman-style deconvolution before interpreting anything, because first-generation clocks are partly tracking a shift in your leukocyte mixture rather than a change within the cells themselves. Then apply clocks with methylclock or dnaMethyAge. Run Horvath multi-tissue and PhenoAge (the methylation version), plus GrimAge if you have access. Also run DunedinPACE, which reports a rate of aging per chronological year instead of an age. Because EPIC v2 renamed and dropped probes, check probe coverage for each clock before trusting its output.

Tissue choice matters more than people expect. Methylation age is tissue-specific in ways that are not always small: sun-exposed skin carries large genomic blocks of hypomethylation that non-exposed skin does not 5. A buccal swab and a blood draw are therefore not two measurements of the same thing. Pick blood and stay with blood.

Reliability is the real constraint

The strongest reason to be skeptical of a single functional age report is technical noise in the assay itself. Work comparing biological and technical reliability of epigenetic clocks found that several widely used clocks have test-retest reliability low enough that a year-over-year change in one person can be mostly measurement error, and that principal-component reformulations of the same clocks substantially improve this 6. If a report says your epigenetic age fell three years, the correct first question is what the replicate standard deviation of that assay is.

Our approach is to measure that noise rather than assume it away, which leads to a few concrete practices:

  • Run technical replicates. Split one blood draw into two aliquots and have both arrayed, at least once, to establish your own noise floor.
  • Sample on a fixed cadence, every six or twelve months, and fit a line rather than comparing consecutive points.
  • Prefer PC-based clocks and DunedinPACE for tracking change, and treat Horvath’s original clock as a cross-sectional descriptor.

Interpretation deserves the same caution as measurement. Epigenetic age acceleration is associated with brain functional connectivity patterns and cognitive performance in older adults 7, and mechanistic work continues to link methylation drift to specific aging processes 8, but these remain population-level associations. They do not tell you why your own number moved. Kim and colleagues found that high self-control, a trait usually treated as protective, was associated with accelerated epigenetic aging in low-income adolescents 9. That is a useful reminder that these outputs are not scoreboards where every good behavior pushes the number down.

Questions people also ask

How do you assess functional age? Combine the five-test physical battery with PhenoAge computed from a fasting blood panel and a methylation clock run on blood, with IDAT files retained. Repeat on a fixed schedule and read the trend rather than the point estimate.

What is the number one predictor of longevity? Among single measurements you can take at home or in a gym, cardiorespiratory fitness is the strongest, with usual gait speed close behind in older adults. Among blood-derived composites, mortality-trained scores like PhenoAge outperform any individual marker in the panel 4.

What is the number one mistake that makes you age faster? There is no single one, and any article claiming otherwise is guessing. Sustained systemic inflammation is the most consistently visible signal across these measures, since CRP, white cell count, and RDW all load into mortality-trained clocks, but the causal direction in any individual is not something a test can establish.

How healthy should you be at 70? Useful benchmarks are usual gait speed above 1.0 m/s, a Timed Up and Go under 12 seconds, and holding a single-leg stance for 10 seconds. Grip strength above the sarcopenia cutoffs is another. Falling below several of those at once is a reason to see a clinician, not a reason to recalculate your functional age.

Oak builds longitudinal molecular profiles of individuals: whole-genome sequencing, RNA sequencing, proteomics, blood biomarkers, and continuous glucose data, integrated into one model of you. Build your profile.

Footnotes

  1. Steve Horvath, Kenneth Raj. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nature Reviews Genetics, 2018. https://doi.org/10.1038/s41576-018-0004-3 ↩

  2. Tiina Föhr, Timo Törmäkangas, Hannamari Lankila, et al. The Association Between Epigenetic Clocks and Physical Functioning in Older Women: A 3-Year Follow-up. The Journals of Gerontology: Series A, 2021. https://doi.org/10.1093/gerona/glab270 ↩

  3. Ilja Demuth. RELATIONSHIP BETWEEN 5 EPIGENETIC CLOCKS, TELOMERE LENGTH, AND FUNCTIONAL CAPACITY ASSESSED IN OLDER ADULTS: CROSS-SECTIONAL AND LONGITUDINAL ANALYSES. Innovation in Aging, 2023. https://doi.org/10.1093/geroni/igad104.0065 ↩

  4. Morgan E. Levine, Ake T. Lu, Austin Quach, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging, 2018. https://doi.org/10.18632/aging.101414 ↩ ↩2 ↩3

  5. Amy R Vandiver, Rafael A Irizarry, Kasper D Hansen, et al. Age and sun exposure-related widespread genomic blocks of hypomethylation in nonmalignant skin. Genome Biology, 2015. https://doi.org/10.1186/s13059-015-0644-y ↩

  6. Raghav Sehgal, Daniel S. Borrus, John Gonzalez, et al. Biological Versus Technical Reliability of Epigenetic Clocks and Implications for Disease Prognosis and Intervention Response. Aging Cell, 2026. https://doi.org/10.1111/acel.70635 ↩

  7. Andrew J. Graves, Joshua S. Danoff, Minah Kim, et al. Accelerated epigenetic age is associated with whole-brain functional connectivity and impaired cognitive performance in older adults. Scientific Reports, 2024. https://doi.org/10.1038/s41598-024-60311-3 ↩

  8. Adam Li, Zane Koch, Trey Ideker. Epigenetic aging: Biological age prediction and informing a mechanistic theory of aging. Journal of Internal Medicine, 2022. https://doi.org/10.1111/joim.13533 ↩

  9. Hyungkyung Kim, Edith Chen, Gregory E. Miller, et al. Does High Self-Control Accelerate Epigenetic Aging in Low-Income Adolescents?. Journal of Adolescent Health, 2025. https://doi.org/10.1016/j.jadohealth.2024.10.012 ↩