Belsky · eLife 2022 · biomarker derivation and cohort validation study · n=?

DunedinPACE, a DNA methylation biomarker of the pace of aging.

Cited 1054 times in the scientific literature.

Level 3 - non-randomized controlled study

Longitudinal birth cohort derivation with validation across secondary cohorts

PubMed 35029144 · doi:10.7554/eLife.73420 · record verified 2026-08-27

What was done

Researchers modeled multi-organ biological aging using longitudinal data from the 1972–1973 Dunedin birth cohort, which tracked within-individual decline across 19 physiological indicators over two decades across four time points. Using elastic-net regression restricted to DNA methylation probes with high test-retest reliability, they trained a single-time-point blood biomarker (DunedinPACE). The algorithm was then tested for reliability and predictive validity across five independent external datasets.

What was found

The abstract does not provide specific numerical values, effect sizes, or confidence intervals. It reports that DunedinPACE demonstrated high test-retest reliability, correlated with childhood adversity in young adults, and predicted incident morbidity, disability, and mortality. Effect sizes were reported as comparable to the GrimAge epigenetic clock, with DunedinPACE providing incremental predictive value when combined with GrimAge.

Why it matters

DunedinPACE provides an epigenetic biomarker designed specifically to measure the current rate of biological aging rather than accumulated chronological or biological age, offering a tool to evaluate anti-aging and geroprotective interventions.

Limits

The abstract provides no sample sizes (n) for the training or validation cohorts and reports no exact numerical statistics, effect estimates, or error margins. The biomarker was derived from a single geographic birth cohort, and the abstract does not describe the demographic or ancestral diversity of the validation cohorts.

Cited by