Two people of the same age can differ markedly in strength, thinking skills and vulnerability to illness. A new study suggests that blood DNA may reflect some of those differences in two complementary ways: changes in average chemical markings, and changes in how widely those markings differ between people.
Researchers analyzed DNA methylation in 1,445 Canadian adults aged 45 to 85. Their findings suggest that looking at variability can reveal signals that an analysis of averages alone would miss. An experimental biomarker combining both approaches was associated with mortality, but it is not a clinical aging test or a way to tell an individual how long they will live.
What the researchers studied
The study, published September 17 in The Journals of Gerontology: Series A, used blood samples from the Canadian Longitudinal Study on Aging. It examined methylation in relation to a Frailty Index based on accumulated health deficits, cognitive function and physical function.
This differs from our earlier report on frailty and circulating DNA. That work investigated cellular damage and DNA fragments outside intact cells, with a methylation analysis involving only five selected participants per group. The new study asks whether methylation averages and between-person variability capture different aspects of health in a larger aging cohort.
Measuring the average and the spread
DNA methylation involves small chemical tags called methyl groups attached to DNA. These tags are associated with regulation of gene activity. Their effects depend on where they occur and the surrounding biological context; they do not simply turn individual genes on or off.
Many methylation studies ask whether a particular location has more or less methylation, on average, in people with poorer health. A variability analysis asks another question: how widely do the measurements spread apart between people?
Imagine two groups of measurements with a similar average. In one, most values sit close together. In the other, some values are much higher and others much lower. An average can conceal that difference in spread. Here, variability refers to differences between people at particular genomic locations, not proof that one person’s entire epigenome is becoming unstable.
Two partly separate sets of signals
The researchers identified 448 differentially methylated regions and 488 differentially variable regions associated with the health measures. The first set concerns average methylation differences. The second concerns differences in variability.
There was little overlap between the two sets, and their gene coverage differed. The findings therefore suggest that the approaches provide complementary information about how health differs across people as they age.
Genes overlapped by regions with average methylation differences were enriched for immune and inflammation-related pathways. Enrichment means that these biological pathways appeared more prominently in the analysis than expected under its statistical comparison. It does not establish that methylation caused inflammation or that inflammation caused the health differences.
The variability analysis highlighted additional localized signals enriched in CpG islands, stretches of DNA containing many of the sequence sites where methylation is commonly studied. The authors interpret the pattern as regionally structured. It should not be expanded into a claim that aging produces universal disorder across the epigenome.
Combining the information
The team combined significant CpG sites from the two analyses into an experimental epigenetic biomarker. The combined measure was associated with all-cause mortality and showed better discrimination than measures built from methylation differences or variability alone. A similar overall pattern was reproduced in the Baltimore Longitudinal Study of Aging.
Discrimination describes a model’s ability to distinguish outcomes in the population being studied. Better discrimination does not, by itself, demonstrate accurate estimates of an individual’s remaining lifespan, useful clinical thresholds or improved medical decisions.
What this study does not establish
This is observational human biomarker research. The associations do not demonstrate that the identified methylation patterns cause frailty, poorer cognitive or physical function, or death. Molecular measurements can reflect health conditions and other influences as well as processes that contribute to them.
The biomarker remains experimental. Replication of an overall pattern is encouraging, but it does not make the measure a validated clinical biological-age test. The authors call for larger cohorts with more mortality events, which are needed to assess how reliably the findings hold up.
The study also did not test whether changing these methylation patterns improves health. It provides no demonstration that altering DNA methylation would slow aging or extend lifespan.
The Lifespan Brief assessment
The contribution is a broader approach to measurement: the spread of methylation values may contain information that their averages leave out. Combining the two could help researchers develop more informative aging biomarkers, provided the additional information proves reproducible and useful in further studies.
Evidence level: Early Research. Human observational biomarker research with an independent-cohort comparison. Clinical use remains unvalidated.
Primary source and reporting scope
Vishnyakova and colleagues. DNA methylation variability provides a complementary epigenetic signature of aging heterogeneity: Findings from the Canadian Longitudinal Study on Aging and the Baltimore Longitudinal Study of Aging. The Journals of Gerontology: Series A, September 17, 2026. Accepted manuscript. DOI: 10.1093/gerona/glag234.
This report is based on the publisher’s abstract. The full manuscript PDF was inaccessible at publication, so detailed methods and performance estimates have not been independently reviewed.
Featured image: AI-generated conceptual editorial illustration, not a molecular measurement or study figure.