DNA-based model predicted death risk in postmenopausal women
An observational analysis in postmenopausal women aged 50–79 found prediction comparable to one existing biological-age clock and better than another.
Independent replication
Limit: External validation covered only postmenopausal women, limiting applicability to demographically similar populations
Edited by Boris Djordjevic · AgingResearch.News
In an observational study, researchers used DNA methylation data from Framingham Heart Study participants to build a model that predicted death from any cause. DNA methylation consists of chemical tags attached to DNA. The team selected 190 sites on DNA and combined five survival prediction models into a single model.
They tested its performance in a separate cohort of postmenopausal women aged 50–79. It predicted death better than PhenoAge, an existing DNA methylation clock used to estimate biological age. Its performance was statistically comparable to GrimAge, another such clock. The study’s performance estimates came from cohorts of European ancestry.
Why it matters
DNA methylation is one layer of age-related change. The study examined whether combining prediction methods could help capture the varied DNA patterns linked to survival.
Caveats
This observational prediction study did not establish that the DNA patterns caused differences in survival. The authors limited applicability to demographically similar populations, with external validation restricted to postmenopausal women aged 50–79.
The paper
Chosun University
Mechanisms of Ageing and Development, Aug 22, 2026
