Using deep learning networks analyzing DNA methylation patterns, scientists at Hebrew University achieved chronological age (defined as the amount of time since birth) predictions with a median accuracy of 1.36-1.7 years for individuals under 50. Their results appear in the journal Cell Reports.
Using ultra-deep sequencing of>300 blood samples from healthy individuals, we show that age-dependent methylation changes occur regionally across clusters of CpG sites either stochastically or in a coordinated block-like manner. Image credit: Ochana et al., doi: 10.1016/j.celrep.2025.115958.
“It turns out that the passage of time leaves measurable marks on our DNA,” said Hebrew University’s Professor Tommy Kaplan.
“Our model decodes those marks with astonishing precision.”
“The secret lies in how our DNA changes over time through a process called methylation — the chemical ‘tagging’ of DNA by methyl group (CH3).”
“By zooming in on just two key regions of the human genome, our team was able to read these changes at the level of individual molecules, then use deep learning to translate them into accurate age predictions.”
In the study, Professor Kaplan and colleagues analyzed blood samples from over 300 healthy people, as well as data from a decade-long longitudinal analysis of the Jerusalem Perinatal Study.
The team’s model worked consistently across a range of variables — like smoking, body weight, sex, and even different signs of biological aging.
Beyond potential medical uses, the method could also revolutionize forensic science by allowing experts to estimate a suspect’s age from just a trace of DNA — something existing tools struggle to do.
“This gives us a new window into how aging works at the cellular level,” said Hebrew University’s Professor Yuval Dor.
“It’s a powerful example of what happens when biology meets AI.”
The researchers uncovered new patterns in how DNA changes over time, suggesting our cells encode age both randomly and in coordinated bursts — like ticking biological clocks.
“It’s not just about knowing your age,” said Hebrew University’s Professor Ruth Shemer.
“It’s about understanding how your cells keeps track of time, molecule by molecule.”
“This research could reshape how we approach health, aging, and identity in the future,” the scientists said.
“From helping doctors tailor treatments based on a person’s true biological timeline to giving forensic investigators a powerful new tool for solving crimes, the ability to read age directly from DNA opens the door to breakthroughs across science, medicine, and law.”
“It also deepens our understanding of how aging works — bringing us one step closer to decoding the body’s internal clock.”
_____
Bracha-Lea Ochana et al. Time is encoded by methylation changes at clustered CpG sites. Cell Reports, published online July 14, 2025; doi: 10.1016/j.celrep.2025.115958
