Irina Conboy proposes measuring failures in gene regulation; Matt Kaeberlein calls for testing whether the measure predicts health
Irina Conboy proposes measuring failures in gene regulation; Matt Kaeberlein calls for testing whether the measure predicts health
On July 29, biogerontologist Matt Kaeberlein discussed with Irina Conboy what a biological age test should measure. Conboy proposes looking for a loss of precision in gene activity. Kaeberlein argues that researchers must determine whether the signal is reproducible within the same person and whether it is associated with subsequent health outcomes.
Epigenetic clocks use methylation, which consists of chemical marks on DNA, to calculate age from the pattern of those marks. The calculation produces a single number indicating whether the body appears younger or older than its chronological age. Conboy proposes using the same data differently by measuring variation in methylation at DNA sites where the mean level remains stable with age.
When applied to the same blood samples, eight epigenetic clocks differed by an average of 17 years. The choice of model changes the resulting age estimate. Conboy wants to use the same data to identify a specific failure of gene regulation.
In her conversation with Kaeberlein, she linked this variation to a loss of regulation. In her approach, the mean methylation level at selected DNA sites remains unchanged, while variation among individuals increases with age. Conboy calls this variation epigenetic noise.
Her group described the approach in a 2023 paper. The authors selected CpG sites at which the mean methylation level remains stable with age, then summed the variation in methylation values. A CpG site is a short region of DNA where a methyl group can attach. Conboy proposes combining these signals into a profile that shows which physiological systems exhibit this variation and which measures should be monitored when evaluating an intervention.
Kaeberlein wants this signal to become a testable measure. Two samples collected from the same person at the same time should produce the same result. Researchers must then compare the measure with subsequent organ function, disease risk, or response to treatment. He therefore proposed evaluating the Generation Lab test using samples collected in parallel.
In a study of 15995 samples from 17 tissues, the authors separately observed age-related shifts in mean methylation, increasing differences among individuals, and molecular disorder. These signals were distributed differently across tissues. A single age estimate compresses them into one overall measure, while a profile shows which type of divergence the sample detected.
Conboy and Kaeberlein are addressing two different questions. Conboy links epigenetic noise to failures in cellular regulation. Kaeberlein asks whether the signal is reproducible in parallel samples and whether it predicts a person’s health. The answer will determine whether the measure remains a description of a biological process or becomes a criterion for evaluating treatment.