Many Epigenetic Clocks Differ by Time of Day
Because epigenetic clocks are produced from DNA methylation data via machine learning approaches, correlating patterns of change with chronological age, it remains unclear as to what exactly they measure. Which processes of aging produce the specific DNA methylation changes used in any given clock? As yet that question has no answer. Thus a discovery process continues, in which researchers uncover clock behaviors such as a dependency on aspects of the circadian rhythm. Determining exactly which aspects will be one small part of a much longer process of understanding the details of the relationship between DNA methylation and the rest of cellular biochemistry. For now it is a caveat for those using aging clocks, epigenetic or otherwise - either pick a clock demonstrated to lack this behavior, or be consistent in time of day when measuring.
Since their introduction, epigenetic clocks have been extensively used in aging, human disease, and rejuvenation studies. In this article, we report an intriguing pattern: epigenetic age predictions display a 24-hour periodicity. We tested a circadian blood sample collection using 17 epigenetic clocks addressing different aspects of aging. Thirteen clocks exhibited significant oscillations with the youngest and oldest age estimates around midnight and noon, respectively. In addition, daily oscillations were consistent with the changes of epigenetic age across different times of day observed in an independant populational dataset.
While these oscillations can in part be attributed to variations in white blood cell type composition, cell count correction methods might not fully resolve the issue. Furthermore, some epigenetic clocks exhibited 24-hour periodicity even in the purified fraction of neutrophils pointing at plausible contributions of intracellular epigenomic oscillations. Evidence for circadian variation in epigenetic clocks emphasizes the importance of the time-of-day for obtaining accurate estimates of epigenetic age.