Improved accuracy in measuring biological age

Biomarkers of aging can be derived from cellular, molecular, functional and physiological measurements and are used to study biological aging in humans. Many studies use so-called “epigenetic clocks” based on DNA methylation data to analyze relationships between biological aging, morbidity, and mortality. Although these clocks are currently considered the best predictors of biological age, they still include technical noise, which leads to significant variance in the measurements.

To make the watches more useful for longitudinal analyzes and clinical studies, it is important to improve the accuracy of the measurements. Here we present a method where principal component analysis (PCA) is used to separate noise from age-related signals. In this way, only biologically relevant signals are prioritized and the reliability of computer clocks is much higher compared to the original clocks.

The study was led by a research team at Yale University School of Medicine. Associate Professor Albert Higgins-Chen, Ph.D., and Morgan Levine, Ph.D., who recently moved to the newly started Altos Labs in San Diego, USA. Dr. Higgins-Chen, first author of the study says: “We’ve observed some epigenetic clocks where you are biologically 50 years old in one test, then 59 years old in the next. We started with the current, commonly used clocks, and found that one change – adding a technique called principal component analysis – could preserve aging signals while reducing noise. Significantly. Now, for most tests, two measurements taken at the same time will differ by less than a year. This also means that a person’s longitudinal change in epigenetic age is more likely to reflect true biological aging rather than technical hype.”