Illustration of a clock showing one side dark and one light to indicate circadian rhythms and the human body clock with clouds in the background. A man is hanging onto the clock hand.
Credit: Natalia Misintseva/Getty Images

Research carried out by the University of Surrey and the University of Groningen shows combining machine learning and metabolite analysis can help predict a person’s individual circadian body clock and assess what healthy sleep patterns are for that person.

“The circadian system in humans influences many behavioral, physiological, and molecular processes in the body, causing considerable variation in body function and cellular constitution over the course of the 24-hour day,” wrote lead investigator Roelof Hut, a professor at the University of Groningen, and colleagues in the journal PNAS.

“Chronomedicine seeks to exploit this 24-hour variation to optimize timing for pharmaceutical application (‘clocking the drug’) or diagnostic sampling for treatment of circadian clock abnormalities (‘drugging the clock’). Taking such ‘body time’ approaches into clinical practice is thought to reduce medication load, improve treatment outcome, and increase accuracy and specificity of diagnosis.”

Many factors, including genetics, lifestyle and location around the world, influence a person’s circadian rhythm. The current gold standard technique for measuring the human body clock is measuring pineal melatonin synthesis under controlled light conditions.

This method is accurate, but has long sampling periods of 8–20 hours and requires specific posture and lighting settings. “Such a body time estimation method is not practical in real-world settings and therefore there is a strong need for validated alternative estimates of human body time,” write the authors.

In this study, Hut and colleagues combined the power of metabolomics with machine learning to accurately predict a person’s body time. They collected blood samples from 12 men and 12 women seven days before attending a clinic session and then in the clinic. The participants were healthy, did not smoke and were asked to abide by regular sleeping schedules for at least a week before they visited the study research facility to undergo a sleep test and further sampling.

The research team measured more than 130 metabolites in the blood samples and used machine learning to estimate body time. “The underlying concept here is that many transcripts exhibit daily patterns that vary widely in phase; as such, it should be possible to estimate phase from a single sample by comparing the relative values of these transcripts when they are treated as ‘features’ in machine learning and other analytical approaches,” explain the authors.

They found that they could estimate body time very accurately (median error of 0.45–0.60 hour) using two optimally timed blood samples (six hours apart for women and 12 hours apart for men) and the machine learning algorithm. They report that this result is more accurate than other molecular approaches to estimate body time.

“We are excited but cautious about our new approach to predicting DLMO—as it is more convenient and requires less sampling than the tools currently available. While our approach needs to be validated in different populations, it could pave the way to optimize treatments for circadian rhythm sleep disorders and injury recovery,” said co-author Debra Skene, a professor at the University of Surrey, in a press statement.

“Smart devices and wearables offer helpful guidance on sleep patterns—but our research opens the way to truly personalized sleep and meal plans, aligned to our personal biology, with the potential to optimize health and reduce the risks of serious illness associated with poor sleep and mistimed eating.”

Also of Interest