In a paper appearing in the Proceedings of the National Academy of Sciences, researchers at the University of Manchester and other centers in the UK turn to blood plasma metabolomics and machine learning to predict a circadian rhythm measurement known as "dim light melatonin onset" (DLMO) in dozens of male and female participants. Using a partial least squares regression machine learning method, the team attempted to estimate DLMO based on targeted metabolomic profiling of 131 compounds in the blood at sampling points every two hours. In parallel, participants' hormone levels were measured hourly. The results suggest the targeted metabolomics machine learning method compared favorably to existing circadian rhythm estimates when tweaked to take sex-specific metabolites into account. More than half of metabolites used for DLMO prediction were informative in both male and female participants, while around one-third were male-specific and roughly 15 percent were female-specific. "Although estimation of circadian body time using blood-targeted metabolomics requires further validation in shift work and other real-world conditions," the authors write, "it currently may offer a robust, feasible technique with relatively high accuracy to aid personalized optimization of behavior and clinical treatment after appropriate validation in patient populations."