Monitoring the circadian clock in human blood using personalized machine learning

Abstract

The circadian clock and the rhythms it produces are crucial for human health, but frequently perturbed by the modern environment. At the same time, circadian rhythms may influence the efficacy and toxicity of therapeutics and the metabolic response to food intake. Measuring the bodys response to treatments for circadian dysfunction, as well as optimizing the daily timing of treatments for other health conditions, requires a simple and accurate method for monitoring the circadian clock. Here we used a recently developed method called ZeitZeiger to predict circadian time (CT, time of day according to the circadian clock) from genome-wide gene expression in human blood. In cross-validation on 498 samples from 60 individuals across three publicly available datasets, ZeitZeiger predicted CT in single samples with a median absolute error of 2.1 h. The predictor trained on all 498 samples used 15 genes, only two of which are part of the core circadian clock. We then extended ZeitZeiger to make predictions for groups of samples, and developed a general framework to personalize predictions using samples from only the respective individual. Each of these strategies improved prediction of CT by ~20%. Our results are an important step towards precision circadian medicine.