Servicenavigation

Project Details

Back

Projekt Print View

Data-driven models of circadian output regulation in mammals

Subject Area Bioinformatics and Theoretical Biology
Cognitive, Systems and Behavioural Neurobiology
Term from 2020 to 2024
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 444137814
 
Final Report Year 2025

Final Report Abstract

Circadian clock controls many features of our bodies according to the time of day, such as when we are most productive or when we are sleepy. There are two important challenges to studying clocks within our bodies. First, to study the clock and its effects we need to make measurement at different times of the day, called a time series, which if often very challenging. Second, the clock in humans differs from person to person, which is colloquially termed ‘morning larks’ and ‘night owls’. In this project, we developed computational methods to overcome either or both challenges to advance the study of circadian rhythms using high-throughput datasets. To study how the clock functions, we commonly compare the effect of changes in clock rhythms with and without the change. We showed that such comparisons were done incorrectly in the literature and several important insights were missed as a result. This general method works well in animal studies, but in human studies, we often only have measurements under a single condition. So, we developed a method that can identify what is rhythmic and what is not also in this situation, which we used to show that circadian rhythms in septic-shock patients in the ICU are severely disturbed. Both these approaches need time series of measurements, which are cumbersome to obtain in humans. We had therefore developed a way of estimating certain clock properties using a single blood sample, which we extended here to the human skin. This study also allowed us to quantify how different the clock rhythms were from human to human. However, the question remains how we can obtain time series to measure rhythms in internal human organs that also have clocks. To get around this, we used machine learning to virtually sample internal tissues across time. Applying this algorithm to data from clinical human cancer biospies, we described for the first time the presence of rhythms in gene and protein expression in cancers, but these rhythms in the cancers were not in tune with rhythms in healthy tissue. In summary, we gained many new insights into circadian rhythms in humans by means of new computational methods (which we also made available to the public as software packages) that can help realize the goal of using circadian rhythms to improve health and cure disease.

Publications

 
 

Additional Information

© 2026 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung