More and more evidence suggest that circadian clock disruption or misalignment is a feature of many diverse chronic diseases including metabolic syndrome, depression but also a number of cancers. For the latter, recent mechanistic studies in cancer models have established an understanding of how the circadian clock influences onset, progression and therapeutic outcomes. Moreover, it has been proposed that tumours might have disrupted circadian oscillators. In patients, however, this is more difficult to establish as usually only single samples, e.g., tumour biopsies, are available. Therefore, novel tools to measure the functional state of the molecular circadian clock are needed.

Here, we introduce TimeTeller, a machine learning tool that analyses the clock as a system and aims to estimate circadian clock function from a single sample’s transcriptome by modelling the multi-dimensional state of the clock. We demonstrate TimeTeller’s utility for analysing experimental in vitro and in vivo, as well as healthy human and patient samples from various platforms (microarray, RNA-Seq and NanoString) and highlight TimeTeller’s potential relevance for advancing circadian medicine. The project is an inter-disciplinary collaboration including significant work by Warwick’s MRCDTP students Laura Usselmann and Vadim Vasilyev and is setting the stage for further applications of TimeTeller in experimental models and human breast tumours.
Read the paper hereLink opens in a new window.