TimeTeller: A tool to probe the circadian clock as a multigene dynamical system

Area of Science:

• Chronobiology
• Systems Biology
• Machine Learning

Background:

• The molecular circadian clock regulates critical physiological processes and influences disease pathogenesis, including cancer and heart disease.
• Accurate assessment of circadian clock function is crucial for understanding disease and developing effective therapies.
• Existing tools are limited in analyzing the clock as a complex, noisy, multi-gene dynamical system.

Purpose of the Study:

• To introduce TimeTeller, a novel machine learning tool designed to analyze the circadian clock as a noisy multigene dynamical system.
• To estimate circadian clock function from single transcriptome data by modeling its multi-dimensional state.
• To demonstrate TimeTeller's utility in assessing clock systems across different species and data types.

Main Methods:

• Development of TimeTeller, a machine learning algorithm for analyzing noisy multigene dynamical systems.
• Application of TimeTeller to mouse, baboon, and human microarray and RNA-seq data.
• Utilizing TimeTeller to visualize and quantify clock structure, stratify samples by clock dysfunction, and compare clocks across diverse contexts.

Main Results:

• TimeTeller successfully models the multi-dimensional state of the circadian clock from transcriptomic data.
• The tool enables visualization and quantification of global clock structure.
• TimeTeller can quantitatively stratify samples based on circadian clock dysfunction and compare clocks across individuals, conditions, and tissues.

Conclusions:

• TimeTeller provides a novel methodology for analyzing the molecular circadian clock as a noisy dynamical system.
• This tool has significant potential for advancing circadian medicine by enabling precise assessment of clock function.
• TimeTeller facilitates a deeper understanding of circadian clock dynamics in health and disease across species.