Kronos: circadian rhythmicity analysis in microbiome and other omics datasets

Circadian rhythms, or daily biological cycles, are key to maintaining health and involved in multiple physiological and behavioural processes, including coordination of metabolic homeostasis, nutrition, physical activity, and sleep. In recent years, it has not only been shown that the microbiome displays diurnal oscillations, but that these oscillations impact host physiology, specifically through the modification of gut-brain communication among other systems. Interest in circadian rhythm research is expanding as our urban environments have increased exposure to factors that can disrupt the normal physiological rhythm of our body, such as delayed bedtimes, shift work, jet-lag, increased screen-time, and exposure to artificial light. Discovering how oscillatory signals respond to both external and internal factors can lead to important biological breakthroughs, but assessing rhythmicity can be both limited by the complexity of statistical models and demanding in terms of coding and statistical expertise.

Due to the rhythmic nature of microbiome data, bespoke analysis pipelines and subsequent statistical testing are required to understand challenging facets of host-microbe interactions. Some methods and frameworks exist, but they are either limited to simple two-group designs or demand an unrealistically extensive theoretical and coding background from their users.

Here, we describe the novel easy-to-use, free and open-source R-based tool, Kronos, to assess circadian rhythms. We benchmarked Kronos against three other commonly-used rhythmicity analysis tools; Cosinor2, JTK Cycle and LimoRhyde. We found that Kronos is a suitable method for assessing rhythmicity that performs similarly or better compared to its peers and does not sacrifice accuracy or reliability. Further, we provide the user with a detailed step-by-step tutorial to ensure easy adoption.