MetaCycle: an integrated R package to evaluate periodicity in large scale data

Area of Science:

• Computational Biology
• Bioinformatics
• Data Science

Background:

• Identifying periodic signals in large-scale time-series data is a significant computational challenge.
• Existing methods often focus on individual algorithms, lacking a generalized approach for robust periodicity detection.

Purpose of the Study:

• To introduce MetaCycle, an R package designed for the generalized detection of periodicity in time-series data.
• To integrate multiple established periodicity detection algorithms into a single, user-friendly package.

Main Methods:

• MetaCycle incorporates three established algorithms: ARSER, JTK_CYCLE, and Lomb-Scargle.
• The package provides two main functions, meta2d and meta3d, for analyzing two- and three-dimensional time-series datasets, respectively.
• Meta2d utilizes N-version programming to combine results from multiple algorithms for enhanced reliability.

Main Results:

• MetaCycle offers a convenient and integrated approach to evaluating periodicity in time-series data.
• The package facilitates the analysis of both 2D and 3D time-series datasets.
• Integration of multiple algorithms aims to improve the accuracy and robustness of periodicity detection.

Conclusions:

• MetaCycle provides a valuable tool for researchers needing to detect periodicity in complex time-series datasets.
• The package simplifies the application of advanced periodicity detection methods.
• MetaCycle enhances the ability to find biological rhythms and other periodic phenomena in large datasets.