FMM: An R Package for Modeling Rhythmic Patterns in Oscillatory Systems

TL;DR

The paper introduces the FMM R package, which models diverse rhythmic patterns in oscillatory signals using a flexible, parametric approach suitable for biological data analysis.

Contribution

It presents a novel, flexible modeling framework and an R package for analyzing rhythmic oscillations, including functions for fitting, exploring, and generating synthetic data.

Findings

Successfully models various biological oscillations

Provides tools for fitting and visualizing rhythmic patterns

Demonstrates applicability to circadian, cardiac, and neuronal data

Abstract

This paper is dedicated to the R package FMM which implements a novel approach to describe rhythmic patterns in oscillatory signals. The frequency modulated M\"obius (FMM) model is defined as a parametric signal plus a gaussian noise, where the signal can be described as a single or a sum of waves. The FMM approach is flexible enough to describe a great variety of rhythmic patterns. The FMM package includes all required functions to fit and explore single and multi-wave FMM models, as well as a restricted version that allows equality constraints between parameters representing a priori knowledge about the shape to be included. Moreover, the FMM package can generate synthetic data and visualize the results of the fitting process. The potential of this methodology is illustrated with examples of such biological oscillations as the circadian rhythm in gene expression, the electrical activity of the heartbeat and neuronal activity.