paper.md
In travisbyrum/disgraph: Graph Distance Utility Package

title: 'Inferring, comparing and exploring ecological networks from time-series data through R packages constructnet, disgraph and dynet' tags: - Time-series networks - network construction - graph distance - network dynamics - ecological networks authors: - name: Anshuman Swain orcid: 0000-0002-9180-2222 affiliation: 1 - name: Travis Byrum affiliation: 1 - name: Zhaoyi Zhuang affiliation: 1 - name: Luke Perry affiliation: 1 - name: Michael Lin affiliation: 1 - name: William F Fagan affiliation: 1 affiliations: - name: Department of Biology, University of Maryland, College Park index: 1 date: 03 May 2021 bibliography: paper.bib

Statement of need

Network inference is a major field of interest for the ecological community, especially in light of the high cost and difficulty of manual observation, and easy availability of remote, long term monitoring data. In addition, comparing across similar network structures, especially with spatial, environmental, or temporal variability and, simulating processes on networks to create toy models and hypotheses – are topics of considerable interest to the researchers.

A large number of methods are being developed in the network science community to achieve these objectives but either don’t have their code available or an implementation in R, the language preferred by ecologists and other biologists.

We provide a suite of three packages which will provide a central suite of standardized network inference methods from time-series data (constructnet), distance metrics (disgraph) and (process) simulation models (dynet) to the growing R network analysis environment and would help ecologists and biologists to perform and compare methods under one roof.

These packages are implemented in a coherent, consistent framework – making comparisons across methods and metrics easier. We hope that these tools in R will help increase the accessibility of network tools to ecologists and other biologists, who the language for most of their analysis.

Introduction

The usage of networks in ecology and allied fields has increased considerably in the past two decades, owing to the explosion of new tools and theoretical findings in network science [@Poisot:2016]. The entities represented by ecological networks can be quite varied – they can be animal social interaction networks (Croft et al., 2008; Krause et al., 2015), species interaction networks (trophic networks, mutualistic networks, parasitic networks, antagonistic networks) (Dunne et al., 2002; Pascual and Dunne, 2006; Kefi et al., 2012; Bascompte and Jordano, 2013; Shaw et al., 2021), or ecosystem flux models (Proulx et al., 2005). The study and exploration of these networks have influenced our understanding of communities and ecosystems in a tremendous way, allowing for a holistic representation of ecological systems (see reviews: Delmas et al., 2019; Guimarães, 2020).

Acknowledgements

We would like to thank Morelle Tchuindjo and Nathan Stiff for their contributions to the project. We acknowledge the role of netrd package in Python in providing a baseline for our R development at all times. AS thanks National Science Foundation award DGE-1632976 for training and support.