bedtoolsr: Bedtools Wrapper

title: 'Bedtoolsr: An R package for genomic data analysis and manipulation' tags: - R - Python - bedtools - genomics - data analysis - metaprogramming authors: - name: Mayura N. Patwardhan affiliation: 1 - name: Craig D. Wenger affiliation: 2 orcid: 0000-0002-7361-8456 - name: Eric S. Davis affiliation: 3 orcid: 0000-0003-4051-3217 - name: Douglas H. Phanstiel affiliation: "1, 3, 4, 5" # (Multiple affiliations must be quoted) orcid: 0000-0003-2123-0051 affiliations: - name: Thurston Arthritis Research Center, University of North Carolina, Chapel Hill, NC 27599, USA index: 1 - name: Independent Scholar, Long Beach Township, NJ 08008 index: 2 - name: Curriculum in Bioinformatics & Computational Biology, University of North Carolina, Chapel Hill, NC 27599, USA index: 3 - name: Department of Cell Biology & Physiology, University of North Carolina, Chapel Hill, NC 27599, USA index: 4 - name: Lineberger Comprehensive Cancer Research Center, University of North Carolina, Chapel Hill, NC 27599, USA index: 5 date: "10 September 2019" bibliography: paper.bib

Summary

The sequencing of the human genome and subsequent advances in DNA sequencing technology have created a need for computational tools to analyze and manipulate genomic data sets. The bedtools software suite and the R programming language have emerged as indispensable tools for this purpose but have lacked integration. Here we describe bedtoolsr, an R package that provides simple and intuitive access to all bedtools functions from within the R programming environment. We provide several usability enhancements, support compatibility with past and future versions of bedtools, and include unit tests to ensure stability. bedtoolsr provides a user-focused, harmonious integration of the bedtools software suite with the R programming language that should be of great use to the genomics community.

Source Code: https://github.com/PhanstielLab/bedtoolsr. Information: http://phanstiel-lab.med.unc.edu/bedtoolsr.html. Contact: douglas_phanstiel@med.unc.edu

Introduction

The sequencing of the human genome and subsequent advances in DNA sequencing technology have transformed modern biological research by producing data sets of ever-increasing size and complexity. While these technologies have led to breakthroughs in genetics research, the incredible throughput and breadth of the resulting data have spurred a reliance on computational tools and programming languages to interpret the results. In 2010, Quinlan et al. developed bedtools, a powerful suite of command-line tools for ‘genome arithmetic’ that has become one of the most widely used and indispensable tools for genomic data analysis [@Quinlan2010-ak]. A year later, pybedtools extended the features of bedtools to the python programming language [@Dale2011-ja]. During that same time period, the use of the programming language R—with a rich trove of libraries for statistical analysis and data visualization—skyrocketed in the biological sciences [@Tippmann2015-yz]. While several R packages have been developed for bedtools-like genome analysis, their usage and functionality differ significantly from that of bedtools [@Riemondy2017-vz; @Lawrence2013-px]. These differences make them more difficult to use for those who are already familiar with bedtools behavior and lacks some of the capabilities of bedtools.

Here we describe bedtoolsr, an R package that allows seamless integration of bedtools functions into the R programming environment. bedtoolsr functionality, inputs, outputs, and documentation perfectly replicate those found in the command-line version of bedtools and offer new features for improved usability within the R environment.

Methods

bedtoolsr is an R package that allows access to all bedtools functions from within the R environment. To support past, current, and future versions of bedtools, we wrote bedtoolsr using a metaprogramming approach. The bedtoolsr package is built by a python script that reads function names, parameters, default settings, and documentation from a local installation of bedtools and constructs a distributable R package custom-built for that bedtools version. bedtoolsr is version controlled and freely available on the software development platform GitHub. To ensure stability, bedtoolsr supports continuous integration and includes unit tests for every function. These unit tests were implemented using the R package testthat [@Wickham2011-or] and can be run immediately after installation to ensure proper functionality. The continuous integration service Travis CI runs every time a code change is posted to GitHub to safeguard against any updates that might introduce flaws, faults, or failures.

Results

bedtoolsr was written with user experience in mind. To minimize the learning curve for those already familiar with bedtools we aimed to perfectly replicate the bedtools experience while adding all of the features of an R package. As such, bedtoolsr supports every currently available bedtools function and all function parameters. Parameters have the exact same names and documentation as those provided by bedtools with code autocompletion support from within RStudio.

bedtoolsr extends bedtools features for improved ease of use. Inputs for bedtoolsr functions can be provided either as file paths or as R objects (e.g. data.frames, data.tables, tibbles, etc). bedtoolsr automatically detects whether the input is a file path or R object and handles the data accordingly. To simplify usage, bedtoolsr comes preloaded with chromosome size files for commonly used genomes that are required by many bedtools functions. Results can be returned as a data frame or written directly to a file. To ensure proper installation of the package, users can run unit tests for most functions which can be executed with a single command following installation.

Discussion

bedtoolsr provides seamless integration of the bedtools software suite into the R programming environment. The package was designed to be as user-friendly as possible and should be intuitive for those already familiar with the bedtools command-line tool. The ability to handle multiple data types, the forward and backward compatibility, and the included unit tests ensure stability and ease of use. The harmonious combination of these two powerful analytical platforms should make bedtoolsr a valuable and widely used tool for genomic analysis.

Acknowledgements

We thank Erika Deoudes for her contributions to the website and logo design and Mike Love for his helpful suggestions.

Funding

D.H.P. is supported by the National Institutes of Health (NIH), National Human Genome Research Institute (NHGRI) grant R00HG008662 and National Institutes of Health (NIH), National Institute of General Medical Sciences (NIGMS) grant R35GM128645. E.S.D. was supported in part by a grant from the National Institute of General Medical Sciences under award 5T32 GM067553