To the Editor: Bioinformatics software comes in a variety of programming languages and diverse installation methods, which are frequently not straightforward. This heterogeneity makes management of a software stack complicated and error-prone while consuming inordinate amounts of time. Moreover, it inhibits reproducible science \cite{Mesirov2010,Baker2016,Munaf__2017}, because it is hard to recreate a software stack on different machines. While deployment of software has traditionally been handled by administrators, reproducibility of data analyses requires that the researcher is able to maintain full control of the software environment and rapidly modify it without administrative privileges.
The Conda package manager (https://conda.io) has become an increasingly popular approach to overcome these challenges for all major operating systems. Conda normalizes software installations across language ecosystems by describing each software with a human readable recipe that defines meta-information and dependencies, as well as a simple build script that performs the steps necessary to build and install the software. Conda builds software packages within an isolated environment, transforming them into relocatable binaries. Importantly, it obviates reliance on system-wide administation privileges by allowing users to generate isolated software environments, within which software versions can be managed per-project, without generating incompatibilities and side-effects (see Supplementary Results S3). These environments support reproducibility, as they can be rapidly exchanged via files that describe their installation state. Conda is tightly integrated into popular solutions for reproducible data analysis like Galaxy \cite{Afgan2016}, bcbio-nextgen (https://github.com/chapmanb/bcbio-nextgen), and Snakemake \cite{Koster_2012}. To further enhance reproducibility guarantees, Conda can be combined with container or virtual machine based approaches (see Supplementary Results S2) and archive facilities such as Zenodo (see Supplementary Results S3.1). Finally, while Conda provides many commonly-used packages by default, it also allows users to optionally include additional, community-managed, repositories of packages (termed channels).
In order to unlock the benefits of Conda for the life sciences, we present the Bioconda project (https://bioconda.github.io). The Bioconda project provides over 3000 Conda software packages for Linux and macOS. Rapid turnaround times (Supplementary Results S1, Fig. S1e) and extensive documentation (https://bioconda.github.io/contributing.html), have led to a growing community of over 200 international scientists working in the project (Supplementary Results S1, Fig. S1a,b). The project is led by a core team, complemented by interest groups for particular language ecosystems. Unlimited (in time and space) storage for generated packages is donated by Anaconda Inc. All other used infrastructure is free of charge. Bioconda provides packages from various language ecosystems like Python, R (CRAN and Bioconductor), Perl, Haskell, Java,  and C/C++ (Fig. \ref{952607}a). Many of these packages have complex dependency structures that require various manual steps to install when not relying on a package manager like Conda (Supplementary Fig. S1c). With over 6.3 million downloads, Bioconda has become a backbone of bioinformatics infrastructure that is used heavily across all language ecosystems (Fig. \ref{952607}b). Bioconda is complemented by the conda-forge project (https://conda-forge.github.io), which hosts software not specifically related to the biological sciences. This separation has proven to be beneficial, because the focused nature of the Bioconda community allows for fast turnaround times and support when contributing packages or fixing problems. Nevertheless, the two projects collaborate closely, and the Bioconda team maintains over 500 packages hosted by conda-forge. 
Bioconda is not the only effort to distribute bioinformatics software (Fig. \ref{952607}c). The alternatives can be categorized into system-wide (Debian-Med, Genotoo Science, Biolinux, Homebrew) and per user (EasyBuild, GNU Guix, BioBuilds) installation mechanisms. The system-wide approaches lack the ability to put the scientist in control of the installed software stack, thus not solving the requirements for reproducibility outlined above. All per-user based approaches provide a similar feature set (BioBuilds is even also using the Conda package manager). However, among all available approaches, Bioconda is by far the most comprehensive, while being among the youngest (Fig. \ref{952607}c).