Install bwa with Homebrew, apt, dnf, Nix

Project links

• Homepage
• Upstream docs
• Package manager page

Commands and aliases

• bwa

Project history

BWA was created by Heng Li and documented through a series of algorithm publications and release notes. The current README describes three major algorithms in the package: BWA-backtrack for short Illumina reads, BWA-SW for longer sequences, and BWA-MEM as the later generally recommended algorithm for many reads longer than about 70 bp.

The README points users to the GitHub source repository and SourceForge release downloads, while NEWS.md records the long-running 0.7.x series. Release notes show important maintenance milestones such as GRCh38 ALT-contig support and bwakit binaries in 0.7.11, ARM64 support in 0.7.18, and the 0.7.19 bug-fix release in 2025.

Adoption history

BWA achieved unusually broad adoption in computational biology rather than only in general-purpose Unix packaging. The official README advertises SourceForge downloads, GitHub downloads, and BioConda install counts, and the supplied package facts show Homebrew, Debian, Ubuntu, Fedora, and Nix packaging.

Its adoption is also bibliographic: the README instructs users to cite the 2009 BWA-backtrack Bioinformatics paper, the 2010 BWA-SW Bioinformatics paper, and the 2013 BWA-MEM arXiv preprint depending on which algorithm they use.

How it is used

A typical BWA workflow builds an FM-index with `bwa index`, then maps reads with subcommands such as `bwa mem`, `bwa aln` plus `samse` or `sampe`, or `bwa bwasw`. Output is usually SAM streamed into downstream tools such as samtools.

The README now notes that BWA-MEM has been superseded for many newer workloads by minimap2 for long reads and is expected to be superseded by minibwa for short reads, while BWA-MEM2 is suggested when exact legacy BWA-MEM behavior is desired.

Why package nerds care

BWA is package-nerd significant because a single small C executable sits at the center of many genomics workflows, yet its behavior is scientifically consequential. Reproducibility often depends on the exact BWA version, algorithm, reference preparation, and SAM output details.

For maintainers, BWA is also a case study in domain tooling that migrated from SourceForge-era distribution to GitHub while remaining heavily consumed through scientific package managers such as BioConda and through operating-system packages.

Timeline

• 2009: BWA-backtrack paper is published for fast short-read alignment with the Burrows-Wheeler transform.
• 2010: BWA-SW paper is published for longer reads.
• 2013: BWA-MEM preprint is published and becomes the recommended algorithm for many common workloads.
• 2014: BWA 0.7.11 adds GRCh38 ALT-contig support and introduces bwakit binaries.
• 2017: BWA 0.7.17 is released.
• 2024: BWA 0.7.18 adds ARM64 support.
• 2025: BWA 0.7.19 is released as a bug-fix release.

Related projects

• minimap2 is the successor recommended by the BWA README for long reads.
• BWA-MEM2 is recommended when users want faster execution while preserving legacy BWA-MEM output.
• minibwa is named by the README as the expected short-read replacement.
• bwakit packages BWA with associated tools and reference-building workflows for some use cases.

Sources

• https://github.com/lh3/bwa
• https://github.com/lh3/bwa/blob/master/NEWS.md
• source_facts.package-manager