Installation

salmon 2.0 is a single self-contained binary. Pick whichever method fits your environment.

The release install script downloads the right prebuilt binary for your OS/architecture and puts it on your PATH:

curl --proto '=https' --tlsv1.2 -LsSf \
  https://github.com/COMBINE-lab/salmon/releases/latest/download/salmon-cli-installer.sh | sh

Prebuilt binaries are provided for Linux and macOS on both x86-64 and aarch64. The x86-64 build targets a broad SIMD floor (SSE4.2-era) and selects AVX2 at runtime where available, so one binary runs on essentially any modern CPU.

If you have a Rust toolchain (≥ 1.91):

cargo install salmon-cli

This builds and installs the salmon binary from crates.io.

conda install -c bioconda -c conda-forge salmon

The salmon package tracks the 2.0 (Rust) line. The final C++ release remains available as salmon-cpp for reproducibility.

Multi-arch images (linux/amd64 + linux/arm64) are published on every release to Docker Hub and GHCR:

docker pull combinelab/salmon:latest          # or pin a version: :2.2.1
# mirror on GHCR:
docker pull ghcr.io/combine-lab/salmon:latest

The image’s entrypoint is salmon, so arguments go straight to it (don’t repeat salmon):

docker run --rm combinelab/salmon:latest --version

Mount your data and run as usual — the first argument is the salmon subcommand:

# build an index
docker run --rm -v "$PWD:/data" -w /data combinelab/salmon:latest \
  index -t transcripts.fa -i salmon_index

# quantify
docker run --rm -v "$PWD:/data" -w /data combinelab/salmon:latest \
  quant -i salmon_index -l A -1 r1.fq.gz -2 r2.fq.gz -p 8 -o out

Tags: latest (newest release), X.Y (e.g. 2.2), and exact X.Y.Z (e.g. 2.2.1) — pin X.Y.Z for reproducible pipelines.

git clone https://github.com/COMBINE-lab/salmon
cd salmon
cargo build --release
# binary at target/release/salmon

The repository ships a .cargo/config.toml that pins the portable SIMD floor; release builds use it automatically.

Verify

salmon --version
salmon --help