Installation

How to install

prot2exon ships as a single package containing both a C++ binary (the index and map commands) and a Python wrapper (the fetch / plot commands, the Mapper API, and the DataFrame helpers). pip and bioconda both give you the whole thing.

Method	Ships the binary	Best for
pip	✅ pre-built wheels (Linux + macOS)	Most users.
bioconda / pixi	✅ built by conda	conda/mamba users; pinned scientific stacks.
build from source	✅ you compile it	Development, custom builds, Windows (via WSL).

pip

pip install prot2exon
# optional extras:
pip install "prot2exon[html]"        # plotly interactive HTML
pip install "prot2exon[all]"         # html + benchmarks + notebooks

The wheel bundles the compiled binary (as prot2exon/_bin/prot2exon-core), so all four commands work immediately — nothing else to install:

prot2exon --version
prot2exon fetch list

Pre-built py3-none wheels are published for Linux (manylinux x86_64) and macOS (Intel + Apple Silicon), and work on any Python 3.9+. On platforms without a wheel — notably Windows — pip falls back to the source distribution, which compiles the C++ on install and therefore needs a C++17 toolchain + CMake (the smoothest Windows path is WSL or conda). To point the wrapper at a binary other than the bundled one, set PROT2EXON_BIN:

export PROT2EXON_BIN=/path/to/prot2exon-core

Discovery order: $PROT2EXON_BIN, then the wheel-bundled _bin/prot2exon-core, then ./build/prot2exon in a source checkout, then prot2exon-core on $PATH.

bioconda

conda install -c bioconda -c conda-forge prot2exon
# or, faster:
mamba install -c bioconda -c conda-forge prot2exon

conda compiles the binary as part of the recipe, so all four commands land on your PATH just like the pip install.

pixi

pixi resolves from the same conda channels:

pixi add -c bioconda -c conda-forge prot2exon      # in a project
pixi global install -c bioconda -c conda-forge prot2exon   # as a global tool

Build from source

Requirements:

• C++17 toolchain (g++ ≥ 9, clang ≥ 10, or MSVC ≥ 2019)
• CMake ≥ 3.16
• OpenMP (optional — enables --threads; the binary still runs single-threaded without it)

git clone https://github.com/SotoLF/Prot2Exon.git
cd Prot2Exon
mkdir build && cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
make -j$(nproc)

The binary lands at build/prot2exon. Then install the Python wrapper editable so fetch/plot and the API are available too:

cd ..
pip install -e .

The repo's bin/prot2exon wrapper finds build/prot2exon automatically, so you can run the four commands straight from the checkout.

Docker

A Dockerfile at the repo root bundles both the binary and the wrapper:

docker build -t prot2exon .
docker run --rm -v "$(pwd):/work" prot2exon \
    map --index /work/human.idx --bed /work/queries.bed \
        --out-dir /work/out --output all

Smoke test

prot2exon --version             # C++ binary, prints the index format version
prot2exon plot --help           # Python plotter
prot2exon fetch list            # pre-built indexes + GTF-build presets

Then run the end-to-end test suite from a source checkout:

python3 tests/run_tests.py      # expects "N passed, 0 failed"

Once installed, head to Building an index. Tuning flags for large runs live on Performance and Benchmarking.