Installation

Installation

Build from source

Requirements:

• C++17 toolchain (g++ ≥ 9, clang ≥ 10, or MSVC ≥ 2019)
• CMake ≥ 3.16
• OpenMP (optional — parallelises per-query processing; the binary still works without it, just single-threaded)

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

You'll find the binary at build/prot2exon.

Python wrapper

The Python package shells out to the C++ binary. Install with the wrapper requirements:

pip install -e python/         # editable install from the repo
# or, if you build a wheel: pip install dist/prot2exon-*.whl

The wrapper auto-discovers the binary via, in order:

1. PROT2EXON_BIN environment variable
2. ./build/prot2exon relative to the repo root
3. prot2exon on $PATH

To pin a specific binary:

export PROT2EXON_BIN=/path/to/build/prot2exon

Docker

A Dockerfile is shipped at the repo root. Build:

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

The image bundles both the C++ binary and the Python wrapper.

Smoke test

./build/prot2exon --version
python3 -m prot2exon.plot --help     # if the wrapper is installed

Then run the end-to-end test suite:

python3 tests/run_tests.py

You should see 109 passed, 2 failed (the two failures need matplotlib + pandas in system python and are not regressions).

See Performance and RAM for tuning flags once you're up and running.