Project build with uv, a modern python package manager (https://docs.astral.sh/uv/).

Installing uv (or follow the installation setup):

curl -LsSf https://astral.sh/uv/install.sh | sh

git clone
uv install
uv run python main.py

Check cli args:

uv run python main.py --help

Datasets are expected to be under ./datasets. Model states are saved under ./saved_models, and figures under ./figures.

2vs3 classes

The master branch classifies TP and FP. The branch three-classes integrates FN as well, as the class of type C for the xgboost to predict.

The inclusion of FN requires ground truth knowledge of the mutations, and can't be used post-vcf, since there isn't vcf of FN in the first place.