Installation¶

Requirements¶

Python 3.12–3.14
PyTorch 2.0+ (with MPS/CUDA support recommended)

Install from Source¶

git clone <repo-url> DNANet-clean
cd DNANet-clean

# Create virtual environment
python -m venv .venv
source .venv/bin/activate

# Install with development dependencies
pip install -e ".[dev]"

Verify Installation¶

# Check CLI is available
dnanet --help

# Run the test suite
pytest tests/ -q

Data Setup¶

DNANet expects forensic HID data to be placed under data/ relative to the project root. Two datasets are supported out of the box:

NFI R&D Dataset (2p/5p)¶

Place the dataset so the directory structure looks like:

data/
  2p_5p_Dataset_NFI/
    Raw data .HID files/
      Mixture dataset 1/...
      Mixture dataset 2/...
      ...
    txt_annotations_2024/
      Dataset 1 DTH_AlleleReport.txt
      ...
    2p_5p_hid_to_annotation.csv
    best_ladder_paths_DTH.csv
    ladder_alleles.csv

PROVEDIt Dataset¶

data/
  PROVEDIt/
    5 sec/
      injection_dir_1/
        *.hid
      ...
    PROVEDIt_RD14-0003 GF Known Genotypes.xlsx

Optional Dependencies¶

Package	Purpose
`mlflow`	Experiment tracking (default logger)
`tensorboard`	Alternative experiment tracking
`openpyxl`	Reading ProvedIt XLSX genotype files
`mkdocs`	Building documentation with Zensical