DNAGAST

The official code repository for DNAGAST: Generative Adversarial Set Transformers for High-throughput Sequencing.

Dependencies

• LMDBM (modified)
• Tensorflow
• tf-utils
• tf-set-transformer
• Weights & Biases

Training & Evaluation

Each model architecture can be trained and evaluated by invoking the appropriate script located in the scripts/ directory. These scripts integrate the Weights & Biases platform directly for version control, thus W&B must be configured appropriately on your system before execution.

Citation

@article{ludwig_dnagast_2025,
    title={DNAGAST: Generative Adversarial Set Transformers for High-throughput Sequencing},
    author={Ludwig II, David W. and Phillips, Joshua L.},
    journal={The International FLAIRS Conference Proceedings},
    volume={38},
    issn = {2334-0762},
    year={2025},
    month={May}
}