🧬 BioReason-Pro
Advancing Protein Function Prediction with
Multimodal Biological Reasoning

BioReason-Pro Protein Structures

AlphaFold backbone structures for the proteins in the BioReason-Pro training and evaluation sets. BioReason-Pro feeds these coordinates to its ESM3 encoder alongside the amino-acid sequence; ESM3 falls back to sequence-only when a structure is absent, so these are optional — but the released checkpoint was trained with them.

Covers 131,838 structures — every structure_path referenced by bioreason-pro-sft-reasoning-data, bioreason-pro-test-data, and the full 133,492-protein superset.

Usage

Do not download this repository by hand. The helper script in the code repo reconstructs the exact directory layout the training and evaluation scripts expect:

git clone https://github.com/bowang-lab/BioReason-Pro.git && cd BioReason-Pro
pip install -e .

python scripts/download_assets.py --dest /data/bioreason
# -> /data/bioreason/structures       (pass as STRUCTURE_DIR)
# -> /data/bioreason/go_embeddings

python scripts/download_assets.py --dest /data/bioreason --verify

The download is resumable and parallel; ~34 GB over the wire, ~60 GB on disk.

Layout

Each shard is a .tar.gz of gzipped mmCIF files (AF-<accession>-F1-model_v4.cif.gz).

Note: the datasets reference the decompressed filename (AF-<accession>-F1-model_v4.cif). download_assets.py gunzips on extraction. If you unpack the shards yourself and leave the .gz extension in place, every structure lookup fails silently — training falls back to empty coordinates with no error. Run --verify to confirm coverage.

Source

Structures are from the AlphaFold Protein Structure Database (Jumper et al. 2021, Varadi et al. 2024), redistributed under CC-BY-4.0.

Citation

If you find this work useful, please cite our papers:

@article {Fallahpour2026.03.19.712954,
    author = {Fallahpour, Adibvafa and Seyed-Ahmadi, Arman and Idehpour, Parsa and Ibrahim, Omar and Gupta, Purav and Naimer, Jack and Zhu, Kevin and Shah, Arnav and Ma, Shihao and Adduri, Abhinav and G{\"u}loglu, Talu and Liu, Nuo and Cui, Haotian and Jain, Arihant and de Castro, Max and Fallahpour, Amirfaham and Cembellin-Prieto, Antonio and Stiles, John S. and Nem{\v c}ko, Filip and Nevue, Alexander A. and Moon, Hyungseok C. and Sosnick, Lucas and Markham, Olivia and Duan, Haonan and Lee, Michelle Y. Y. and Salvador, Andrea F. M. and Maddison, Chris J. and Thaiss, Christoph A. and Ricci-Tam, Chiara and Plosky, Brian S. and Burke, Dave P. and Hsu, Patrick D. and Goodarzi, Hani and Wang, Bo},
    title = {BioReason-Pro: Advancing Protein Function Prediction with Multimodal Biological Reasoning},
    elocation-id = {2026.03.19.712954},
    year = {2026},
    doi = {10.64898/2026.03.19.712954},
    publisher = {Cold Spring Harbor Laboratory},
    URL = {https://www.biorxiv.org/content/early/2026/03/20/2026.03.19.712954},
    eprint = {https://www.biorxiv.org/content/early/2026/03/20/2026.03.19.712954.full.pdf},
    journal = {bioRxiv}
}

@misc{fallahpour2025bioreasonincentivizingmultimodalbiological,
      title={BioReason: Incentivizing Multimodal Biological Reasoning within a DNA-LLM Model}, 
      author={Adibvafa Fallahpour and Andrew Magnuson and Purav Gupta and Shihao Ma and Jack Naimer and Arnav Shah and Haonan Duan and Omar Ibrahim and Hani Goodarzi and Chris J. Maddison and Bo Wang},
      year={2025},
      eprint={2505.23579},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2505.23579}, 
}