DeProt

Code for DeProt. A protein language model with quantizied structure and disentangled attention

1 Install

git clone https://github.com/ginnm/DeProt.git
cd DeProt
pip install -r requirements.txt
export PYTHONPATH=$PYTHONPATH:$(pwd)

2 Structure quantizer

from deprot.structure.quantizer import PdbQuantizer
processor = PdbQuantizer(structure_vocab_size=2048) # can be 20, 128, 512, 1024, 2048, 4096
result = processor("example_data/p1.pdb", return_residue_seq=False)

Output:

[407, 998, 1841, 1421, 653, 450, 117, 822, ...]

3 DeProt models have been uploaded to huggingface 🤗 Transformers

from transformers import AutoModelForMaskedLM, AutoTokenizer
model = AutoModelForMaskedLM.from_pretrianed("AI4Protein/DeProt-2048", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("AI4Protein/DeProt-2048", trust_remote_code=True)

3.1 Available models

Model	Description
AI4Protein/Deprot-20	structure vocab size = 20
AI4Protein/Deprot-128	structure vocab size = 128
AI4Protein/Deprot-512	structure vocab size = 512
AI4Protein/Deprot-1024	structure vocab size = 1024
AI4Protein/Deprot-2048	structure vocab size = 2048
AI4Protein/Deprot-4096	structure vocab size = 4096
AI4Protein/Deprot-2048-NO_AA2SS	Ablative
AI4Protein/Deprot-2048-NO_SS2AA	Ablative
AI4Protein/Deprot-2048-NO_AA2POS	Ablative
AI4Protein/Deprot-2048-NO_POS2AA	Ablative

4 Zero-shot mutant effect prediction

4.1 Example notebook

Zero-shot mutant effect prediction

4.2 Run ProteinGYM Benchmark

Download dataset from Google Driver. (This file contains quantized structures within ProteinGYM).

cd example_data
unzip proteingym_benchmark.zip

python zero_shot/proteingym_benchmark.py --model_path AI4Protein/DeProt-2048 \
--structure_dir example_data/structure_sequence/2048

5 Representation

6 Transfer-Learning

Citation

If you use DeProt in your research, please cite the following paper:

@article {Li2024.04.15.589672,
	author = {Mingchen Li and Yang Tan and Bozitao Zhong and Ziyi Zhou and Huiqun Yu and Xinzhu Ma and Wanli Ouyang and Liang Hong and Bingxin Zhou and Pan Tan},
	title = {DeProt: A protein language model with quantizied structure and disentangled attention},
	elocation-id = {2024.04.15.589672},
	year = {2024},
	doi = {10.1101/2024.04.15.589672},
	publisher = {Cold Spring Harbor Laboratory},
	URL = {https://www.biorxiv.org/content/early/2024/04/17/2024.04.15.589672},
	eprint = {https://www.biorxiv.org/content/early/2024/04/17/2024.04.15.589672.full.pdf},
	journal = {bioRxiv}
}