PLM-DBPPred

Cite the code:

What is PLM-DBPPred?

A DNA binding protein prediction tool based on proteinBERT, ESM2_30 and protT5.

ProteinBERT Env

ProteinBERT environment https://github.com/nadavbra/protein_bert.

# python==3.8.0 Tensorflow==2.5.0 CUDA Version: 11.4 
conda env create -f ProteinBERT_env.yml
conda activate PLM-DBPPred_ProteinBERT

Using ProteinBERT

ProteinBERT

Training

python train.py -train_set ../DBP_dataset/DBP_Predict_refine_13289_230726.train.csv -test_set ../DBP_dataset/DBP_Predict_PDB.test.csv -exp_name test_py

Predict

python predict.py -test_set ./DBP_dataset/DBP_Predict_PDB.test.csv -o ./ -m ./DBP_model_param/

ESM and ProtT5 Env

ESM2 environment https://github.com/facebookresearch/esm. protT5 environment https://github.com/HannesStark/protein-localization.

# python==3.8.0 Pytorch==2.0.1 CUDA Version: 11.4 
conda env create -f ESM_ProtT5_env.yml
conda activate PLM_DBPPred_ESM_ProtT5
pip install -r ESM_ProtT5_requirements.txt
pip install bio-embeddings[all]
pip install fair-esm  # latest release, OR:
pip install git+https://github.com/facebookresearch/esm.git

Using ProtT5

ProtT5

Used the bio-embedding to generate the .h5 file

#Change the file path (sequences_file: ./Train_dataset.fasta prefix: ./Train_dataset_emb)
bio_embeddings light_attention.yml

Training

#Change the file path of .h5 and .fasta
python train.py --config ./configs/DBP_light_attention_train.yml

Predict

#Change the file path of .h5 and .fasta
python predict.py --config ./configs/DBP_test_predict.yml

Using ESM

ESM

Training

python train.py -train_set ../DBP_dataset/DBP_Predict_refine_13289_230726.train.csv -test_set ../DBP_dataset/DBP_Predict_PDB.test.csv -exp_name outputs

Predict

python predict.py -test_set ../DBP_dataset/DBP_Predict_PDB.test.csv  -o run_result/ -m ESM2_30_model_param.t7

Citing PLM-DBPPred

@article{hu2024systematic,
  title={Systematic discovery of DNA-binding tandem repeat proteins<? mode longmeta?>},
  author={Hu, Xiaoxuan and Zhang, Xuechun and Sun, Wen and Liu, Chunhong and Deng, Pujuan and Cao, Yuanwei and Zhang, Chenze and Xu, Ning and Zhang, Tongtong and Zhang, Yong E and others},
  journal={Nucleic Acids Research},
  pages={gkae710},
  year={2024},
  publisher={Oxford University Press}
}