jhl8041/UniKP

UniKP

Description

• Visit original repository for detailed information
• This is simplified version of original repo

Setup

1. Clone git repostiroy

cd {directory-you-want-to-clone}
git clone https://github.com/jhl8041/UniKP.git

2. Setup Anaconda environment

conda create -n Uni_test python=3.7 # can be higher than 3.7
conda activate Uni_test

3. Install dependencies

cd {project-directory}
pip install torch torchvision torchaudio
pip install -r requirements.txt
pip install scikit-learn==1.2.2 # if you have error with scikit-learn

4. Download pretrained model(1)

• Download following files from https://huggingface.co/Rostlab/prot_t5_xl_uniref50/tree/main

pytorch_model.bin
config.json
special_tokens_map.json
spiece.model
tokenizer_config.json

• Place them in prot_t5_xl_uniref50 folder

4. Download pretrained model(2)

• Download following files from https://huggingface.co/HanselYu/UniKP/tree/main

UniKP for Km.pkl
UniKP for kcat.pkl
UniKP for kcat_Km.pkl

• Place them in trained_dataset folder

5. Fill in input data

• Fill sequence and SMILE in inplut/input.xlsx
• Note: When SMILE is longer than 218, it will be truncated

6. Run

python run.py