TEPCAM

Prediction of T cell receptor-epitope binding specificity via interpretable deep learning

Requirements

TEPCAM is constructed using python 3.8.16. The detail dependencies are recorded in requirements.txt.

To install from the requirements.txt, using:

pip install -r requirements.txt

Using a conda virtual environment is highly recommended.

conda create -n TEPCAM
conda activate TEPCAM
conda install -r requirements.txt

Model Training

For training TEPCAM, run command below:

python ./scripts/train.py --input_file="./Data/TEP-merge.csv" --model_name="TEPCAM" --epoch=30 --learning_rate=5e-4 --GPU_num=0

Model Evaluation

Model evalutation using test.py, run command below:

python ./scripts/test.py 
--file_path="./Data/ImmuneCODE.csv" 
--model_path="./model/TEPCAM_test.pt" 
--output_file="./output.csv" 
--metric_file="./metric_file.csv"