🎮 CONTRA-IL6 🎮

📝 Paper | 🚩 Model & Dataset

The official implementation of paper: CONTRA-IL6

Abstract

Update soon!

News

Update soon!

TOC

This project is summarized to:

• Package installation
  • Pip
  • How to use
• Training
  • Installing environment
  • Preparing datasets
  • Training models
• Citation

Package installation

Pip

Require: Python 3.8

pip install contra-il6

or

pip install -U 'contra-il6 @ git+https://github.com/duongttr/CONTRA-IL6.git'

How to use

contra-il6 [OPTIONS]

Options:
  -i, --input PATH          Path to a FASTA file containing peptide sequences.
  -t, --threshold FLOAT     Threshold for classification (default: 0.46).
  -b, --batch_size INTEGER  Batch size for processing (default: 4).
  -o, --output PATH         Output file to save predictions.
  --help                    Show this message and exit.

Example:

contra_il6 -i data/Validate_positive.txt -o data/Validate_positive_result.csv

Training

Installing environment

Create environment by using conda:

conda create -n contra_il6 python=3.10
conda activate contra_il6

Clone the repo, then install the required packages:

cd CONTRA-IL6/
python -m pip install -r requirements.txt

Preparing datasets

You can check the raw dataset inside data folder. Or you can download 12 extracted features mentioned in paper at this 🔗 link and extract to data folder.

Training models

To reconstruct the checkpoints, you can run following commands:

python train.py --config_path configs/top_4_features.yaml --save_config

Citation

Update soon!