qsarKit

⚛️ QSAR Predictive Modeling for Evaluating Contaminant Transfer

Table of Contents

1. About the project
2. Installation
3. Use cases
4. Tutorials
5. Documentation
6. Contact

📝 About the project

qsarKit is a Python package that offers robust predictive modeling using QSAR for evaluating the transfer of environmental contaminants in breast milk. Developed by the dedicated team led by Professor Nadia Tahiri at the University of Sherbrooke in Quebec, Canada. This open-source integrates multiple predictive models, provides synthetic data generation via GANs, and is tailored for researchers and health professionals.

⚒️ Installation

Miniconda is used to handle the environment dependencies.

Once miniconda is installed, the environment can be created and activated with the following commands:

conda env create -f environment.yaml
conda activate qsar_env

If you encounter any issues activating the environment, try sourcing the Conda script first and then retry activation:

source ~/miniconda3/bin/activate qsar_env

or if you installed Anaconda instead of Miniconda:

source ~/anaconda3/bin/activate qsar_env

⚠️ We currently only support Python 3.10 due to some dependencies that are not yet compatible with Python 3.11+. We will update the package as soon as the dependencies are updated.

🚀 Use cases

The qsarKit package can be encapsulated in other applications or used as a standalone package. You can refer to the tutorials on how to use the package functionalities, or use the package as a standalone application. To perform a quick test, you can run the package with only one model by executing the following command:

python main.py --config ridge.yaml --output results/

For a more generic way of running the package as a standalone application, you can execute the following command by specifying the <config_file> (path to the YAML configuration file) and <output_dir> (path to the output directory).

python main.py --config <config_file> --output <output_dir>

Both arguments are optional. If not provided, the default values are config/compare_all_models.yaml and results/, respectively.

We can also generate synthetic data using GANs by including the gan flag in the configuration file. You can explore examples of the different options provided by the package in the config/ folders. And you can refer to the gan tutorial.

📚 Tutorials

We provide several tutorials to help you get started with the package. You can find them in the tutorials/ folder. You can explore the tutorials/models/, tutorials/gan/, and tutorials/preprocessing/ folders to learn more about the different functionalities of the package.

📖 Documentation

You can also refer to the documentation for more details.

We generated the documentation using Sphinx. To generate the documentation locally, you can run the following command:

Linux/Mac:

cd docs/
make html

Windows:

cd docs/
.\make.bat html

The documentation will be generated in the docs/build/html/ folder. You can open the index.html file in your browser to view the documentation.

📧 Contact

Please email us at: [email protected] for any questions or feedback.

Go to Top