pKaPredict 0.1.1

pKaPredict project

pKaPredict

📦 Overview

This package provides a streamlined pipeline for predicting the pKa values of molecules from their SMILES strings using machine learning. It includes functionalities that enable descriptor generation via RDKit, and model training using LightGBM and other regressors. The package is designed to be easily pip-installable and modular, making it ideal for cheminformatics applications and molecular property prediction tasks.

👩‍💻 Installation

1. Clone the repository:

git clone https://github.com/anastasiafloris/pKaPredict.git
cd pKaPredict

2. Create and activate the conda environnement:

If your project includes an environment.yml file:

conda env create -n pkapredict_env -f environment.yml
conda activate pkapredict_env

If not, you can create one manually:

conda create -n pkapredict_env python=3.10 -y
conda activate pkapredict_env

3. Install the package:

pip install pKaPredict

4. Install jupyter lab:

pip install jupyterlab

🍏 For macOS users (⚠ required for LightGBM to work):

5. Install the system library libomp:

brew install libomp

If brew is not installed, follow the instructions here: https://brew.sh

6. The package is yours 🎁:

Run the test_package.py file in a python or jupyter environment to predict the pKa of a molecule of your choice, using its SMILES string.

🪪 License

This project is licensed under the MIT License.
You are free to use, modify, and distribute it with proper attribution.

📗 References

The dataset used in this project is the test_acids_bases_descfinal_nozwitterions.csv file from the cbio3lab repository.
It was originally extracted from the Harvard Dataverse.

👯‍♀️ Authors

This project was completed as part of the EPFL course Practical Programming in Chemistry.

• Anastasia Floris
• Candice Habert