qsarify

qsarify is a library of tools for the analysis of QSAR/QSPR datasets and models. This library is intended to be used to produce models which relate a set of calculated chemical descriptors to a given numeric endpoint. Many great tools will take the geometry or string data of a given chemical and compute descriptors, which are numeric measures of the properties of these, but you can generate some of these with another one of my scripts, Free Descriptors.

Dependencies

• Python 3
• numpy
• pandas
• scikit-learn
• matplotlib

Installation

pip install qsarify

What is included right now?

• Data preprocessing tools: data_tools
• Dimensionality reduction via clustering: clustering
• Feature selection:
  • Single threaded: feature_selection_single
  • Multi-threaded: feature_selection_multi
• Model Export and Visualization: model_export
• Cross Valiidation: cross_validation

How to use

The best way to learn how to use this library is to look at the example notebook in the examples folder. This notebook will walk you through the workflow of using this library to build a QSAR model.

Future Plans

• Massively parallel feature selection methods:
  • CUDA acceleration
  • MPI acceleration
• Include Shannon Entropy as a dimensionality reduction metric in clustering
• Embedded kernel methods
• More visualization tools
• More cross validation tools
• Feature selection tools for categorical data

Contributing

If you would like to contribute to this project, please feel free to fork this repository and submit a pull request. Otherwise, you may also submit an issue. I will try to respond to issues as quickly as possible.

License

This project is licensed under the GNU GPLv3 license. See the LICENSE file for more details.

Citation

If you use this library in your work, please cite it as follows:

Szwiec, Stephen. (2023). qsarify: A high performance library for QSAR model development.

BibTex:

@misc{szwiec2023qsarify,
  author = {Szwiec, Stephen},
  title = {qsarify: A high performance library for QSAR model development},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/stephenszwiec/qsarify}},
  }