Introduction

USP-inhibition is a Python package for the analysis of publically available enzyme inhibition data.

In this project, we build and use quantitative structure-activity relationships (QSAR) models for the prediction of a desired interaction between enzymes and small drug molecules. The data describes the inhibition of USP1 - an enzyme essential to DNA-repair in proliferating cancer cells. Descriptors of the molecular structures of these drugs are computed to populate a working data set from the raw data in the high-throughput screen.

Objectives:

1. Engineer molecular features and build machine learning models to predict the inhibition activity of small molecules.
2. Use genetic algorithms to tease out optimal values of descriptors that contribute to high inhibitory action.
3. Create an reusable open-source tool for cheminformaticians that acts as the first step to intelligent drug design prior to synthesis and testing in a lab.

Resources

• Dataset: view latest version on PubChem or the version used in this project on Amazon S3
• Previous work using this dataset

Dependencies

• ChemoPy
• MOPAC

• Open Babel

• Pybel
• PyChem

• RDKit

• Significant Python packages:
  • nolearn
  • scikit-learn

  How to use this package

  Package requirements are listed here.

  Setup instructions for this package:

  • Ubuntu
  • Windows

  
  

  Call graph for this Python application: