PROTAC-Degradation-Predictor: Documentation and Overview

PROTAC-Degradation-Predictor is a Python package designed to predict the activity of PROTAC molecules using advanced machine learning techniques. The tool aims to assist researchers in evaluating the potential effectiveness of PROTACs, a novel class of drugs that target protein degradation.

The package Github repository can be found here.

Introduction

PROTACs (Proteolysis Targeting Chimeras) are a class of molecules that induce the degradation of specific proteins. This package allows researchers to predict the activity of PROTACs by leveraging a variety of machine learning models, including XGBoost and PyTorch-based pretrained neural networks.

The primary functionalities of this package include: - Predicting PROTAC activity using different machine learning models. - Accessing curated datasets for training and evaluation. - Hyperparameter tuning and model training using Optuna.

Features

• Machine Learning Models: Utilize XGBoost, PyTorch, and scikit-learn models to predict PROTAC activity (refer to the protac_degradation_predictor.protac_degradation_predictor.get_protac_active_proba() function).

• Dataset Handling: Load and manage datasets specific to PROTAC research (refer to the protac_degradation_predictor.data_utils.load_curated_dataset() function).

• Customizability: Tune model hyperparameters and experiment with different model configurations (refer to the protac_degradation_predictor.optuna_utils.hyperparameter_tuning_and_training() function).

Quickstart

To get started with PROTAC-Degradation-Predictor, follow these steps:

1. Installation: Install the package using pip:

2. Basic Usage: Here’s an example of how to predict PROTAC activity:

   from protac_degradation_predictor import get_protac_active_proba
   
   smiles = "CC(C)C1=CC=C(C=C1)C2=NC3=CC=CC=C3C(=O)N2"
   e3_ligase = "Q9Y6K9"
   target_uniprot = "P04637"
   cell_line = "HCT116"
   
   prediction = get_protac_active_proba(
       protac_smiles=smiles,
       e3_ligase=e3_ligase,
       target_uniprot=target_uniprot,
       cell_line=cell_line,
       device='cpu',
       use_models_from_cv=False,
       use_xgboost_models=True,
       study_type='standard'
   )
   print(prediction)
   

For more detailed usage and customization, refer to the relevant sections below.

Contents

Documentation Contents:

• protac_degradation_predictor
  • protac_degradation_predictor package
• protac_degradation_predictor package
  • Subpackages
  • Submodules
  • protac_degradation_predictor.config module
  • protac_degradation_predictor.data_utils module
  • protac_degradation_predictor.optuna_utils module
  • protac_degradation_predictor.optuna_utils_xgboost module
  • protac_degradation_predictor.protac_dataset module
  • protac_degradation_predictor.protac_degradation_predictor module
  • protac_degradation_predictor.pytorch_models module
  • protac_degradation_predictor.sklearn_models module
  • Module contents

Getting Help

If you encounter any issues or have questions, please refer to the following resources:

• Documentation: Full API documentation and user guide.

• GitHub Issues: Report bugs or request features on the GitHub Issues page.

• Contributing: Learn how to contribute to the project by reading our Contribution Guidelines.

License

This project is licensed under the MIT License. See the LICENSE file for details.

About

Author: Stefano Ribes

Version: v1.0.2

Built with Sphinx using the Read the Docs theme.

---

This documentation was last updated on August 27, 2024.