Improved Models

Trained new models with improved performance.

Stable Release

This release includes changes implemented after the completion of the peer review process of our work. In particular, it adds the following changes and improvements:

• Bug-fixes on the data used for training and testing models
• Improved the API to include XGBoost models
• Improved the API to include models trained in different studies
• Improved the API to include cross-validation models
• Added additional experiments and evaluation of models

Initial Release: PROTAC-Degradation-Predictor v1.0

We're excited to announce the first release of the PROTAC-Degradation-Predictor, a machine learning-based tool designed to predict PROTAC protein degradation activity.

This release includes:

• A comprehensive data curation process, detailed in the data_curation.ipynb notebook.
• A open-source training pipeline for the deep learning models, evaluated in the run_experiments.py file.
• Easy installation process with minimal dependencies.
• A user-friendly API that allows you to predict the activity of a PROTAC molecule with just a few lines of code.
• A detailed tutorial in the protac_degradation_tutorial.ipynb notebook, guiding you through the usage of the package.
• Support for batch computation, allowing you to predict the activity of multiple PROTACs at once (also on GPU).

This tool has been developed on a Linux machine with Python 3.10.8. We recommend using a virtual environment to avoid conflicts with other packages.

We look forward to your feedback and contributions!

Full Changelog: https://github.com/ribesstefano/PROTAC-Degradation-Predictor/commits/v1.0.0