OpenChem

OpenChem implements deep learning methods for computational chemistry, drug design, and materials informatics using the PyTorch framework to model chemical structures and properties.

Key Features:

• PyTorch integration: Implements models and development routines using the PyTorch framework.
• Deep learning models: Supports deep learning architectures for modeling chemical structures and predicting properties.
• Modular architecture: Provides a modular software design to enable flexibility and scalability across research applications.
• Data preprocessing modules: Includes modules for preparation and preprocessing of chemical datasets.
• Workflow integration: Facilitates integration of deep learning into computational chemistry workflows for model development.
• Comparison to traditional ML: Positions deep learning approaches relative to traditional machine learning methods such as random forest, support vector machines, and nearest neighbor methods.

Scientific Applications:

• Computational chemistry: Uses deep learning to model chemical properties and explore structure–property relationships.
• Drug design: Applies deep learning models to predict properties relevant to drug discovery and design.
• Materials informatics: Applies deep learning techniques to materials datasets for property prediction and analysis.

Methodology:

Implemented with the PyTorch framework and comprising a modular architecture and data preprocessing modules for development of deep learning models on chemical and materials datasets.