CGRdb2

CGRdb2.0 provides a PostgreSQL-backed Python package for storing, searching, and analyzing molecules, reactions, and related chemical data for cheminformatics and computational chemistry applications.

Key Features:

• Python/PostgreSQL integration: Implemented as a Python package that interfaces with PostgreSQL for database management of chemical entities.
• Molecular similarity search: Implements similarity search algorithms for molecules.
• Molecular substructure search: Implements substructure search functionality for molecules.
• Reaction search — component-based: Supports reaction searches based on reaction components.
• Reaction search — CGR-based: Supports reaction searches using the Condensed Graph of Reaction (CGR) representation.
• CGR performance advantage: Uses the CGR approach to improve reaction search performance relative to component-based methods.
• Benchmarking with RDKit: Benchmarking studies with the RDKit database cartridge reported faster search speed on smaller datasets compared to traditional methods.

Scientific Applications:

• Computational chemistry data retrieval: Retrieval and querying of molecular structures and reaction records for computational chemistry workflows.
• Reaction pattern identification: Identification and retrieval of reaction patterns using component-based and CGR representations.
• Cheminformatics search benchmarking: Comparative benchmarking of search algorithms and database cartridges such as RDKit for performance evaluation.

Methodology:

Implemented as a Python package interfacing with PostgreSQL; implements molecular similarity and substructure searches; supports reaction searches using reaction component-based and Condensed Graph of Reaction (CGR) representations; benchmarking performed using the RDKit database cartridge reporting improved search speed on smaller datasets.