canSARchem

canSARchem standardizes and registers chemical structures from public resources such as ChEMBL and PubChem and organizes related compounds into hierarchies to enable integrated medicinal chemistry and bioactivity analysis.

Key Features:

• Integration and Standardization: Integrates data from ChEMBL and PubChem and standardizes chemical structures for consistent representation.
• Hierarchical Organization: Groups structurally related compounds into hierarchies to link alternative forms and associated bioactivity data.
• Canonicalization Process: Performs canonicalization prior to parent structure assignment, aligning with PubChem's OpenEye approach to reduce redundancy.
• Pipeline Components: Implements structure checking, standardization, generation of canonical tautomers and representative structures, salt stripping, and abstract-structure generation.
• Performance and Efficiency: Exhibits a lower compound rejection rate compared to PubChem and ChEMBL.
• Impact on Bioactivity Data Exploration: Enhances exploration of bioactivity data, with FDA-approved drugs showing sensitivity to canonicalization.
• Large-Scale Application: Has standardized over 3 million compounds in the canSAR knowledgebase to enable integrated data analysis.

Scientific Applications:

• Academic drug discovery: Facilitates identification of alternative compound forms and integration of medicinal chemistry data to support candidate selection.
• Translational research: Enables integration of disparate public chemical datasets to support cross-dataset interpretation in translational studies.
• Bioactivity data exploration: Improves retrieval and comparison of bioactivity information across related compound forms and hierarchies.

Methodology:

The pipeline applies structure checking; chemical standardization; generation of canonical tautomers and representative structures; salt stripping; generation of an abstract structure to form compound hierarchies; and canonicalization before parent-structure assignment consistent with PubChem's OpenEye approach.