Metabolic Disassembler

Metabolic Disassembler predicts biosynthetic units (BUs) from natural product chemical structures to identify starting materials for biosynthetic pathway analysis.

Key Features:

• Automatic Prediction: Disassembles target molecule structures into substructures corresponding to biosynthetic units (BUs) that represent starting materials in biosynthetic pathways.
• Biosynthetic Unit Library (BUL): Utilizes a comprehensive library containing 765 registered BUs derived from the KEGG PATHWAY database, including starting materials, key intermediates, and derivatives.
• Predictive Model: Employs a model that predicts BU assignments based on matches to the Biosynthetic Unit Library.
• Implementation: Implemented in Python.
• High Accuracy: Achieved 95.8% accuracy on 943 molecules from KEGG PATHWAY secondary metabolism maps, compared to ~52.0% for a genetic algorithm–based workflow.
• Efficiency: Completed calculations within one minute for 90.7% of tested molecules.

Scientific Applications:

• Biosynthetic pathway elucidation: Automates identification of likely starting materials to support reconstruction and analysis of natural product biosynthetic routes.
• Natural product discovery and design: Assists exploration of functional materials such as medicines derived from natural sources by revealing constituent biosynthetic units.
• Synthetic biology production planning: Provides insights into starting materials and intermediates to inform synthetic biology strategies for producing target natural products.

Methodology:

Disassembles target molecules using a predictive model based on the Biosynthetic Unit Library (BUL) of 765 BUs from KEGG PATHWAY; implemented in Python and optimized for accuracy and rapid computation, allowing input of molecules not present in existing databases.