Inferring Biochemical Ontology from Metabolic Databases
Author Information
Author(s): D. C. McShan
Primary Institution: University of Colorado Health Sciences Center
Hypothesis
Biochemical reactivity of arbitrary compounds can be inferred directly from their structure using existing metabolic databases.
Conclusion
The study identified over 120,000 substructural relationships, revealing that substructural composition is not sufficient for functional classification.
Supporting Evidence
- Identified 120,455 substructural relationships from 9660 molecular structures.
- Only 433 out of 835 abstract compounds were substructures of at least one compound.
- Proline was not classified as an amino acid despite being functionally one.
Takeaway
The researchers looked at how to classify compounds based on their structure and found that just knowing the structure isn't enough to understand how they work.
Methodology
The study parsed data from the KEGG Ligand database and performed substructural searches to identify relationships between compounds.
Limitations
The study did not validate all identified relationships, and the classification of proline as an amino acid was problematic.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website