Unique Identifiers for Chemicals
Two approaches:
-
Registry number
- Assigned by an authority, Applies to any substance, Semantically void,
Usually copyrighted and may require payment
- Examples: CAS, RTECS
-
Graph canonicalisation
- Unique atom labelling assigned algorithmically
- "Layers" manage differences between, Hydrogens and tautomerism, Stereochemistry, Isotopes
- Semantically rich
- IUPAC/NIST is Open and OpenSource, assimilated into CML
Example of INChI v0.932Beta
C7H6O2,1H3-5-4H-6(8)2H-3H-7(5)9