Classes and Metaclasses

We need an alternative classification of chemical atoms and elements that is different than the existing CHEBI. There are deep problems with classification of atoms and molecular entities in CHEBI. These problems arise from a basic confusion about classes and metaclasses. Sources like IUPAC represent chemical species, instances of which are chemical elements, NOT instances in the OBO sense. This fundamental confusion means CHEBI does not work for OBO.

The idea here is to provide a schema in which the instances are classes in a chemical ontology. The schema could be thought of as an ontology of metaclasses, but care must be taken here due to mixing of different metamodel formalisms (this schema is closed-world and thus not well represented in OWL).

The schema language used here is linkml, which provides powerful modeling framework that incorporates aspects of UML, ShEx, and OWL, and is semantic-web ready.

The goal is to create instances of these schema classes (e.g. carbon-14) which would be OWL classes, auto-populated with standardized names and logical axioms. This could be done via a compilation to dospd or robot, or direcly within the LinkML framework (see this issue ). TBD.