Add new CURIE prefixes for instance-level entities (e.g., person, device_exposure) to Biolink Model context #1627
Description
Is your feature request related to a problem? Please describe.
Yes. When creating knowledge graphs that follow the Biolink Model, validation tools such as KGX report errors for instance identifiers that use prefixes like person, visit, observation, or device_exposure.
For example:
Node property 'id' has a value 'person:7604828642657013760' with a CURIE prefix 'person'
that is not represented in the Biolink Model JSON-LD context.
This occurs because these prefixes—representing entities that are already defined as Biolink classes—are not currently registered in the Biolink Model JSON-LD context.
As a result, instance-level identifiers cannot be validated as Biolink-compliant even though the entity types themselves exist in the model.
Describe the solution you'd like
We request either:
The addition of CURIE prefixes (e.g., person, visit, condition_occurrence, device_exposure, measurement) to the Biolink Model JSON-LD context for use with instance-level entities, or
Guidance on namespace conventions or URI patterns for representing locally defined instance identifiers that correspond to existing Biolink classes.
If existing namespaces should be reused, please clarify the recommended mappings or canonical URI patterns to maintain consistent Biolink compliance across implementations.
What working group (or team) did this request originate from?
This request arises from data modeling and integration work aligned with the NCATS Translator ecosystem, where Biolink-compliant instance-level representations are required for validation and interoperability.
Additional context
This enhancement would improve validation and interoperability for Biolink-aligned knowledge graphs that include instance-level data.
The issue is commonly encountered during validation (KGX) and transformation (Koza) workflows where locally scoped identifiers are needed for entities corresponding to existing Biolink classes.
Tag relevant members for discussion
@biolink/maintainers
@NCATSTranslator/translator-team
@sierra-moxon @mbrush