PIDapalooza 2018

Determining identifier equivalency is key to data integration and to realizing the scientific discoveries that can only be made by collating our vast disconnected data stores.
There are two key problems in determining equivalency - conceptual and syntactic alignment. Conceptual alignment often relies on Xrefs and string-matching against synonyms. There is indeed a better way! Algorithmic determination of identifier equivalency across different sources can use a combination of Xrefs, priors rules, existing semantic relations, and synonyms to create equivalency cliques than can highlight the discrepancies in conceptual definitions for manual review. This is especially useful for data sources annotated with concept drift and differences, such as diseases. Syntactic issues are that there are so many variations of the same identifier, making data joins difficult. We present a framework to reconcile and provide authoritative and integration-ready prefixed identifiers (CURIES), to capture and consolidate prefixes and to build links across key resource registries. The combination of JSON-LD context technology with a prefix metadata repository provides the basis for the infrastructure to handle identifiers in a consistent fashion. Finally, this architecture also allows resources to be self describing "beacons" with respect to their identifiers.