The Bioregistry is a metaregistry of identifiers resources, meaning that it harmonizes across entries of multiple external registries via mappings. Each mapping is between a Bioregistry prefix and a specific prefix within one of the external registries the Bioregistry integrtes with. The initial creation of such mappings is largely automated. The vast majority of these mappings is correct but some correspond to prefixes that do not represent the same resource (or more specifically, the same semantic space) and should therefore not be mapped to each other.

The following TSV file contains mappings reviewed by curators. This file follows the SSSOM standard, a community standard for sharing semantic mappings. When reviewing a mapping, curators should update the TSV file with the following information:

  • subject_id: The CURIE corresponding to Bioregistry prefix from which this is a mapping. Even when a mapping is symmetric, as a convention, this should always be the Bioregistry prefix not the external one.
  • predicate_modifier: A modifier on the predicate, typically Not if the mapping is negative.
  • predicate_id: The CURIE representing the predicate of the mapping, for instance skos:exactMatch.
  • object_id: The CURIE corresponding to the external prefix to which this is a mapping.
  • creator_id: The ORCID of the curator reviewing the mapping represented as a CURIE.
  • mapping_justification: A CURIE representing how the mapping was created, typically semapv:ManualMappingCuration for manual curation.
  • comment: Any additional notes or comments regarding the mapping that was reviewed.

Semi-automated Review of Cross-registry Mappings

The Bioregistry uses a machine learning-based approach to automatically identify semantic similarity between the metadata associated with the Bioregistry entry and the external registry entry connected by a mapping. The resulting similarity scores are available in this TSV file. Curation can be prioritized to entries in this file that have low similarity scores but have not been curated yet.

Curating Mappings to Providers in External Registries

The Bioregistry ingest and harmonizes 30+ external registries that contain metadata describing ontologies, databases, and other resources that mint semantic spaces.

Some registries, such as the UniProt Database registry, contain many records for databases that use UniProt IDs, but aren’t the UniProt resource themselves. The Bioregistry currently does not have a mechanism for ingesting this information, but for some applications, it’s useful to have.

These can be curated in the curated_mappings.sssom.tsv SSSOM file using the has provider (bioregistry.schema:0000030) predicate. For example, bioregistry:uniprot has the provider uniprot:DB-0262, which corresponds to the AlphaFoldDB.