I submitted an abstract to the 3rd Base4NFDI User Conference (UC4B2026) on behalf of the NFDI Section Metadata Working Group for Ontology Harmonization and Mapping describing the current uses and future prospects for the Semantic Farm throughout the NFDI and Base4NFDI. This post contains an extended version of the original submission.

I’d like to extend a special thanks to co-authors Mark Doerr (University Greifswald), Ulrik Stervbo (Ruhr-Universität Bochum), Benjamin Zapilko (GESIS), and Sonja Herres-Pawlis (RWTH Aachen). Their interest and support for the Semantic Farm has been key in its integration in the NFDI Section Metadata Working Group for Ontology Harmonzition and Mapping and next steps towards its adoption into other NFDI sections, working groups, and services.

We posted the original abstract to Zenodo such that it can be cited by https://doi.org/10.5281/zenodo.21686485. The following abstract below has minor embellishments (and many added links for context) and an extra section.

Abstract

Title: Operationalizing identifier interoperability across the NFDI with the Semantic Farm

A key challenge in achieving the interoperability facet of the FAIR data principles is the consistent identification of researchers, chemicals, paintings, and other entities relevant for NFDI. Often, this means choosing the correct standard uniform resource identifier (URI) or compact URI (CURIE) for an entity from an ontology, controlled vocabulary, persistent identifier (PID) service, or database.

This presents a challenge when multiple CURIEs or URIs can be constructed for the same entity, which is compounded by the independent evolution of tools and services relevant for different NFDI consortia. For example, the entry for water in the Chemical Entities of Biomedical Interest (ChEBI) can be identified by either the URIs https://www.ebi.ac.uk/chebi/CHEBI:15377 or http://purl.obolibrary.org/obo/CHEBI_15377 or by the CURIEs chebi:15377, CHEBI:15377, or CHEBIID:15377. An organization-wide policy is required to determine and communicate which is correct.

The NFDI has not yet adopted an actionable, organization-wide policy for standardizing CURIEs and URIs. We present the Semantic Farm (previously called The Bioregistry) on behalf of the NFDI Section Metadata Working Group for Ontology Harmonization and Mappings as a pre-existing, mature solution for the standardization of CURIEs and URIs. It acts as a centralized index of metadata for resources that mint identifiers that can be readily adopted on the NFDI-level and beyond.

Importantly, the Semantic Farm is a foundational service that can directly support complementary base services (PID4NFDI, TS4NFDI, and KGI4NFDI), NFDI sections, their respective working groups, and researchers in NFDI consortia towards improving interoperability. We highlight several existing applications of the Semantic Farm within NFDI:

  1. TS4NFDI uses both the flagship instance of the Ontology Lookup Service (hosted by the European Bioinformatics Institute) and the TIB Terminology Service instance, which both use the Semantic Farm for URI compression, CURIE expansion, and generation of web links for database cross-references.
  2. Section Metadata Working Group for Ontology Harmonization and Mapping uses the Semantic Farm capture ontology lists used by each consortium at https://semantic.farm/nfdi.
  3. The Semantic Farm supports the construction of knowledge graphs with standardized URIs such as in Section EduTrain’s DALIA platform for open educational resources and Section International Engagement Working Group for Landscaping and Outreach’s bibliometric knowledge graph. Standardization facilitates the integration of external data from ORCiD, ROR, CORDIS, and Wikidata.
  4. The Semantic Farm’s codebase is used by the LinkML runtime, which has growing adoption across consortia such as NFDI4Chem (e.g., MS DCAT AP), NFDI4Cat (e.g., CoreMeta4Cat), and GHGA (e.g., ghga-metadata-schema). It can be further used to standardize the prefix maps in LinkML schemas
  5. The Semantic Farm supports consortia like NFDI4Chem that must standardize and ultimately harmonize a variety of ontology, database, and structural identifiers like InChI.

Further information on the Semantic Farm’s KPIs can be found here. We aspire to submit Semantic Farm to Base4NFDI Service Support Track and to support its integration in more NFDI services and usage in data resources.

Additional Context

This section was not part of the original abstract.

Semantic Farm as a data standard The Semantic Farm induces a standard prefix map that is simultaneously a reflection of real-world usage as well as community-curated standards for CURIE prefixes to URI prefixes. The flagship Semantic Farm software package is implemented in Python, but the prefix map can be operationalized as a data standard through the lower-level curies.rs software package, which has bindings to Python, Java, R, Rust, and JavaScript (via WASM) .makes this an actionable standard. For example:

  1. LinkML schemas can be checked to use correct CURIE prefixes and URI prefixes, see this previous post.
  2. Prefix maps in triple stores, such as the NFDI4Culture Culture Knowledge Graph, can be contextualized and checked for standard CURIE prefix usage, see this previous post.
  3. Turtle (and other RDF) artifacts, such as those produced for SKOS vocabularies and knowledge graphs, can be validated similarly, see this previous post.

Further, there is an evolving (but not yet well-documented) notion of NFDI standards where we would like to promote the Semantic Farm prefix map.

Semantic Farm enables data integration The Semantic Farm was a key tool that enabled the integration of the NFDI4Chem and NFDI4Culture’s KGs that wasn’t possible using other NFDI tools like TS4NFDI and KGI4NFDI. This scenario was described in detail here.

Semantic Farm has demonstrated widespread impact While measuring the impact of a data resource or software package is challenging, we collated a combination of direct and indirect usages of the Semantic Farm across programming languages, use cases, and countries in this post.

Semantic Farm for KG4NFDI

KG4NFDI could proactively adopt the Semantic Farm’s prefix map as a standard, contribute towards the design and implementation of new validation workflows, and use its organizational energy to promote the standard across consortia, sections, and section working groups that produce (and consume) knowledge graphs.

This could directly unlock opportunities for federation and interdisciplinary querying and exploration.

Semantic Farm for PID4NFDI

The idea of persistent identifiers (PIDs) is an important facet of the standardized identification of entities, but frequent narrow focus on popular PID providers (ORCiD, ROR, DOI, etc.) often obscures the other important aspects of standardized identification. For example, while the question of how to identify entity types outside the popular PID providers is often eschewed, the Semantic Farm is a database that describes thousands of resources that mint PIDs for other entity types (e.g., proteins).

The Semantic Farm can be used as an informational resource by PID-minded NFDI consortia, sections, and working groups to better understand the challenges and best practices surrounding standardized identification of entities and how they fit into the larger goals of data standardization, data integration, and ultimately application to scientific endeavors.

The Semantic Farm can also provide a more straightforward way of documenting popular PID providers and what they do. Accordingly, the Semantic Farm has a collection describing the popular PID providers listed by the PID4NFDI on their website.