Unifying the identification of biomedical entities with the Bioregistry

Charles Tapley Hoyt, Meghan A. Balk, Tiffany J. Callahan, Daniel Domingo‐Fernándéz, Melissa Haendel, Harshad Hegde, Daniel Himmelstein, Klas Karis, John Kunze, Tiago Lubiana, Nicolas Matentzoglu, Julie A. McMurry, Sierra Moxon, Chris Mungall, Adriano Rutz, Deepak Unni, Egon Willighagen, Donny Winston, Benjamin M. Gyori

Abstract

The standardized identification of biomedical entities is a cornerstone of interoperability, reuse, and data integration in the life sciences. Several registries have been developed to catalog resources maintaining identifiers for biomedical entities such as small molecules, proteins, cell lines, and clinical trials. However, existing registries have struggled to provide sufficient coverage and metadata standards that meet the evolving needs of modern life sciences researchers. Here, we introduce the Bioregistry, an integrative, open, community-driven metaregistry that synthesizes and substantially expands upon 23 existing registries. The Bioregistry addresses the need for a sustainable registry by leveraging public infrastructure and automation, and employing a progressive governance model centered around open code and open data to foster community contribution. The Bioregistry can be used to support the standardized annotation of data, models, ontologies, and scientific literature, thereby promoting their interoperability and reuse.

Introduction: The Biomedical "Tower of Babel"

In the life sciences, the precise identification of entities—ranging from proteins and chemicals to diseases and clinical trials—is fundamental to data integration. To achieve Findable, Accessible, Interoperable, and Reusable (FAIR) data, researchers utilize Compact Resource Identifiers (CURIEs), such as uniprot:P0DP23. However, a "Tower of Babel" problem has historically hindered these efforts.

The landscape of identifier registries is characterized by significant fragmentation. Different registries often assign different prefixes to the same resource. For instance, the NCBI Taxonomy is identified as taxonomy in Identifiers.org but as NCBITAXON in BioPortal. Furthermore, many existing registries suffer from metadata gaps, lack of versioning, and restrictive governance models that prevent community-driven updates. Often built on private institutional infrastructure, these registries are also prone to "link rot" and sustainability issues. Without a unified source of truth, integrating disparate datasets remains a manual and error-prone challenge.

Methods: A Metaregistry Approach

To address these challenges, we developed the Bioregistry, an open-source "metaregistry" designed to integrate and harmonize 23 existing identifier registries. The Bioregistry is built upon three core pillars:

  1. Integrative Alignment: The system automatically imports records from 23 external registries. A multi-stage alignment algorithm identifies equivalences between records, resolving conflicts in capitalization (e.g., go vs. GO) and punctuation to create a single, reliable hub.
  2. Governance-as-Code: Moving away from closed committees, the Bioregistry utilizes a community-driven model hosted on GitHub. Anyone can propose new prefixes or metadata corrections via Pull Requests, which are then validated through automated quality assurance workflows (GitHub Actions).
  3. Technical Accessibility: The Bioregistry provides a comprehensive technical stack, including a Python package for programmatic access, a web portal, a Docker image, and a RESTful API documented with OpenAPI.
The Bioregistry acts as a central hub, mapping conflicting prefixes (e.g., for NCBI Taxonomy) from various external registries to a unified standard.

Results and Key Insights

The Bioregistry serves as a superset of the current identifier ecosystem. As of version 0.5.132, it contains 1,504 individual records and 7,150 cross-registry mappings. Notably, 192 of these records are entirely novel, curated by the community to fill gaps in existing major registries.

Expansion and Novelty

The Bioregistry represents a massive expansion in coverage, containing 81% more records than Identifiers.org and nearly double those of Prefix Commons. It provides new curated metadata for 74% of all existing records, significantly enriching the global knowledge base.

Registry coverage comparison (a) and the distribution of identifier resources across external registries (b), illustrating the 'long tail' of fragmented data.

The "Long Tail" of Science

Analysis of the integrated data revealed a highly fragmented landscape: only seven resources appeared in more than 13 of the 23 registries. This "long tail" distribution highlights the necessity of a metaregistry to bridge the gaps between siloed standards.

Sustainability and Adoption

By leveraging public infrastructure such as GitHub Actions, Zenodo, and Amazon Web Services, the technical maintenance cost of the Bioregistry has been reduced to approximately $60 per year. This low-cost model ensures resilience against funding fluctuations. The system has already seen immediate adoption, integrated into major pipelines like INDRA for automated model assembly, BridgeDb for identifier mapping, and the OBO Foundry.

Discussion and Future Work

While the Bioregistry provides a robust framework for standardization, challenges remain. It cannot retroactively enforce standards on legacy databases or external registries. Furthermore, while it identifies equivalent resources, it does not currently provide a ranking system to guide users on which resource to prioritize.

Future work will focus on expanding the metaregistry methodology to other scientific domains and continuing to grow the contributor base. The Bioregistry demonstrates that a community-driven, automated approach can effectively solve long-standing interoperability issues in the life sciences, providing a template for sustainable data infrastructure.

The Bioregistry can be accessed at https://bioregistry.io. Its source code and data are available under MIT and CC0 Licenses at https://github.com/biopragmatics/bioregistry.

Summary updated

Abstract

The standardized identification of biomedical entities is a cornerstone of interoperability, reuse, and data integration in the life sciences. Several registries have been developed to catalog resources maintaining identifiers for biomedical entities such as small molecules, proteins, cell lines, and clinical trials. However, existing registries have struggled to provide sufficient coverage and metadata standards that meet the evolving needs of modern life sciences researchers. Here, we introduce the Bioregistry, an integrative, open, community-driven metaregistry that synthesizes and substantially expands upon 23 existing registries. The Bioregistry addresses the need for a sustainable registry by leveraging public infrastructure and automation, and employing a progressive governance model centered around open code and open data to foster community contribution. The Bioregistry can be used to support the standardized annotation of data, models, ontologies, and scientific literature, thereby promoting their interoperability and reuse. The Bioregistry can be accessed through https://bioregistry.io and its source code and data are available under the MIT and CC0 Licenses at https://github.com/biopragmatics/bioregistry .
BibTeX
@inproceedings{hoyt2022unifying,
  author = {Charles Tapley Hoyt and Meghan A. Balk and Tiffany J. Callahan and Daniel Domingo‐Fernándéz and Melissa Haendel and Harshad Hegde and Daniel Himmelstein and Klas Karis and John Kunze and Tiago Lubiana and Nicolas Matentzoglu and Julie A. McMurry and Sierra Moxon and Chris Mungall and Adriano Rutz and Deepak Unni and Egon Willighagen and Donny Winston and Benjamin M. Gyori},
  title = {{Unifying the identification of biomedical entities with the Bioregistry}},
  booktitle = {Scientific},
  year = {2022},
  doi = {10.1038/s41597-022-01807-3},
}
Figure