Beyond the Lab Website: Reusing Structured Research Metadata Across Lab Operations
Year: 2026
Type: article
Abstract: We compare two ways to deploy ORINOCO, a system for collaboratively managing and reusing structured research metadata. One way uses dedicated services to store records and review changes; the other uses a GitHub repository and pull requests. Research groups repeatedly need facts about people, projects, and publications for CVs, grant applications, and yearly reporting. Those facts are scattered across email, spreadsheets, institutional systems, and web pages, so each task begins by recovering what the lab already knows. Systems such as ORCID and SciENcvmanage specific record types, but research groups still need an open collection they control that brings these records together and describes their relationships. Such infrastructure improves efficiency and accuracy in lab operations and provides a foundation for lab websites. Research-information management is an established field [1]-[3]. CERIF and systems such as VIVO model institutional records, while scholarly knowledge graphs connect metadata at global scale [4], [5]. Many use Linked Data: shared identifiers and formally defined terms make relationships machine-readable across independently maintained sources, allowing information to be connected and reused without every system maintaining its own copy. A schema defines the types of records, their relationships, and the rules used to validate them. Identifiers such as ORCID and DOI let a lab connect its records to this broader landscape while describing local roles, relationships, and tools. This explicit modeling of lab metadata becomes especially valuable with AI: as AI makes it easier to extract and generate candidate content, the harder task is organizing, validating, and reviewing that content so that only trustworthy information enters the lab’s official record. Our long-time collaborator Michael Hanke and colleagues in the Psychoinformatics group at the INM-7 Institute of Forschungszentrum Jülich in Germany developed ORINOCO (Organized Research Information: Ontology-mapping, Curation, Orchestration). ORINOCO addresses this need at research-group scale through an open, self-hostable set of interoperating components [6]. It is grounded in LinkML schemas which provide machine-readable definitions of people, projects, grants, research outputs, and their relationships. These schemas drive browser-based forms for entering records and a service that validates submissions and stages them for review before they join the lab’s curated collection. Query and rendering tools can then follow the relationships among approved records and reshape them for websites, reports, catalogs, discovery, and other lab operations. The Lab-in-a-Box deployment toolkit places ORINOCO alongside other lab-operated services [7]. The Psychoinformatics group website demonstrates how linked records organize pages, related-item lists, backlinks, and graph navigation; the TRR379 research consortium applies the same approach, with schema-generated interfaces accepting records from people and automated processes [6]. At the Center for Open Neuroscience (CON), we are adopting ORINOCO for our lab website and broader research-information needs, retaining its schemas and processing tools while adapting their operation to GitHub-centered collaboration and review in ORINOCO-Lite [8]. Records stored as YAML text files in Git constitute the official collection. Whether edited directly or prepared automatically from existing lab sources such as our Zotero publication group, all proposed changes are reviewed through pull requests. A GitHub Action uses ORINOCO components to validate the records and regenerate the website. For RSEs, the central opportunity is to treat lab metadata as shared infrastructure rather than maintain it separately for each application. A website provides an immediate and visible use for that infrastructure, but its larger value is an enduring body of curated, structured information that can be reused as the lab’s needs and tools evolve.
Authors: Lee, John, To, Isaac, Halchenko, Yaroslav
Institutions: Dartmouth College, Center for Open Neuroscience
Language: English
Topic: Optics and Image Analysis
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7213662830