Reusing the NCBO BioPortal Technology for Agronomy to Build AgroPortal

Clément Jonquet, Anne Toulet, Elizabeth Arnaud, Sophie Aubin, Esther Dzalé Yeumo Kaboré, Vincent Emonet, John B. Graybeal, Mark A. Musen, Cyril Pommier, Pierre LarmandeICBO/BioCreative 2016

TL;DR

Researchers have adapted the successful NCBO BioPortal, a biomedical ontology repository, to create AgroPortal, a similar platform for agronomic sciences. This initiative aims to streamline data integration and knowledge management in agriculture by providing a centralized, semantic web-compliant infrastructure for ontologies.

What Problem Does This Paper Address?

The field of agronomy, much like biomedicine, is generating vast amounts of data from diverse sources. To effectively integrate, annotate, and analyze this data, a standardized way to represent and understand concepts is crucial. This involves using ontologies – structured vocabularies that define relationships between terms. However, the proliferation of these ontologies across various agronomic sub-domains (plant breeding, genetics, environment, etc.) has led to a fragmented landscape. Researchers and data managers face challenges in discovering, accessing, and utilizing these ontologies, as well as managing their versions and alignments. The paper addresses the need for a unified, user-friendly platform that can host, search, visualize, and annotate using these agronomic ontologies, mirroring the success of existing biomedical resources.

Methods

The core methodology of this paper is the reuse and adaptation of existing technology. The authors leveraged the NCBO BioPortal, a well-established, open-source platform for biomedical ontologies. This involved:

  • Re-platforming: Deploying the NCBO BioPortal's virtual appliance, which contains the complete code and deployment environment, to host agronomic ontologies.
  • Customization and Extension: While the underlying technology is domain-independent, the authors are working on specific customizations for the agronomic domain. This includes organizing content, enhancing multilingual support, interconnecting with the BioPortal, improving annotation scoring, and supporting various ontology formats relevant to agriculture (e.g., SKOS).
  • Semantic Web Compliance: Ensuring the platform adheres to semantic web standards, storing ontologies, mappings, and annotations in an RDF triple store for robust data management and interoperability.
  • API Development: Providing REST web service and SPARQL endpoints to allow programmatic access and integration with other systems.

Datasets

The paper does not explicitly list specific datasets used for training or evaluation in the traditional machine learning sense. Instead, the "datasets" are the agronomic ontologies themselves. The AgroPortal aims to host and manage a growing collection of these ontologies.

  • Initial Content: The prototype platform (v1.0 beta) hosts 49 ontologies, including 28 not originally present in BioPortal.
  • Candidate Ontologies: An additional 37 candidate ontologies are being worked on.
  • Existing Resources: The paper mentions several prominent agronomic vocabularies and ontologies that are either hosted or intended to be hosted, such as AGROVOC, NAL Thesaurus, CAB Thesaurus, Crop Ontology, and Planteome.

Key Insights and Ideas

  • Leveraging Biomedical Success: The paper highlights the significant value of reusing mature, well-tested technologies from adjacent scientific domains (biomedicine) rather than reinventing the wheel for agronomy. This accelerates development and ensures robustness.
  • Domain Independence of NCBO Technology: The NCBO BioPortal's underlying technology is shown to be adaptable to different scientific fields, demonstrating its versatility and potential for broader impact.
  • Bridging the Gap: AgroPortal aims to create a crucial bridge between the agronomic and biomedical communities by offering similar tools and services, facilitating cross-domain data integration and research.
  • Community-Driven Development: The platform is designed to be open and generic, encouraging user participation in uploading content, commenting, and contributing to the ontology ecosystem.
  • Focus on Interoperability: By supporting semantic web standards and providing APIs, AgroPortal directly addresses the critical need for data interoperability in agronomic research.

Key Results

The paper presents a functional advanced prototype platform (v1.0 beta) that is live and accessible. Key results include:

  • Platform Deployment: The AgroPortal is operational and hosts a significant number of agronomic ontologies.
  • User Adoption: The platform has already registered 38 users and is being actively used in four driving agronomic use cases.
  • Feature Set: The portal offers core BioPortal functionalities like ontology search, annotation, versioning, visualization, and community support, adapted for agronomy.
  • Interoperability Endpoints: The availability of REST and SPARQL endpoints enables programmatic access and integration.
  • Use Case Integration: The platform is already supporting specific projects and working groups, such as Agronomic Linked Data (AgroLD), RDA Wheat Data Interoperability (WDI), INRA Linked Open Vocabularies (LovInra), and the Crop Ontology project.

Limitations and Future Work

  • Thematic Boundaries: The authors acknowledge that the thematic boundaries of the portal are not precisely defined yet, noting that agriculture also includes animal sciences. Future work will involve users expressing their needs to shape the portal's scope.
  • Ontology Formats: While the platform supports formats like SKOS, further work is needed to fully accommodate the diverse range of ontology formats used in agronomy.
  • Metadata Enrichment: The authors plan to conduct an extensive description of metadata for each ontology to improve comprehension of the agronomical ontology landscape.
  • User Interface Improvements: While the prototype is functional, the paper mentions that existing applications like Crop Ontology require improvements in versioning, curation, multilingual aspects, and user interface. AgroPortal aims to address these.
  • Expanding User Base: Future work will focus on identifying more potential users and supporting new research scenarios.

Summary updated

BibTeX
@inproceedings{jonquet2016reusing,
  author = {Clément Jonquet and Anne Toulet and Elizabeth Arnaud and Sophie Aubin and Esther Dzalé Yeumo Kaboré and Vincent Emonet and John B. Graybeal and Mark A. Musen and Cyril Pommier and Pierre Larmande},
  title = {{Reusing the NCBO BioPortal Technology for Agronomy to Build AgroPortal}},
  booktitle = {ICBO/BioCreative},
  year = {2016},
}
Figure