CIDO: The community-based coronavirus infectious disease ontology

Yongqun He, Hong Yu, Edison Ong, Yang Wang, Yingtong Liu, Anthony Huffman, Hsin-Hui Huang, John Beverley, Yu Lin, William D. Duncan, Sivaram Arabandi, Jiangan Xie, Junguk Hur, Xiaolin Yang, Luonan Chen, Gilbert S. Omenn, Barry SmithPhilPapers 2020

TL;DR

The COVID-19 pandemic has highlighted the urgent need to integrate vast amounts of heterogeneous coronavirus data. This paper introduces CIDO, a community-based ontology designed to standardize and integrate knowledge about coronaviruses, their associated diseases, treatments, and medical devices, enabling more effective research and development.

What Problem Does This Paper Address?

The global health crises caused by coronaviruses like SARS, MERS, and most recently COVID-19, have generated an overwhelming amount of complex and diverse data. This data is often siloed in non-integratable and non-interoperable repositories, creating a significant bottleneck for research. This lack of integration hinders our ability to understand these viruses and develop effective countermeasures like vaccines and drugs. The paper addresses this by proposing CIDO, an ontology to provide a standardized, computer-interpretable framework for coronavirus-related knowledge.

Methods

The development of CIDO follows the principles of the Open Biomedical Ontology (OBO) Foundry and an eXtensible Ontology Development (XOD) strategy. Key methods include:

  • Data Collection: Gathering data from literature and publicly available databases, focusing on coronavirus taxonomy, host data, phenotypes, drugs, and vaccines.
  • Ontology Development: Reusing and aligning terms and relations from existing ontologies (e.g., Basic Formal Ontology (BFO), Human Disease Ontology (DOID), Human Phenotype Ontology (HPO), Infectious Disease Ontology (IDO)) using tools like Ontofox and editing with Protégé-OWL.
  • Design Patterns: Employing specific ontological design patterns to logically represent complex relationships, such as the "COVID-19 disease process" and its various attributes and causal factors.
  • Querying: Utilizing Description Logic (DL) and SPARQL queries for ontology content exploration and analysis.

Datasets

The paper does not explicitly list specific datasets used for training or evaluation in the traditional machine learning sense. Instead, it describes the sources of data used to build the ontology:

  • Literature and openly available databases related to coronaviruses.
  • Existing ontologies such as:
  • Basic Formal Ontology (BFO)
    • Chemical Entities of Biological Interest (ChEBI)
    • Human Disease Ontology (DOID)
    • Human Phenotype Ontology (HPO)
    • Infectious Disease Ontology (IDO)
    • NCBI Taxonomy Ontology (NCBITaxon)
    • Ontology for Biomedical Investigations (OBI)
    • Ontology for General Medical Science (OGMS)
    • Ontology of Host-Pathogen Interaction (OHPI)
    • Ontology of Precision Medicine and Investigation (OPMI)
    • Vaccine Ontology (VO)
    • Ontology of Adverse Events (OAE)
    • National Drug File Reference Terminology (NDF-RT)
    • Drug Ontology (DrON)

Key Insights and Ideas

  • Ontologies as a Solution to Data Silos: The paper strongly advocates for ontologies as a crucial tool to overcome the fragmentation of biomedical data, enabling integration and interoperability.
  • Community-Driven Development: CIDO is designed as a "community-based" ontology, emphasizing open access and encouraging community involvement for its ongoing development and improvement.
  • Standardized Representation of Complex Entities: CIDO provides a structured way to represent diverse entities related to coronavirus diseases, from the viruses themselves and their phenotypes to drugs, vaccines, and medical devices like ventilators.
  • Drug Repurposing Potential: The ontology can be used to identify potential drug candidates for repurposing by querying for compounds with specific roles (e.g., antiviral, antimalarial, anti-inflammatory).
  • Modeling Disease Processes: Representing "COVID-19 disease process" as a distinct entity, rather than just a disposition, allows for more nuanced modeling of its occurrence, causes, and interactions with treatments.

Key Results

  • CIDO Development: The paper presents the initial development and architecture of CIDO, which has been accepted as an OBO library ontology.
  • Ontological Representation: Demonstrations of CIDO's ability to represent coronaviruses, host phenotypes, anti-coronavirus drugs, and medical devices (specifically ventilators) are provided.
  • Drug Repurposing Application: A use case demonstrates how CIDO can be used to identify potential drug candidates for COVID-19 treatment. A DL query identified six chemical compounds with antiviral and antimalarial roles, with amodiaquine also possessing anti-inflammatory properties.
  • Statistics: As of August 30, 2020, CIDO contained 6,758 terms, including imported terms from over 40 ontologies, and 190 CIDO-specific terms.

Limitations and Future Work

  • Limitations:
    • The paper focuses on the initial development and architecture of CIDO, with ongoing work planned.
  • While drug repurposing is demonstrated, the full scope of its application is still being explored.
  • Future Work:
    • Modeling Mechanisms: Modeling the intricate mechanisms of molecular and cellular interactions between hosts and coronaviruses to further support drug repurposing and vaccine development.
  • Standardization: Using CIDO as a platform to standardize metadata types for coronavirus research and enhance the analysis of specific experimental and clinical study conditions.
  • Application Development: Identifying and developing more application tools that implement CIDO for various purposes.
  • Community Engagement: Encouraging broader community involvement in the development and critique of the ontology.

Summary updated

Abstract

Current COVID-19 pandemic and previous SARS/MERS outbreaks have caused a series of major crises to global public health. We must integrate the large and exponentially growing amount of heterogeneous coronavirus data to better understand coronaviruses and associated disease mechanisms, in the interest of developing effective and safe vaccines and drugs. Ontologies have emerged to play an important role in standard knowledge and data representation, integration, sharing, and analysis. We have initiated the development of the community-based Coronavirus Infectious Disease Ontology (CIDO). As an Open Biomedical Ontology (OBO) library ontology, CIDO is an open source and interoperable with other existing OBO ontologies. In this article, the general architecture and the design patterns of the CIDO are introduced, CIDO representation of coronaviruses, phenotypes, anti-coronavirus drugs and medical devices (e.g. ventilators) are illustrated, and an application of CIDO implemented to identify repurposable drug candidates for effective and safe COVID-19 treatment is presented.
BibTeX
@inproceedings{he2020cido,
  author = {Yongqun He and Hong Yu and Edison Ong and Yang Wang and Yingtong Liu and Anthony Huffman and Hsin-Hui Huang and John Beverley and Yu Lin and William D. Duncan and Sivaram Arabandi and Jiangan Xie and Junguk Hur and Xiaolin Yang and Luonan Chen and Gilbert S. Omenn and Yongqun He and Barry Smith},
  title = {{CIDO: The community-based coronavirus infectious disease ontology}},
  booktitle = {PhilPapers},
  year = {2020},
}
Figure