CIDO, a community-based ontology for coronavirus disease knowledge and data integration, sharing, and analysis

Yongqun He, Hong Yu, Edison Ong, Yang Wang, Yingtong Liu, Anthony Huffman, Hsin-Hui Huang, John Beverley, Junguk Hur, Xiaolin Yang, Luonan Chen, Gilbert S. Omenn, Barry SmithScientific 2020 · DOI

TL;DR

The COVID-19 pandemic has highlighted the urgent need for better ways to organize and integrate the vast amount of research data being generated. This paper introduces CIDO, a community-driven ontology designed to standardize knowledge about coronavirus diseases, enabling more effective data integration, sharing, and analysis. CIDO aims to accelerate research into understanding, diagnosing, and treating these diseases.

What Problem Does This Paper Address?

The rapid proliferation of data and research related to coronavirus diseases, particularly COVID-19, presents a significant challenge. This "big data" era is characterized by the "five V's" (volume, velocity, variety, veracity, value), leading to disintegrated and non-interoperable data. This fragmentation hinders computer-assisted reasoning and artificial intelligence applications, slowing down the discovery of effective treatments and preventive measures. Existing bioinformatics tools struggle to efficiently integrate and analyze this heterogeneous information. The paper addresses this gap by proposing CIDO, an ontology that acts as a structured vocabulary to represent and link various aspects of coronavirus diseases, from their etiology and transmission to pathogenesis, diagnosis, and treatment.

Methods

The core of this paper is the introduction and description of the Coronavirus Infectious Disease Ontology (CIDO). CIDO is built following the principles of the Open Biomedical and Biological Ontologies (OBO) Foundry, emphasizing collaboration, openness, and the use of standardized definitions. It leverages an extensible ontology development strategy and aligns with the Basic Formal Ontology (BFO), an ISO/IEC standard for upper-level ontologies. This alignment ensures CIDO's interoperability with over 300 other ontologies that also adhere to BFO. CIDO integrates existing terms from other relevant ontologies (e.g., Disease Ontology, Human Phenotype Ontology, Chemical Entities of Biological Interest) and develops new, CIDO-specific terms to address the unique needs of coronavirus research. The paper also highlights the use of logical axioms to define relationships between different ontology classes, such as linking a disease process to its causative agent.

Datasets

Not discussed in the paper. The paper focuses on the ontology itself and its potential to integrate data from various sources, rather than specific datasets used for its development or evaluation.

Key Insights and Ideas

  • Ontologies as a Solution to Data Fragmentation: The paper strongly advocates for ontologies as a critical tool to overcome the challenges posed by disintegrated and non-interoperable data in complex scientific domains like infectious diseases.
  • The Power of Interoperability: CIDO's adherence to OBO Foundry principles and its alignment with BFO are crucial for achieving interoperability, allowing it to seamlessly integrate with a vast ecosystem of existing biomedical ontologies.
  • Beyond Simple Hierarchies: CIDO goes beyond basic "is_a" relationships by incorporating a rich set of logical axioms and relations (e.g., "caused by," "susceptibility to") to capture complex biological interactions and causal pathways.
  • Accelerating Drug and Vaccine Development: By providing a standardized framework for representing knowledge about diseases, pathogens, drugs, and vaccines, CIDO can significantly aid in the rational design and repurposing of therapeutics and the development of vaccines.
  • Community-Driven Development: The emphasis on CIDO as a community-based, open-source project underscores the importance of collaborative efforts in building comprehensive and up-to-date knowledge resources for rapidly evolving scientific fields.

Key Results

The paper primarily describes the development and design of the CIDO ontology. It reports that CIDO currently contains over 4,000 terms, imported from approximately 20 existing ontologies, with additional CIDO-specific terms developed. The paper illustrates how CIDO can logically link components related to coronavirus diseases, such as diseases, viruses, drugs, and their mechanisms of action, using examples like the causal link between COVID-19 and SARS-CoV-2. It also mentions ongoing work using CIDO for analyzing anti-coronavirus drugs and designing COVID-19 vaccines, citing related pre-print papers.

Limitations and Future Work

The authors acknowledge that the rapid evolution of knowledge surrounding COVID-19 necessitates a readily updatable ontology. They plan to address this by following OBO Foundry guidelines for term requests and using issue tracking on the CIDO GitHub site. Future work includes further developing CIDO to support the fundamental understanding of host-coronavirus interactions, precision medicine research, and rational vaccine design. They also aim to extract a lightweight, independent ontology for coronavirus metadata from CIDO to support data integration and knowledge discovery. The authors actively encourage community participation in CIDO's development and applications, anticipating the development of innovative computational and statistical algorithms and tools.

Summary updated

Abstract

The Coronavirus Infectious Disease Ontology (CIDO) is a community-based ontology that supports coronavirus disease knowledge and data standardization, integration, sharing, and analysis.
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 Junguk Hur and Xiaolin Yang and Luonan Chen and Gilbert S. Omenn and Yongqun He and Barry Smith},
  title = {{CIDO, a community-based ontology for coronavirus disease knowledge and data integration, sharing, and analysis}},
  booktitle = {Scientific},
  year = {2020},
  doi = {10.1038/s41597-020-0523-6},
}
Figure