Understanding ecological systems using knowledge graphs: an application to highly pathogenic avian influenza.

Area of Science:

• Ecology
• Computational Biology
• Epidemiology

Background:

• Ecological systems present complex data integration challenges due to heterogeneity across subdisciplines and sources.
• Knowledge graphs (KGs) effectively organize disparate data and predict new linkages in complex systems.
• KGs offer significant potential for ecological research, especially amidst rapid environmental changes.

Purpose of the Study:

• To demonstrate the utility of KGs for ecological problems using highly pathogenic avian influenza (HPAI) as a case study.
• To develop a KG integrating HPAI-related data, including pathogen-host associations, species distributions, and population demographics.
• To validate the KG method and identify ecological patterns of HPAI.

Main Methods:

• Developed a semantic ontology to define relationships within and between ecological datasets.
• Constructed a knowledge graph incorporating pathogen-host associations, species distributions, and population demographics for HPAI.
• Performed proof-of-concept analyses on the KG to validate the methodology and uncover ecological patterns.

Main Results:

• The developed KG successfully integrated heterogeneous data relevant to HPAI ecology.
• Analyses using the KG revealed patterns in HPAI dynamics and validated the KG approach.
• The study demonstrated the generalizable value of KGs in ecology for uncovering relationships and generating hypotheses.

Conclusions:

• Knowledge graphs provide a powerful framework for organizing complex ecological data and advancing mechanistic understanding.
• The KG approach is broadly applicable to ecological research, aiding in the discovery of previously unknown relationships and testable hypotheses.
• This work highlights the potential of KGs to deepen our understanding of ecological systems facing rapid change.