SEA CDM: Study-Experiment-Assay Common Data Model and Databases for Cross-Domain Data Integration and Analysis

Area of Science:

• Biomedical Informatics
• Immunology
• Data Science

Background:

• Increasing biomedical data volume presents challenges in standardization and integration.
• Heterogeneous experimental data across domains requires robust solutions for sharing and analysis.
• Existing data models may not adequately support cross-domain integration and knowledge inference.

Purpose of the Study:

• To develop an ontology-supported common data model (CDM) for standardizing and integrating biomedical experimental data.
• To build a relational database and knowledge graph (OSEAN) based on the SEA CDM for enhanced data analysis.
• To apply the developed system to large-scale immune study datasets for scientific discovery.

Main Methods:

• Developed the Study-Experiment-Assay (SEA) common data model (CDM) using object-oriented modeling with 10 core and 3 auxiliary classes.
• Utilized interoperable ontologies within the SEA CDM for data standardization and knowledge inference.
• Constructed the Ontology-based SEA Network (OSEAN) relational database and knowledge graph, incorporating ETL and query tools.

Main Results:

• Successfully represented 1,278 immune studies with over two million samples from VIGET, ImmPort, and CELLxGENE using the OSEAN system.
• Identified scientific insights into sex-specific immune responses, including neutrophil degranulation and TNF binding, following influenza vaccination.
• Demonstrated the utility of simple, robust queries and analyses on the integrated data.

Conclusions:

• The novel SEA CDM system provides a foundational framework for an integrative biodata ecosystem.
• The OSEAN system facilitates data standardization, sharing, and knowledge discovery across diverse biomedical domains.
• This approach enables robust analysis of complex immunological data, revealing important sex-specific vaccination responses.