From manual entry to machine precision: challenges and evolution of metadata schema development in collaborative...

Area of Science:

• Biomedical Research
• Data Standardization
• Metadata Management

Background:

• Collaborative biomedical research requires metadata standards that balance broad interoperability with specific domain needs.
• Existing schemas often lack flexibility for diverse research areas like nephrology, tumor immunology, and perinatal immunology.

Purpose of the Study:

• To describe a parent-template approach for adapting a baseline metadata schema across different collaborative research centers.
• To demonstrate how a single schema can be modified to meet the unique requirements of distinct research domains.

Main Methods:

• A baseline schema from the Nephrology-focused CRC 1453 NephGen was used as a parent template.
• The schema was adapted for the tumor-immunology CRC OncoEscape and the perinatal-immunology CRC Pilot.
• Adaptations involved expanding level lists and incorporating new, CRC-specific query dimensions.

Main Results:

• Three structurally compatible but vocabulary-divergent schemas were generated for OncoEscape, NephGen, and Pilot.
• Granularity varied, with Pilot requiring the most levels (324), followed by NephGen (287) and OncoEscape (283).
• Vocabulary reuse from the baseline schema was limited (35-47%), with significant additions of domain-specific terms and dimensions like "Oncogenes" and "Timeline."

Conclusions:

• The parent-template approach is effective for creating interoperable yet domain-specific metadata schemas in collaborative biomedical research.
• This method facilitates AI-assisted data extraction by defining schemas as "instruction sets" with detailed specifications.
• Adaptable metadata frameworks are crucial for managing complex data across diverse research consortia.