ADOPT BBMRI-ERIC

ADOPT BBMRI-ERIC harmonizes colorectal cancer clinical and biobank data across European biobanks to enable standardized integration and access to high-quality clinical and specimen data for research.

Key Features:

• Data Harmonization and Integration: Semantic mapping aligns local biobank terminologies with a centralized ADOPT BBMRI-ERIC terminology to produce a common data representation.
• Lexical Bag-of-Words Matcher: A semiautomatic lexical bag-of-words matcher performs term mapping using fuzzy matching, synonym utilization, and sentiment tagging.
• Data Transformation Application: A transformation application processes anonymized instance data according to established mappings to produce integrated datasets.

Scientific Applications:

• Cross-biobank harmonization: Applied across 10 European biobanks to map 1,492 local terms with an automatic mapping accuracy of 78.48%, with remaining mappings completed by human experts.
• Colorectal cancer data integration: Harmonized and processed 147,608 data records from 3,415 patients to create a unified dataset for colorectal cancer research.
• Generic harmonization approach: Establishes a standardized data representation to support cross-institutional research collaborations.

Methodology:

Semantic mapping using metadata repository semantics with semiautomated human expert refinement; a lexical bag-of-words matcher employing fuzzy matching, synonym utilization, and sentiment tagging; and a data transformation application that processes anonymized instance data based on the mappings.