Pan-European Data Harmonization for Biobanks in ADOPT BBMRI-ERIC

Area of Science:

• Biobanking and Biomedical Research
• Data Science and Informatics

Background:

• High-quality clinical data and biological specimens are crucial for advancing medical research and personalized medicine.
• The Biobanking and Biomolecular Resources Research Infrastructure-European Research Infrastructure Consortium (BBMRI-ERIC) was established to improve access to these vital resources.
• The ADOPT BBMRI-ERIC project initiated BBMRI-ERIC by gathering colorectal cancer data from European biobanks.

Purpose of the Study:

• To develop a uniform approach for data integration and harmonization, transforming diverse biobank data into a common representation.
• To design and implement a toolset capable of achieving this data harmonization.

Main Methods:

• Developed a lexical bag-of-words matcher for semi-automatic mapping of local biobank terms to a central terminology, incorporating fuzzy matching, synonyms, and sentiment tagging.
• Created a data transformation application to process anonymized instance data based on the established mappings.
• Utilized a metadata repository's semantics to guide the mapping process.

Main Results:

• The toolset processed data from 10 European biobanks.
• The lexical matcher achieved an automatic and correct mapping rate of 78.48% for 1,492 local biobank terms.
• Expert curation completed the remaining mappings, enabling successful processing of 147,608 data records from 3,415 patients.

Conclusions:

• A generic and successful cross-institutional data harmonization approach was created for 10 European biobanks.
• The developed software tools for data harmonization are available as open source, promoting wider adoption and collaboration.