MOLGENIS

MOLGENIS enables integration, harmonization, and model-driven generation of biomedical data management and analysis infrastructures to support pooled analyses of omics and clinical datasets.

Key Features:

• Data Integration and Pooling: MOLGENIS/connect semi-automatically pools data from multiple sources such as biobanks and patient registries, using ontology-based query expansion to identify relevant source attributes.
• Algorithmic Mapping and Transformation: The system generates algorithms that transform source attributes into a common target DataSchema, handling unit conversion, categorical value matching, and complex calculations such as BMI.
• Model-Driven Development: A simple modeling language defines biological data structures which the generator suite translates into databases, exchange formats, and scriptable interfaces.
• Customization and Scalability: A plug-in mechanism allows customization of the generator suite and generated products and supports rapid prototyping for next-generation sequencing, GWAS, QTL analysis, proteomics, and biobanking applications.
• Development Efficiency: Models typically of around 500 lines of XML replace thousands of lines of hand-written code, enabling reuse of modeling languages and generators across applications.
• Optimized Back-end and Interfaces: Each MOLGENIS application includes an optimized database back-end and provides programming interfaces in R and Java, with SOAP, REST/JSON, and RDF endpoints and a tab-delimited file format for data exchange.

Scientific Applications:

• Data pooling and harmonization: Pooling and harmonization of cohort, biobank, and patient registry data to increase statistical power in multi-center and retrospective studies.
• Genomics and omics analyses: Support for next-generation sequencing, GWAS, QTL analysis, and proteomics data management and prototyping.
• Biobanking and cohort management: Integration and standardized exchange of biobanking datasets and associated clinical metadata.
• Automated data transformation: Generation of attribute-mapping algorithms and standardized exchange formats for downstream bioinformatics workflows.

Methodology:

MOLGENIS/connect performs semi-automatic pooling using ontology-based query expansion; it generates attribute-mapping algorithms that perform unit conversions, categorical value matching, and calculations (e.g., BMI) to map to a target DataSchema; a model-driven approach uses a simple modeling language and a generator suite to produce databases, exchange formats, and scriptable interfaces, with customization via plug-ins and programmatic access through R, Java, SOAP, REST/JSON, RDF and a tab-delimited file format.