dsSwissKnife

dsSwissKnife enables privacy-preserving federated statistical analysis of biomedical datasets to support multi-cohort studies and biomarker discovery.

Key Features:

• R package implementation: Implemented as an R package that complements and extends DataSHIELD's analytical functions.
• Integration with DataSHIELD: Extends DataSHIELD to perform additional federated analyses within the DataSHIELD framework.
• Privacy-preserving federated analysis: Performs statistical analyses without requiring direct access to raw individual-level data at participating sites.
• Support for harmonized data: Operates on harmonized datasets and supports data standardized using the CDISC SDTM format.
• Advanced statistical functionality: Provides additional analytical methods for complex statistical analyses across distributed datasets.

Scientific Applications:

• Biomarker discovery: Enables identification of potential biomarkers from harmonized multi-cohort datasets, as applied in RHAPSODY.
• Multi-cohort type 2 diabetes studies: Applied to harmonize and analyze data from 10 observational cohorts related to type 2 diabetes.
• Large-scale collaborative research: Supports federated analyses across multiple sites for multi-center or international studies without centralizing raw data.

Methodology:

Extends DataSHIELD to perform federated statistical analyses on harmonized datasets standardized to the CDISC SDTM format while avoiding transfer of raw individual-level data.