ViPAR: a software platform for the Virtual Pooling and Analysis of Research Data.

Area of Science:

• Bioinformatics
• Computational Biology
• Health Informatics

Background:

• Research on disease determinants requires robust statistical power, often achieved by pooling data from multiple sites.
• Centralized data pooling faces significant ethical, privacy, and data ownership hurdles.
• There is a critical need for methods that facilitate secure research data sharing and flexible analysis.

Purpose of the Study:

• To introduce ViPAR (Virtual Pooling and Analysis of Research data), a novel software platform.
• To provide a web-based solution for analyzing datasets housed in disparate locations.
• To overcome challenges associated with traditional centralized data pooling.

Main Methods:

• ViPAR utilizes database federation and the Secure Shell protocol for secure data exchange.
• It enables 'virtual pooling,' where data is temporarily pooled in memory for analysis.
• Remote sites manage their own harmonized datasets, with a central server handling data federation and analysis.

Main Results:

• The ViPAR platform facilitates secure retrieval and virtual pooling of requested data from remote sites.
• Analysis is performed on the virtually pooled data using statistical software.
• Results are returned to the user, and the temporarily pooled data is removed from memory upon completion.

Conclusions:

• ViPAR is a secure, flexible, and powerful analysis platform built on open-source technology.
• The platform is currently utilized by large international consortia.
• ViPAR is publicly available to facilitate collaborative research.