Imagene-DEAP: Imaging and Genetics Data Exploration and Analytics Portal

20.02.2024

The Imagene-DEAP Advanced User Support (AUS) project was started in the late autumn of 2023 as an effort to establish a portal for imaging and genetics data exploration and analytics.

Large-scale datasets containing sensitive, individual-level genetic and imaging information are currently being collected, stored, processed, and analyzed on secure research infrastructures.

Illustration of human silhoutte with abstract elements.

These trusted infrastructures are located at various institutions and utilize high-performance storage and computing facilities. This necessitates the use of (pseudo-)federated solutions for analysis and model development, particularly when dealing with non-shareable data like individual-level genomics.

A Comprehensive Approach to Sensitive Data Management and Analysis

One component of Imagene-DEAP involves offering a nationwide, accessible catalogue and metadata service for sensitive data. This service is integrated with the REDCap database software, which is deployed at each site to facilitate data registration and the provision of non-sensitive metadata. This service is hosted by Sigma2.

Bilde
Illustration of federated analysis in different HPC environments
Figure 1: Illustration of federated analysis in different HPC environments.

Another component is providing a portal for initiating primary analyses to be run on sensitive datasets located in separate facilities. This ensures consistent software environments through software containers distributed to each site. Output data, such as non-sensitive summary statistics can be exported, combined with similar output computed in the other sites, and used for final meta-analyses by the core analyst(s), as illustrated in Figure 1.

The last component is to provide an online catalogue and metadata service and analytics portal combining aspects of the other components, for user authentication, presenting metadata, initiation of distributed analyses, retrieval, and meta-analysis of the results. 

Results

Imagene-DEAP may improve our position to facilitate novel discoveries in clinical research by exploiting the large, distributed datasets and at the same time improve research data management in Norway by establishing a solution for the archival of multimodal and sensitive data, starting with the Sensitive Data Services (TSD), HUNT Cloud and SAFE. The project is still in an early stage, where TSD/USIT personnel aid in planning, investigating and incorporating the technical requirements for the project, such as deployment of tools.

About the project

The project is a collaboration with the Precision Psychiatry group (PI: Ole A. Andreassen, UIO and OUS), NRIS (TSD/USIT, UIO) and Sigma2; with support from SAFE (UIB), HUNT Cloud (NTNU), and ELIXIR Norway.