Work package 2: Secure cloud-based data store
Lead: Reykjavík University
The goal of this work package is to build secure cloud-based facilities to store and analyse data related to the project.
The work in this work package can be divided in to four parts.
Task 2.1: Generate a secure cloud-based data store
In this task, we will construct a computer infrastructure for all the research data in the project. The equipment provides a secure, access-controlled storage, including a high-performance GPU-based cluster resource for researchers to process massive amounts of data and train deep neural networks. This includes providing an access-controlled repository for retrospective data (WP3) and prospective data (WP6-10). This infrastructure is based at RU and will be maintained and updated throughout the project.
Task 2.2: Integration with the European Open Science Cloud
We intend to actively participate as an early adopter of the European Open Science Cloud (EOSC) initiative throughout the project, and provide credentialed access using the EOSC for all members of the project.
Task 2.3: Online scoring and algorithm analysis
We will construct an authorisation system to manage access to the data so that researchers can readily obtain permissions to work directly on specific datasets related to their research. This system supports all General Data Protection Regulation (GDPR) rights such as medical subjects right to withdraw their permission and request data destruction. Thereby, we will not only address existing security concerns for tracking access and copying of large, medically sensitive datasets, but also allow researchers to more readily share and validate data analysis, whilst working simultaneously in different locations with the same sleep study scoring and algorithms (WP 4-5). As a result, other research participants do not need a dedicated and expensive computer infrastructure at their own institution but can work directly with the SLEEP REVOLUTION infrastructure instead, with access levels to different data sets. This structure also allows for the data management plan and public data sharing as planned in WP1.
Task 2.4: Access right integration and monitoring for testing phase
We will provide a novel computer architecture, within a virtualised framework that integrates and monitors the access rights for research subjects, researchers with different access levels, and health care professionals for relevant data needed in the clinical phase of the project (see further in regards to the data management platform in WP7). This task expands the storage from sleep studies and health outcomes, to included data from the PROMs, wearables, and objective daytime functioning testing. It will provide a shared common resource for algorithms developed for the project, allowing all project partners to use these data processing tools. Any use will be validated, improved, and tracked with appropriate source control, within an automated testing and verification environment.
University Of Eastern Finland
Nox Medical Ehf