Linking of research data
Within the laboratory workstream, it was important to integrate and link the data of the different involved consortium partners and stakeholders, as well as link public data sets. Linking of data within big consortia is needed in order to break the research silos and to facilitating the sharing of data, insights, and results. Ontoforce aimed to link and integrate research data in their DISQOVER platform with the goal to improve data integration for large and various data sets utilized in the PRISMA project (Figure 13).
Within PRISMA, several aspects have been the focus: improve platform scalability and interoperability of data ingestion, improve distribution mechanisms and coverage of public data assets, develop a plugin framework to improve user journey, connecting DISQOVER with third party platforms.
Four specific activities were performed in function of this goal:
1) Optimizing data integration. This includes (i) developing a new user interface (UI) or the data ingestion platform, allowing for better handling of large size and number of datasets and improving complexity, (ii) improving disk space footprint and memory ingestion speed.
2) Optimizing data sharing by developing data synchronization “Remote Data Subscription” technology and out-of-the box public data packages. This includes i) creating the Remote Data Subscription technology that allows sharing data from one DISQOVER system to another, (ii) optimizing this technology on disk space, memory and UI usability, and (iii) creating different data packages with this technology for different public data sources, allowing PRISMA to subscribe to these.
3) Optimizing data mining by developing plugins. This includes (i) developing a new functionality in the platform allowing python-customized plugins (i.e. customized features) which allow new “downstream analytics” features and (ii) with this functionality creating customized plugins which sent a finalized result set of DISQOVER data to a analytical application.
4) Consortium partner data collection, integration and visualization. This includes (i) research of different FAIR data representations and common data models (FHIR, OMOP, ...) (ii) setup of DISQOVER server with the latest version, (iii) stakeholder interviews to understand their data, patient identifiers used and which data is suitable to upload in the platform, (iv) share and integrate data the stakeholder data in DISQOVER and (v) create usable dashboards for the stakeholders to explore the integrated and linked data. The Stakeholders received accounts and training on how to use the platform.
Several technical and functional improvements were made to the DISQOVER platform during the PRISMA project such as improved data provenance, data synchronization speed improvements, plugin development, and the embedding of Remote Data Subscription technology in the platform. Moreover, several PRISMA specific achievements were reached during the project. We were able to succesfully integrate consortium partner data within the platform. Furthermore, gene-association data packages were successfully integrated at scale for KU Leuven and VIB target identification use cases and visualizations for public data packages, e.g., showing correlations between AD and specific gene-related biomarkers, were made possible.
Ontoforce blog: DISQOVER supports the PRISMA consortium in the development of precision medicine for Alzheimer’s - https://www.ontoforce.com/blog/disqover-supports-the-prisma-consortium-for-alzheimers
Ontoforce blog: Ontoforce blog: Improving target identification with a knowledge graph - https://www.ontoforce.com/blog/improving-target-identification-with-a-knowledge-graph
Ontoforce webinar: Optimize your multi-omics research: data management challenges and best practices - https://www.ontoforce.com/webinar/optimize-your-multi-omics-research-data-management-challenges-and-best-practices
Ontoforce webinar: Disqover for integrated omics data - https://www.ontoforce.com/webinar/disqover-for-integrated-omics-data
Ontoforce webinar: How to accelerate biomarker identification with DISQOVER - https://www.ontoforce.com/webinar/how-to-accelerate-biomarker-identification-with-disqover