The German Biobank Node (GBN) and the German Biobank Alliance (GBA) have established an IT infrastructure that connects biobanks and enables a cross-biobank search for samples and associated data. With ”The journey to establishing an IT-infrastructure within the German Biobank Alliance“, representatives of the IT team of GBN and GBA have now published a description of this process in the renowned journal PLoS ONE. In it, they discuss the challenges of the project, lessons learned and particular successes. “We have gained valuable insights that can be of great use for similar projects,“ first author Christina Schüttler and her colleagues explain in their publication.

A federated network for central search queries

The IT infrastructure of GBN/GBA was built by a team of developers at eleven locations between 2017 and 2020. The project was managed by two development centres, one at the German Cancer Research Center (DKFZ) in Heidelberg and the other at the Friedrich Alexander University Erlangen-Nuremberg. GBN was responsible for coordination.

The infrastructure consists of a federated network of so-called bridgeheads – data integration servers that are established and filled with data by each biobank. This infrastructure forms the basis for the “Sample Locator“, which researchers can use to search for samples and associated data across biobanks. A detailed description of the IT infrastructure, its components and the underlying technology is the subject of another publication, which is currently in preparation led by the DKFZ Heidelberg.

Taking stakeholder requirements into account

In “The journey to establishing an IT-infrastructure within the German Biobank Alliance“, the authors emphasise the importance of stakeholder engagement for their work. Considering the requirements of different stakeholders, especially potential users, and testing realistic use cases with exemplary search queries are indispensable for the development of IT tools. “Another important aspect is user-friendliness. A search tool like the Sample Locator should be as intuitive to use as possible – this increases its acceptance. That's why we conducted several evaluations and adapted the interface accordingly,“ says Dr. Björn Kroll, one of the two senior authors.

Build on developments while staying flexible

For the design of local components, the developers were able to build on existing IT solutions. The “bridgehead“ architecture of the German Consortium for Translational Cancer Research (DKTK) was adopted as a central element. To harmonise the heterogeneous data, the developers were also able to draw on existing terminologies and standards, such as MIABIS (Minimum Information About BIobank data Sharing) and SPREC (Standard PREanalytical Code). Based on this, they agreed on a core data set for donor data, sample data and disease-specific data.

In addition to using existing resources, flexibility played a role in the success of the IT developments. Contrary to the original planning, the team decided to adapt to HL7®FHIR®, as this was becoming more and more popular as a transfer standard for healthcare data at that time. Thus, within the Medical Informatics Initiative (MII), all university hospitals agreed to use this standard. For the sake of interoperability, GBN/GBA therefore also chose HL7®FHIR®.

Coordinating with other networks

Of particular importance was the cooperation with related initiatives at national and international level in order to achieve interoperability and avoid parallel structures. An exchange took place both with the European biobank organisation BBMRI-ERIC and the German MII. The latter resulted in the follow-up project ABIDE_MI (Aligning Biobanking and DIC Efficiently) in 2021, for which the developments of GBN/GBA are successfully used.

Publish open source, facilitate installation

“To make our developments sustainable, we make them available open source,“ says Dr. Cecilia Engels, second senior author of the publication. “Interested parties can thus build on and use our IT components.” In addition to this easy access for developers, the IT team also optimised the installation process for participating biobanks. ”Not all biobanks have their own IT staff. That is why we provide our tools in a lean 'container version', including a detailed implementation guide,” says Christina Schüttler. Within a short time, 14 biobanks could already be connected to the Sample Locator in this way.

Further information and links:

• Schüttler C, Prokosch H-U, Hummel M, Lablans M, Kroll B, Engels C (2021) The journey to establishing an IT-infrastructure within the German Biobank Alliance. PLoS ONE 16(9):e0257632. https://doi.org/10.1371/journal.pone.0257632
• IT infrastructure
• Sample Locator
• Evaluation of the Sample Locator
• Implementation guide
• ABIDE_MI