LIV Contributes to the Expansion of AI Infrastructures Across the Leibniz Association

Modern life sciences generate vast amounts of data, ranging from genome sequences and metabolic profiles to high-resolution imaging data. Artificial intelligence (AI) offers new possibilities for analysing these data and identifying complex biological relationships. As one of the six key technologies of Germany’s High-Tech Agenda, AI is of central importance to the future competitiveness of Germany as a research and innovation hub. Harnessing this potential requires powerful research infrastructures. The Leibniz Association is expanding its capacities in this area. The Leibniz Institute of Virology (LIV) is involved in two of the new infrastructure initiatives, contributing its expertise in data-intensive virology research.

Analysing large and complex datasets using AI requires substantial computing power. Graphics processing units (GPUs) are particularly well suited to AI applications because they can perform many computational operations simultaneously. In a GPU cluster, numerous processors are interconnected to form a powerful computing system. This enables the training of large AI models and the analysis of extensive biological datasets within a reasonable timeframe.

Developing AI infrastructures for the life sciences together

To meet these growing computational demands, the LIV is involved in an initiative coordinated by the Leibniz Institute of Plant Biochemistry (IPB) in Halle (Saale) to develop a high-performance AI infrastructure. At the heart of the planned infrastructure is a high-performance GPU cluster designed to train state-of-the-art AI models for life sciences. The aim is to link large and highly diverse datasets from genomics, proteomics, metabolomics, imaging and environmental research and to use AI to identify biological patterns and relationships. The underlying datasets may comprise several petabytes, equivalent to several million gigabytes.

The LIV is also participating in another Leibniz-wide initiative in the fields of health and prevention research lead by the Leibniz Institute of Analytical Sciences – ISAS in Dortmund.

For the LIV, these collaborations provide an opportunity to combine its own expertise and infrastructures with the capacities of other Leibniz institutions and to expand its computing resources significantly. This will further accelerate the development of Viral Data Science, a key strategic area of research at the Leibniz Institute of Virology.

Infrastructure at the LIV

The new initiatives are part of the LIV’s ongoing efforts to expand its infrastructure for data-intensive research. A key element of these efforts is the Scientific Computing technology platform, through which the institute continuously develops and expands its computing capabilities. The platform includes an in-house high-performance computing (HPC) cluster, GPU-accelerated applications, and tools for big data, cloud computing and AI.

For example, the early development of the in-house HPC already supports a range of research activities, including molecular simulations, the analysis of sequence and structure data, and the training of AI models. Moreover, as AI becomes increasingly integrated into research, the LIV is also addressing the organisational and ethical aspects of its use. An institute-wide AI committee has developed a dedicated policy, while internal and external training sessions and workshops provide guidance on topics such as data protection, transparency and bias. To facilitate the practical use of AI, the LIV also provides researchers and staff with a dedicated, secure closed system (Chat AI) for scientific and research-related tasks.

Learn more about the LIV’s activities as part of Germany’s High-Tech Agenda here.