High Performance Computing-Platform (HPC)

The DRFZ High-Performance Computing (HPC) Platform is the institute’s central computational hub, providing the scalable resources necessary to accelerate discovery across multi-omics, advanced microscopy, and high-throughput translational immunology research.

Infrastructure and Strategic Partnership
Our strategic hybrid computing cluster is hosted in partnership with the Zuse Institute Berlin (ZIB), ensuring high availability and professional sustainable maintenance.

• Dual Cluster Access: Researchers access two distinct computational tiers: a high-throughput general cluster and a private, high-security DRFZ-exclusive cluster optimized for processing sensitive data.
• Integrated User Environment: The system enhances researcher productivity through seamless web-based interfaces to key development environments, including RStudio, Python Notebooks (Jupyter/JupyterLab), and Visual Studio Code (VS Code).

Services and Computational Expertise
Beyond the hardware, the HPC Platform team operates as a group of computational biologists and data engineers, providing essential infrastructure support alongside high-level collaborative analysis services.

• Collaborative Bioinformatics and Data Analysis: We partner with research groups to develop and execute advanced mathematical and ML/AI analyses, supporting diverse biological fields such as rheumatology, systems medicine, neuroscience and microbiology. Our technical depth includes dominant programming languages such as Python, R, C++, MATLAB, Nextflow, and Snakemake.
• Custom Tool Development: We actively create custom software applications, including specialized diagnostic tools for medical doctors and bespoke GUIs, ensuring practical research outputs.
• Agentic AI Platform for Multi-Omics:As a prime example of our custom engineering, we have developed a proprietary agentic AI ecosystem to autonomously analyze the complex multi-omics datasets generated by the Robotics Facility. This initiative positions the DRFZ at the avant-garde of computational immunology, bridging the gap between high-throughput data and automated clinical insight.
• Reproducibility & Standards: We enforce workflow standardization, through resources such as Nextflow and containerization, to ensure all research is consistent, reproducible, and compliant with FAIR Data principles.
• Training and Teaching: Training and Teaching: Mentorship and computational literacy are central to our mission. We welcome trainees from bachelor’s students to postdocs, having already successfully trained several interns who actively contributed to collaborative projects. Our scientific team also provides extensive 1:1 bioinformatics support and troubleshooting to researchers across the DRFZ and our liaison groups (e.g., Charité, Zuse Institute).

Innovation and Data Modernization
The HPC Platform drives innovation by implementing next-generation data standards and tools across the DRFZ:

• Robotics Project Highlight: The team develops and supports the core data analysis pipelines for diagnostic products emerging from the DRFZ’s Robotics Facility (Kitanovic) and AG Latz merger venture. Lab process engineering, automation and modernization is a big aspect of our work.
• Next-Generation Biotech Automation: Supported by the BMBF DATIpilot grant (2025), we are developing a Digital Twin of the Robotics facility for advanced protocol simulation. Concurrently, we are implementing blockchain technology to establish immutable audit trails, elevating process documentation and data provenance to cryptographic standards.
• In-House LLM Deployment: In collaboration with the Zuse Institute Berlin, we are implementing secure, localized Large Language Models (LLMs) for all DRFZ users. This provides a local, private AI environment for all staff (administrative and wet lab) to make use of these tools without compromising institutional data security.
• Scientific Data Lake Initiative: We are currently establishing a comprehensive, institute-wide scientific data lake with an integrated platform to facilitate streamlined data access and discovery.
• EU IHI AURORA Consortium: Central Hub & Federated Learning In partnership with Prof. Dr. Gerhard Krönke, the HPC Platform serves as the central data management hub for the pan-European IHI AURORA consortium. As the lead for Work Package 2, our platform is overseeing the deployment of federated learning architectures across the EU. This infrastructure enables international institutions to collaboratively train models on complex multi-omics clinical datasets, driving predictive immune profiling while strictly maintaining decentralized data privacy.