Medical Data (Privacy)

COMSYS is invested in interdisciplinary research activities that tackle life sciences for several years. In this context, our research activities are not limited to a specific theme. Instead, we look at a wide range of research challenges (with our applied cooperation partners). In particular, we have prior experience with the following topics:

  • Interoperability: FAIRifying medical (research) data and ensuring interoperability between different processing steps are timely challenges. Addressing them promises to advance Big Data applications and boost privacy research.
  • Consent: For healthcare research, informed consent is a key principle. Unfortunately, obtaining consent is a challenging endeavor given the lack of usable and accepted solutions in the area. Thus, novel solutions are needed.
  • Confidentiality: Privacy is a critical aspect when handling (sensitive) patient data. Exploiting data globally and handling different data types calls for future research activities. Specifically, we apply well-known concepts from information security to provide confidentiality guarantees for involved stakeholders.
  • Representation: Given the diversity of medical research data (e.g., *-omics) and data formats (tabular data, image data, etc.), finding, linking, and joining (sparse) data is not a trivial task. Accordingly, COMSYS is looking into promising approaches and new directions in the area.
  • Distributed Analytics: Concepts like federated learning are particularly important for medical contexts, where data might be scattered across institutions. Hence, we look into these concepts to identify attack vectors, close privacy leaks, and showcase new application areas.
  • Data Quality: New findings depend on a solid foundation. In the life sciences setting, data quality and corresponding guarantees are essential (particularly in distributed settings). Consequently, we also research technical building blocks that promise to improve data quality (in distributed settings).

Please reach out via email health@comsys.rwth-aachen.de if you have any questions or would like to collaborate.

A Selection of Projects and Teaching Activities

Projects:

  • CALCIPROTECT: A novel intervention against osteoporosis and vascular calcification for CKD patients (ERS SFASIA008, 2025-2026)
  • COAT: Computational ecosystem for clinical applications of organ crosstalk (ERS PFExC005, 2022-2024)
  • myneData: Self-determined Utilization of Personal Data with Inherent Protection of Privacy and Data (BMBF, 2016-2019)
  • RUST: Latent Patient Representations using Single-Cell Transcriptomics (ERS SFFAIR002, 2025)
  • SYNCLIVER: SYNCLIVER: A Utility Assessment of Synthetic Liver Cancer Data (ERS SFFAIR003, 2025)

Teaching Activities:

Publications

Transformer-Based Integrative Patient Representations from Single-Cell RNA Data. Learning Meaningful Representations of Life Workshop (LMRL ‘25), Apr 28 - Apr 28, 2025, Singapore, Singapore. Event co-located with the 13th International Conference on Learning Representations (ICLR '25). April 2025.
Complementing Organizational Security in Data Ecosystems with Technical Guarantees. Proceedings of the 1st Conference on Building a Secure and Empowered Cyberspace (BuildSEC ‘24), Dec 19 - Dec 20, 2024, New Delhi, India. December 2024.
PASTA-4-PHT: A Pipeline for Automated Security and Technical Audits for the Personal Health Train. arXiv. December 2024.
scE(match): Privacy-Preserving Cluster Matching of Single-Cell Data. Proceedings of the 23rd IEEE International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom ‘24), Dec 17 - Dec 21, 2024, Sanya, China. December 2024.
mcBERT: Patient-Level Single-cell Transcriptomics Data Representation. bioRxiv. November 2024.
Toward Technically Enforceable Consent in Healthcare Research. Research Papers of the Platform Privacy, vol. 4, Oct 17 - Oct 18, 2024, Berlin, Germany. October 2024.
The Unresolved Need for Dependable Guarantees on Security, Sovereignty, and Trust in Data Ecosystems. Data & Knowledge Engineering, vol. 151. May 2024.
BLOOM: BLoom filter based Oblivious Outsourced Matchings. BMC Medical Genomics, vol. 10, no. Suppl 2, Nov 11 - Nov 11, 2016, Chicago, IL, USA. July 2017.