Biomedical Data Intelligence

The Biomedical Data Intelligence business area covers the entire data- and knowledge-driven value chain of translational biomedical research. Core strengths include the systematic integration and representation of biomedical knowledge and multimodal data, as well as the development of advanced, problem-oriented AI algorithms to support decision-making in early drug discovery and clinical research.

The business area comprises two departments with complementary mandates:

• The Bioinformatics Department builds on long-standing expertise in biomedical text mining, semantic technologies, and knowledge graphs to develop computable models of disease mechanisms and therapeutic hypotheses. These knowledge-based representations enable mechanism-oriented interpretation of experimental and clinical data, particularly in complex disease areas such as neurodegeneration.
• The Biomedical AI & Data Science Department focuses on machine learning, causal inference, and scientific AI for real-world and clinical study data. Key topics include individualized time-to-event risk prediction, modeling of longitudinal disease trajectories, synthetic data generation, causal treatment-effect estimation, and digital twins. The department places strong emphasis on trustworthiness, with a particular focus on explaining and contextualizing AI model predictions.

The business area is well-connected internationally and collaborates closely with leading universities, university hospitals, research institutions, and industrial partners from the pharmaceutical, biotechnology, and software sectors. As part of an institute of applied research, it operates at the interface of academic excellence and industrial innovation. In this role, it provides tailored algorithm design, applications to customer-defined use cases, innovative software solutions, systematic literature reviews, training, and strategic partnerships.

Furthermore, the business area contributes to the academic training of students in the “Life Science Informatics” program at the Bonn-Aachen International Center for Information Technology (B-IT). 

1. Unstructured information mining

35% of all data in the world are health-related data. About 80% of this data is unstructured data.

2. Knowledge-graph technologies

Our graph models represent data and knowledge in computable form. Shared semantics lay the foundation of data and knowledge interoperability.

3. Models combining data and knowledge

Models integrating data and knowledge form the basis for our approaches towards precision medicine. Now you can analyse patient-level data given the state of knowledge about disease mechanisms.

4. Actionable insights & decision support

AI and Data Science  help us to not only identify »new signals« in data, but also to perform data interpretation in a way that we can use the insights from our analysis to support decision making in R&D and clinical therapy.