Thousands of genes and proteins control the fate of cells and organisms. Understanding their cooperation is crucial for developing new therapies for complex diseases. This is like putting together a complex puzzle without knowing its picture, and is the aim of the field of systems biomedicine.

 To assemble this puzzle, advanced data analysis and modeling technologies are required to extract the relevant information out of the deluge of data generated today. Developing tools to put together the pieces is the research focus of the Joint Research Center for Computational Biomedicine. We strive to tackle the challenge integrating both data driven modeling and mechanistic modeling methods.

Data-driven modeling

Mathematically, putting together the puzzle is a so-called inverse problem. It requires novel combinations of statistics and pattern recognition, inverse problems and multi-scale modeling methods. This research on mathematical methods has multiple applications. Examples are novel assessment methods for induced pluripotent stem cells using their genome-wide expression patterns, unravelling of the mechanisms of drug action in cancer cell lines, tracking of the dynamics of cellular evolution processes or predicting disease propagation by appropriate projections of data onto physiologically motivated subspaces of the data patterns.

Future developments will focus on the development of BigData methods for modeling complex biomedical processes: induced reorganisation of cellular mechanisms controlling the quantitative, long-term response of a biological system on stress and environmental changes, , the development of multi-scale disease models to the level of full size organisms or the identification of critical states of disease progression in chronic and acute diseases complement the research program on this area.

Mechanistic modeling

Mechanistic modelling aims to integrate experimental data with our knowledge of the biochemistry of the cell, typically casted in the form of networks. In particular, we construct mechanistic and predictive models of signaling networks by integrating pathway information with proteomic and transcriptomic data sets. Application of these methods in our institute aims at obtaining a functional understanding of the alteration of signalling, regulatory and proteome quality control networks in disease, and to apply this knowledge toward the development of novel therapeutics.


Big biomedical and clinical data

We are interested in the inference of models from big data: we aim to extract the information contained in  heterogeneous biological data from multiple sources to gain insight into the functioning of biological systems. To reach this goal, we have a focus  on hybrid modeling methods to merge biomedical mechanisms with tools from statistics, machine learning, and complex systems theory. Examples include:

Network modeling

Across many spatial and temporal scales, networks provide a powerful paradigm to understand the behavior of biological systems. Networks enable the integration of complex building blocks and their dynamic interactions into a concise mathematical framework. Thus building accurate networks and understanding their functioning are a powerful tool to address many biomedical questions. Examples include: