Systems biology

Systems Biology is an interdisciplinary approach with the goal to increase the understanding of biological systems.

Systems Biology is an interdisciplinary approach with the goal to increase the understanding of biological systems. It is a consequential application of tools and methods from systems and control engineering to biological systems. At the heart of this approach is the development of mathematical models to simulate and analyze cellular systems.

The biological systems under investigation at the IST range from intracellular processes like signal transduction in mammalian cells to cell-cell interactions in specific tissues like a tumor. While biological expertise is introduced by a number of long-term collaborations with experimental biologists, the mathematical modeling of the considered processes is done in our group. The wide range of applications is reflected by the diverse modeling frameworks that are applied, including stochastic models, ordinary differential equations, models where only the network structure is considered, and models for heterogeneous cell populations.

Mathematical analysis methods aim at a mechanistic understanding of the underlying processes and, based on this, at deriving model based predictions and hypotheses for new scenarios. The IST systems biology group focuses on applied projects and on the development of model analysis methods that are inspired by statistical learning methods and control engineering. In order to support the modeling process, parameter identification from experimental data is a very relevant issue. In our group, various new approaches to solve this problem are being studied. Models for biological systems are often subject to a large degree of uncertainty. To deal with this problem, we develop methods for uncertainty and robustness quantification which allow deriving reliable statements from uncertain models. A specific interest of the IST systems biology group also lies in the analysis of complex dynamical behavior of biological systems, such as the switch-like behavior encountered in bi-stable systems.


Please find below all our recent research fields at the Institute for Systems Theory and Automatic Control referring to Systems Biology.


Systems Biology

Regulation mechanisms of DLC1 and their role in tumour cell migration

Understanding regulation mechanisms in intracellular signaling pathways and malfunctions that frequently occur in cancer cells is important to improve tumour therapeutics and to find appropriate drug targets. Our project focuses on the regulation mechanisms of the protein 'Deleted in Liver Cancer 1' (DLC1), which acts as a tumour suppressor protein by serving as a GTPase activating protein (GAP) for members of the Rho family of GTPases. These GTPases are critically involved in various cellular processes such as the cytoskeletal organization and cell migration and thus control the invasiveness of cancer cells.

The aim of this project is to get a profound understanding of regulation mechanisms of DLC1 and downstream mechanisms through which DLC1 acts as a tumour suppressor. This is realized via a systems biological approach combining mathematical models and experimental data. Experiments are performed by the group of Prof. Dr. Monilola Olayioye, Institute of Cell Biology and Immunology, at the University of Stuttgart.

Moreover this project also focuses on further methodological development of parameter estimation techniques and subsequent uncertainty quantification for those biological models using statistical approaches like Maximum Likelihood estimation and Bayesian methods.


The effects of error model choice on dynamic modeling in systems biology

One of the main goals of research in systems biology is the development of quantitative models to describe intracellular reaction pathways and their dynamical behaviors. Nowadays the most common approach is to model one “single typical cell”, mainly by using dynamical models in the form of parametrized ODE systems. This research project is focusing in particular in the statistical description of biological data, needed for model calibration. 

In order to use Bayesian methods for parameter estimation, such as Maximum Likelihood estimation or MCMC sampling from posterior distribution, we have to define a specific statistical error model for the measured noisy data (e.g. normal, log-normal, etc).
This mathematical hypothesis influences the results of model calibration, and, most importantly, the predictions of new experimental conditions. In particular, I am interested in analyzing new different error models for normalized experimental data, which is a very common necessary processing of biological data, for example in the case of Western Blot data. By taking into account the variability in the reference condition used for normalization, we end up with ratio distributions as possible error models for normalized data. 

Different normalizations of measurementRatio distributions often have the property to be heavy-tailed distributions, and this fact has significant repercussions on model fitting.
Within this project, I aim to analyze how different error models influence the results of state estimators [1], model calibration and model predictions. In our group, we have also investigated the effect of data normalization on model sensitivity analysis [2]. First results hint to the fact that the outcome of sensitivity analysis changes when we consider normalized model output, and this fact complicates the interpretation of the results. 

Another important focus in systems biology is the development of adequate mathematical methods for the analysis of the structural properties of biological intracellular networks. One interest of my research in this respect is the investigation of the role of (positive and/or negative) feedback interactions on the transient behavior of signaling cascades, on the number and stability properties of fixed points of the system, on the robustness of the network, and other properties.


  • Publications:
  • C. Thomaseth, N. Radde.
    Normalization of Western blot data affects the statistics of estimators.
    Accepted for publication in Proceedings of FOSBE Conference on Foundations of Systems Biology in Engineering 2016. [1]
  • J. Kirch, C. Thomaseth, A. Jensch, N. Radde.
    The effect of model rescaling and normalization on sensitivity analysis on an example of a MAPK pathway model. EPJ Nonlinear Biomedical Physics. 4:3, 2016. [2]
  • Cooperations:
  • Systems Biology Ireland, University College Dublin (Prof. Boris Kholodenko, Dirk Fey)

Synthetic Biology

Scheme of a cell with genetic controllerIn contrast to the field of systems biology – where a top-down approach drives the understanding of biological systems – synthetic biology tries to follow a bottom-up approach to understand and actively control biological systems.

By performing system theoretic analysis on dynamical models – built up by bottom-up or top-down approaches – one can re-engineer existing or design novel biological systems with pre-specified functionalities. In the long run, these engineered systems enable us to tackle unsolved challenges in fields of medicine, energy production or environmental monitoring.

Our main interest at the IST is the application of automatic control principles, especially focusing on the optimal and robust design of genetic regulatory networks (GRNs). Such GRNs consist of several interconnected genes where they interact with each other on many levels, e.g. on the level of transcription or translation via transcription factors or silencing RNAs. These different types of interactions yield different genetic control mechanisms – producing wide ranges of nonlinear dynamical behavior and thus offering challenging tasks for control engineers.

The current research at the IST aims to characterize and analyze such genetic control mechanisms with respect to their system theoretic properties, e.g. transient dynamics, stability and the capacity of operational environment. Living cells introduce limits to the dynamics, therefore the characterization of resource consumption is an essential aspect of our research.

The systemic understanding of genetic control mechanisms and resource consumption may bring us closer to predictive mathematical models of GRNs. Such understanding will produce novel synthetic components and eventually lead to rational design of larger modules to tackle the unsolved challenges.


  • Publications:
  • Halter, W., Montenbruck, J. M., Tuza, Z. A., Allgöwer, F.
    A resource dependent protein synthesis model for evaluating synthetic circuits.
    Journal of Theoretical Biology, 420, 267-278, 2017.
  • Halter, W., Tuza, Z. A., Allgöwer, F.
    Signal differentiation with genetic networks.
    IFAC-PapersOnLine, 50(1), 10938-10943, 2017.
  • Halter, W., Montenbruck, J. M., Allgöwer, F.
    Systems with integral resource consumption
    Proc.\ 56th IEEE Conf.\ Decision and Control (CDC), 2017.
  • Halter, W., Montenbruck, J.M., Allgöwer, F.
    Geometric stability considerations of the ribosome flow model with pool.
    Proc.\ 22nd Int.\ Symp.\ Mathematical Theory of Networks and Systems (MTNS), 2016.

Modeling and analysis of apoptosis in cancer cell populations

Apoptosis, a mechanism of programmed cell death, is crucial for developmental processes in multicellular organisms. Unbalances of apoptosis are related to diseases such as immunodeficiency and cancer. One of the inductors of apoptosis is the TNF related apoptosis inducing ligand (TRAIL). TRAIL binds to receptors on the membrane of a cell and activates a signaling cascade that leads to activation of caspase-3 and subsequently to apoptotic cell death. Because of its selectivity for transformed cells and its low toxicity for normal cells, the ligand is currently under investigation for a treatment of cancer. Despite great proceedings referring to the mathematical and experimental analysis of the signaling pathway in the past decade, the complex regulation of apoptosis still raises questions.

scheme of TRAIL pathwayThis project is embedded in the BMBF supported project 'Predictive Cancer Therapy' and focuses on the mathematical modeling and analysis of the cellular response to TRAIL. The model structure is built from and continuously adapted to experimentally determined interactions and parameters are fitted to measurements. Within the project, a close collaboration with experimental partners from the Institute for Cell Biology and Immunology at the University of Stuttgart is maintained. The work flow is promoted by an active exchange of ideas from modeling and experimental side.


A important issue to consider with respect to a cancer treatment is the heterogeneity of cells. Although cells in a population are genetically identical, differences in expression levels of proteins emerge and affect the apoptotic response behavior. In order to understand TRAIL mediated death of cells in a population, it is necessary to consider factors of heterogeneity. This requires methods in stochastic modeling and the development and application of respective analysis tools. Long-term goal of the project is the support and improvement of anti-cancer strategies.


  • Publications:
  • Imig, D., Pollak, N., Strecker, T., Scheurich, P., Allgöwer, F., Waldherr, S.
    An individual-based simulation framework for dynamic, heterogeneous cell populations during extrinsic stimulations. J. Coupled Syst. Multiscale Dyn., Vol. 3(2), pp. 143–155, 2015.

Sustainable and Effective Biosyntheses

In the face of decreasing organic and fossil resources, alternative production routes for important chemical and pharmaceutical raw materials have to be developed in near future. For solving this task, nature is not only a role model but also contributes to the solution by seeking other raw materials and recycling possibilities.

By the close cooperation of the working groups present in the BW² network, we are able to build an internationally visible and excellent focus. Through this network, the disciplines of biotechnology, chemistry and engineering are linked; resulting in integrated research methods which could be used accordingly. The main goal of this network is to develop novel methods for biosynthesis for the efficient production of products. These products are characterized by lower costs compared to conventional methods in terms of their production costs, use of raw materials, generation of waste and emissions and their risk potential. For this purpose, three product-oriented priorities are defined within three research clusters:

  • Manufacture of basic chemical products on the basis of synthesis gas
  • Production of terpenes and terpenoids
  • Preparation of functional peptides

To accomplish these objectives, efficient and resource-conserving biosynthesis will be developed, the spectrum of bio catalytic reactions will be expanded, chemical and biochemical reactions will be combined and host strains are selected and adjusted and subsequently checked for their genetic stability and heterogeneity.

The work at the IST in detail focuses on mathematical modelling of biological and chemical cascade reactions, their analysis and optimization and ultimately in the development of new synthesis routes applying methods from automatic control theory.


  • Publications:
  • Halter, W., Kress, N., Otte, K., Reich, S., Hauer, B., Allgöwer, F.
    Yield-Analysis of Different Coupling Schemes for Interconnected Bio-Reactors.
    Proceedings of SIAM Conference on Control and its Applications 2015.
  • Cooperations:
  • Institute of Biochemical Engineering (IBVT), University of Stuttgart (Konrad Otte, Sabrina Reich, Niko Kreß und Bernhard Hauer)

Cell decision processes in heterogeneous cell populations

Multiple aberrations in various cellular signaling pathways accumulate in the development of cancer. Many dysregulations occur in cellular pathways, that are activated in response to stress related stimuli such as DNA damage, nutrient starvation or death receptor ligands. Activity in these cellular signaling pathways causes cell fate decisions towards cell cycle arrest, senescence, autophagy or apoptosis. The signaling pathways furthermore affect each other and this interaction represents cell intrinsic control mechanisms.Together with our collaborators  we aim at a deeper understanding of these cellular control mechanisms.

Fluorescence microscopy image of 3D multicellular tumor spheroidWe follow this aim by an integrative approach, where we combine methods from data analysis, biological system modeling and system theory.
The focus of our data analysis, in particular single cell and population data from fluorescence microscopy or FACS experiments, is to identify subpopulations (e.g. drug sensitive or resistant) within the biological sample which behave differently in response to some treatment (e.g. drug sensitive or resistant). The data from different subpopulations is then analyzed with the help of mathematical models that describe the process. The model classes that we use for this purpose ranges from classical ordinary differential equation (ODE) to partial differential equations (PDE) and stochastic differential equations (SDE) models.
The comparison of the model parameters, structure and predictions, that we get from the various subpopulations, allows us to identify crucial processes that cause the emergence of subpopulations. The knowledge that we ultimately obtain from this work flow helps to identify new drug targets, and to optimize treatment strategies.


  • Publications:
  • Kuritz K., Stöhr, D., Pollak, N., Allgöwer, F.
    On the relationship between cell cycle analysis with ergodic principles and age structured cell population models.
  • Cooperations:
  • Institute of Cell Biology and Immunology, University of Stuttgart (Markus Morrison, Nadine Pollak, Daniela Stöhr)

  • El-Samad Lab, UCSF (Hana El-Samad, João Fonseca, Alain Bonny)

Structural design with biological methods: optimality, multi-functionality and robustness

Natural design principle can be described as Tunneling Through the Cost Barrier - Amory Lovins, which means focusing on radical optimization of the whole system rather than incremental improvements of individual components and at the same time maintaining robustness. Robustness against perturbations and imperfections is a crucial property that systems optimized towards one single goal typically do not possess. In this project, we are aiming at transfer of biological design principles to technical systems in an abstract way based on graph theory or matrix representations, rather than the phenomenological or geometrical appearance of the biological and technical systems. This should lead the transfer of biological principles beyond the original functionality. To be more precise, a biological example for a load bearing structure does not have to be a load-bearing structure itself but can be any biological entity exhibiting a similar formal structure. In particular, recently we have shown in [1] that signaling cascade and load-bearing truss structure share a common design principle based on complexity and in both the cases robustness can be measured under sensitivity towards perturbations or imperfection.

In a more in-depth study [2], we have investigated robustness of ubiquitous signaling network by comparing three different models of phosphorylation-dephosphorylation cascade in terms of sensitivity (global and local) towards input parameter and filtering properties. We have found that signaling cascade with higher complexity acts as an efficient low pass filter and robust towards variation in input parameter.


  • Funding: The project is funded by the Collaborative Research Center SFB-TRR 141 and is a joint work with Prof. Manfred Bischoff of Institute of Structural Mechanics (IBB), University of Stuttgart.
  • Publications:
  • Paul, D., Dehkordi, L. K. F., von Scheven, M., Bischoff, M., and Radde, N. (2016). Structural design with biological methods: optimality, multi-functionality and robustness. In Biomimetic Research for Architecture and Building Construction (pp. 341-360). Springer International Publishing. [1]
  • Paul, D., and Radde, N. (2016). Robustness and filtering properties of ubiquitous signaling network motifs. IFAC-PapersOnLine, 49(26), 120-127. [2]

Modeling of the MAPK signaling pathway and investigation of feedback interconnections in cancer cells

MAPK pathway

Cancer cells present many aberrations, in particular abnormal proliferation, reason why it is very important to develop new therapies that are able to arrest the uncontrolled growth of tumor cell populations in human tissues. A promising cancer treatment alternative to the commonly used chemo- or radiotherapy is the development of protein targeted therapies. The idea is to target some specific key proteins whose biosynthesis needs to be controlled in cancer cells, while leaving normal cells unperturbed.

Focus of this project is the mathematical modeling and analysis of the mitogen-activated protein kinase (MAPK) signaling pathway. Growth factors bind to receptors on the cell membrane and activate a downstream cascade of reactions that culminate in the activation of MAPK, whose concentration controls the growth and proliferation of the cell. A great amount of data-driven models and results about this pathway appears in the literature, but many aspects regarding the complexity of interactions concealed behind the apparent simplicity of the cascade are still unclear.

A main characteristic of the considered pathway is in fact the cross talk with many other pathways such that many interconnections create a robust response of the cell to changes in external stimuli. Being able to model this complicated and interconnected signaling network, thanks to the methods of systems theory and control engineering, would provide an important insight of the possibility to control the life and death of cancer cells.

To describe the dynamics and interactions of the proteins of interest, we develop parameterized ordinary differential equation (ODE) models. Another important task is the estimation of the parameters of the model, which are often unknown or cannot be precisely measured in laboratory, by fitting the simulated results of the model to experimental datasets obtained in real cancer cells. In this regard, a close collaboration with experimental partners from the Institute for Cell Biology and Immunology at the University of Stuttgart is maintained, who supply us in vitro cell measurements. Different approaches for parameter estimation are implemented, like the Maximum Likelihood Estimation (MLE) or Bayesian approaches of sampling from the posterior distribution.

Another interesting goal is the development of general methods to investigate, with the help of theoretical approaches of systems and graph theories, the presence of positive and negative feedback interconnections contained in the considered signaling network.


  • Project: BMBF funded e:Bio project PREDICT
  • Publications:
  • Jensch A, Thomaseth C, Radde N. Sampling-based Bayesian approaches reveal the importance of quasi-bistable behavior in cellular decision processes on the example of the MAPK signaling pathway, BMC Syst Biol (2017), 11:11, doi:10.1186/s12918-017-0392-6.


To the top of the page