What We Do

We work with systems biology, which means that we use mathematical modelling as a tool to analyse biological data. Such modelling is centered around two key components: experimental data and mechanistic hypotheses. The experimental data can in principle be of any type, but it typically describes time-series of different players in a biological network. Such networks can be different locations of a single player (such as in the modelling of contrast agent uptake in the liver), but are typically made up of different interacting proteins or metabolites, e.g. in an intracellular signalling network. The mechanistic explanations are then hypotheses regarding how such networks are structured. These hypotheses are based in prior knowledge, but usually this prior knowledge is not unique, but contains various co-existing hypothesis. The analysis then tests which of these existing hypotheses that can and cannot describe the existing data (the first box). An equally important step is to analyse which new experiments that would give optimal information, e.g. to distinguish between different not rejected hypotheses.

More specifically, the main theoretical system we use are ordinary differential equations, but we also have experience of general machine learning techniques such as Bayesian networks. Regarding the development of new methods, we study e.g. bootstrap approaches to statistical tests regarding the first box above. Regarding the second step we are deeply involved in the analysis of uncertainty measures for predictions, for which we have developed a core-prediction methodology that is used in most of our projects.

More specifically, the main experimental systems we study concern systems related to type 2 diabetes (e.g. insulin signalling in adipocytes, and multi-level modelling of glucose and lipid homeostasis), modelling to aid diagnosis of diffuse liver disease, systems related to neurological questions (e.g. facilitation in intracellular calcium dynamics and modelling of fMRI data), desensitization phenomena in heart cells, and modelling of basic processes in yeast.