The Swedish Foundation for Strategic Research, SSF, funds a five year project entitled Hierarchical Mixed Effects Modeling of Dynamical Systems with 20 million SEK.

The project, which is a collaboration between Mathematical Sciences (MS), the Fraunhofer-Chalmers Centre (FCC), and the Sahlgrenska Academy (SA), will address methods, models, and computational tools for biological and biomedical problems involving large number of individual subjects (such as cells in a yeast colony, tumor cells or patients in a clinical study) that exhibits complex time-dependent behaviors to external stimuli. The project team consists of Bernt Wennberg (main grant holder, MS), Mats Jirstrand (FCC), Martin Adiels (MS, SA) and Philip Gerlee (H. Lee Moffitt Cancer Center, USA).

Complex diseases such as cancer and metabolic diseases are the results of complicated and multi-factorial changes often with genetic and environmental components. It is therefore acknowledged that there is a great intra individual variability within disease phenotypes. Variability also exists on the cellular level which may both provide robustness or, when perturbed, introduce pathological changes. Studying only average behavior of systems may lead to false conclusions regarding the system of interest. Taken together, the variability and heterogeneity of the underlying systems of cells, individuals and population makes it hard to extract relevant information and to handle the growing amount of experimental data.

Techniques to study and handle biological variability, such as mixed effects modeling, have primarily been applied in pharmcodynamics/pharmacokinetics research, in which models are comparably simple. The lack of penetration of these theories into biology and medicine may in part be explained by the complexity of the computer software, as the initial threshold is too steep for ‘non modelers’.

The project focus on three research applications which will either benefit from a mixed effect modeling approach, or that may exclusively be studied using mixed effects modeling. The applications are: (A) Proliferation and migration dynamics in brain tumor cells. (B) Human lipid metabolism in metabolic diseases. (C) Regulation of glucose metabolism in yeast.

The main methodological aims of the project are: (1) Framework for mixed effects modeling for both cross sectional and longitudinal data. (2) Extension of homogenization techniques for multi-cellular systems. (3) Formalization of handling of uncertainty using stochastic differential equations. (4) Software tools for model development and analysis, and parameter estimation.