Multi-scale cell-based modelling of gene regulation in development

To provide insight into the dynamical behavior of gene regulation in development, detailed dynamical models are required, which can be tested against available experimental observations. The dynamics of these regulatory networks varies across space and time and is influenced by biomechanical events (e.g. formation of cell layers, cell migration, cell death, and cell division). Here we focus on spatio-temporal models of gene regulation linked to the dynamics of cells, and aimed to capture key quantitative aspects of pattern formation during embryo development. A main issue of such complex models of pattern formation is the large number of parameters, which are often difficult or impossible to measure. Instead, they have to be inferred by fitting models to data using global, non-linear optimization. This type of reverse engineering approach poses significant challenges. Usually, the model accuracy of predicting observed expression patterns is measured by a cost function based on the sum of squared differences between model and data. We propose to develop and apply new methods for multi-objective optimization in models of pattern-forming gene regulatory networks, in which distinct optimization criteria, will be integrated such as the accuracy of a fit and the robustness of a solution towards parameter perturbation. In this project we want to couple the model of gene regulation with a physically based model predicting the collective behaviour of cells at both the microscopic and macroscopic level. We will test these new methods on available data sets of gene expression in embryogenesis of the cnidarian Nematostella vectensis. This work will be done in collaboration with Prof. M. Martindale (University of Hawaii)