Automated model-driven generation of software components for the simulation of epithelial tissues

Quantitative in silico modeling is a powerful means to enhance our understanding of complex biological systems. Accordingly, intuitive and flexible computational tools are needed to support the development of such models. We previously developed the platform EPISIM for graphical modeling and simulation of cellular behavior in epithelia. In this work we demonstrate how computationally efficient software components for epithelial tissue simulations can be automatically generated. We introduce a model-driven workflow to generate extendable and exchangeable software components for both the modeling and the simulation of epithelial tissues. We distinguish two levels of abstraction in our workflow and thus two kinds of models: (i) the meta-model of our modeling language and (ii) particular systems biological cell behavioral models. The model-driven component generation allows optimization of the underlying code and the automated integration in our EPISIM platform. We evaluated the computational performance and the correctness of the generated software components. In this work we focus on the evaluation of the computational performance. It could be shown that the execution time increases nearly linearly with the size of the generated component's underlying model.