A Cross-scale Model of Tumor Growth: Do We Need to Model Molecular Interactions in Separate Artificial Compartments within a Cell?

We modified an existing cross-scale model of avascular tumor growth that couples a gene-protein interaction network with an agent based model.

In the original model each biological cell is artificially subdivided into four geographic compartments to account for a spatial polarity of the molecules. Since these artificial compartments result in eight ordinary differential equations (ODEs) — instead of one — that need to be solved for each molecular species we tried to reduce the related computational burden.

We renounced to include these artificial compartments and assume a homogeneous distribution of molecular concentrations within each cell. By adapting the conditions for nutrient uptake and the neighborhood options for cell migration and division we achieve results that are comparable to the original model. However, for this modified model the number of ODEs that need to be evaluated for each molecular species within each cell and thus the computing time could be significantly reduced.