Overview of the workflow in MetaReg. (a) The available knowledge about the biological system is represented by a mathematical model. (b) The model can be manually improved using simulations. (c) The model can be integrated with experimental data. For each experiment we modify the model according to the specific treatment and attach the measurements to the model variables. (d) MetaReg predicts the variable states based on the experimental treatment and the predictions are visually compared with the measurements. (e) The algorithmic engine proposes refinements to the model in order to increase its consistency with the data. The refinement process can be iterated after accepting certain model changes.