Toward predictive modeling of large and complex biological signaling networks

• Define two types of model predictions: predictions in data-rich and data-poor realms.
• Discuss advantages and disadvantages of current modeling strategies.
• Explain our approach as a case study for a new strategy.

The main goal of this article is to identify critical issues that arise when building in silico models that can predict the behavior of complex biological signaling networks. We discuss practical approaches to overcome, circumvent, or moderate these issues. Here we focus on modeling spatially homogenous systems, such as those consisting of cells of the same type receiving the same input. Proper modeling of spatially heterogeneous systems requires additional model features and added modeling scales, beyond those required for spatially homogenous systems. These different modeling scales are defined here as different layers in the hierarchical relationships in a network. For example, protein molecules and cells that contain the protein molecules belong to different scales.