Reduction of a detailed biological signaling model

Biological signaling is complex. Even if only few components are involved models for biological signaling, in general, comprise a large number of variables and parameters to achieve predictive power. This is due to the many states that can be attained even with few components due to the formation of (allosteric) complexes. This phenomenon is generally referred to as combinatorical complexity. Although the detailed parameterized and validated models can be analysed to reveal regulatory principles these models are, in general, too complex to achieve an intuitive understanding. Methods are urgently needed to achieve meaningful model reduction. Ideally, biologists would like models that retain the simplicity of the typical signaling cartoon, yet provide novel insight. We suggest a 2-step process to achieve this. In a first step a large detailed model is developed and tested based on experimental data. In a second step the detailed information gained from the validated model is used to develop a realistic phenomenological model. The procedure is illustrated by example of σF activation during sporulation in Bacillus subtilis. The reduced model indeed successfully reproduces key regulatory aspects of the detailed model and shows how the the exceptional sensitivity of the regulatory network results from the particular allosteric interactions between SpoIIAB (AB), SpoIIAA (AA), and σF and from the sequestration of AB in inactive AB-ADP-AA complexes