Control of complex networks requires both structure and dynamics.

Area of Science:

• Systems biology
• Network science
• Control theory

Background:

• Complex systems exhibit organizational signatures through network structure.
• Controlling complex systems, like reverting diseased cells, requires understanding critical variables.
• Structure-only methods (structural controllability, minimum dominating sets) aim to predict controllability from interaction graphs.

Purpose of the Study:

• To evaluate the efficacy of structure-only methods in predicting system controllability.
• To investigate the role of system dynamics and logic functions in determining controllability.
• To compare structure-only predictions with actual control strategies in biological models.

Main Methods:

• Analysis of Boolean network ensembles and network motifs.
• Study of three biochemical regulation models: Drosophila melanogaster segment polarity network, yeast cell cycle, and Arabidopsis thaliana floral development.
• Comparison of structure-only controllability predictions with dynamics-informed control analyses.

Main Results:

• Structure-only methods fail to accurately characterize controllability when system dynamics are considered.
• These methods both underestimate and overestimate the number and identity of critical control variables.
• The canalizing nature of automata transition functions significantly impacts the predictive power of network structure.

Conclusions:

• Network structure alone is insufficient for predicting the controllability of complex systems.
• System dynamics and the logic of regulatory functions are essential for accurate control strategy identification.
• The findings underscore the need to integrate dynamic information for effective control of biological networks.