PLoS Comput. Biol., published online 19 July 2012; doi: 10.1371/journal.pcbi.1002579

Credit: HAFNER, MILLER & WEISS

Synthetic biologists often seek to create gene networks that execute defined tasks, but the application of these networks in undefined systems offers important orthogonal tests of our ability to engineer biology. Miller et al. now explore the interplay of design and disorder in a conceptual framework suitable for creating a stable population of stem cells and differentiated β-cells. The authors first designed and modeled a system including four known modules that determine whether stem cells should undergo renewal or differentiation into β-cells but observed that delays in the differentiation process could cause unwanted oscillations in population homeostasis. To counter this, a second system included a 'commitment' module, but this model was also prone to oscillations because of the potential for multiple simultaneous commitments to drain the pool of stem cells. To avoid this synchronization, two final systems explored mechanisms to generate variability in decision making, using either a newly designed and tested toggle switch or an oscillator. The authors further developed several computational methods such as an 'intermodular coupling analysis' to define the best integration of the varying time scales of the different modules and a 'phenotypic sensitivity analysis' to identify whether input modules would achieve a desired phenotype even though specific parameters of the system could not be defined. These results offer new guidelines for systems where heterogeneity is a feature, not a bug.