Evolving designs for gene oscillators

View all posters

Cyrus Sahba, Chris Barnes

UCL Research Department of Cell and Developmental Biology, United Kingdom

Synthetic biology has vast potential in many fields of application, but has so far failed to live up to its promise. The main reason for this is that current design and modelling strategies are greatly lacking. The design paradigms of electronic engineering, and even of network motifs, do not fully account for the complexity and inherent stochasticity of biological systems, and current modelling techniques are inefficient. As we attempt to design more and more complex systems, these difficulties are compounded exponentially, leading to synthetic designs that do not possess the robustness they require to perform and a great deal of trial and error. A more holistic approach to design is the use of Bayesian model selection, a form of in silico evolution, where a prior set of high-dimensional design space is tested and selected for using a desired outcome. ABC-SysBio is used to infer ‘designs’ for a robust genetic oscillator system from all possible 3 node systems. With this technique the design space can be limited by practical considerations, such as the immutable characteristics of available genes, and can also infer kinetic parameters of the system on top of node connectivity. The leading candidates are compared computationally and experimentally to a previously characterised synthetic oscillator that was consciously designed.