Treloar, Neythen J.; (2022) Towards the implementation of distributed systems in synthetic biology. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text NT_thesis_final.pdf - Other Download (20MB) | Preview

Abstract

The design and construction of engineered biological systems has made great strides over the last few decades and a growing part of this is the application of mathematical and computational techniques to problems in synthetic biology. The use of distributed systems, in which an overall function is divided across multiple populations of cells, has the potential to increase the complexity of the systems we can build and overcome metabolic limitations. However, constructing biological distributed systems comes with its own set of challenges. In this thesis I present new tools for the design and control of distributed systems in synthetic biology. The first part of this thesis focuses on biological computers. I develop novel design algorithms for distributed digital and analogue computers composed of spatial patterns of communicating bacterial colonies. I prove mathematically that we can program arbitrary digital functions and develop an algorithm for the automated design of optimal spatial circuits. Furthermore, I show that bacterial neural networks can be built using our system and develop efficient design tools to do so. I verify these results using computational simulations. This work shows that we can build distributed biological computers using communicating bacterial colonies and different design tools can be used to program digital and analogue functions. The second part of this thesis utilises a technique from artificial intelligence, reinforcement learning, in first the control and then the understanding of biological systems. First, I show the potential utility of reinforcement learning to control and optimise interacting communities of microbes that produce a biomolecule. Second, I apply reinforcement learning to the design of optimal characterisation experiments within synthetic biology. This work shows that methods utilising reinforcement learning show promise for complex distributed bioprocessing in industry and the design of optimal experiments throughout biology.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONCSVXML

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item