Analysis and Design of Molecular Machines

Biologically inspired computation has been recently used with mathematical models towards the design of new synthetic organisms. In this work, we use Pareto optimality to optimize these organisms in a multi-objective fashion. We infer the best knockout strategies to perform specific tasks in bacteria, which involve concurrent maximization/minimization of multiple functions (codomain) and optimization of several decision variables (domain). Furthermore, we propose and exploit a mapping between the metabolism and a register machine. We show that optimized bacteria have computational capability and act as molecular Turing machines programmed using a Pareto optimal solution. Finally, we investigate communication between bacteria as a means to evaluate their computational capability. We report that the density and gradient of the Pareto curve are useful tools to compare models and understand their structure, while modelling organisms as computers proves useful to carry out computation using biological machines with specific input–output conditions, as well as to estimate the bacterial computational effort for specific tasks.