Selection of control inputs for enhancing the long-term performance of synthetic gene circuits

Engineered biological systems fail when mutations arise that inhibit their intended function. Such mutations introduce uncertainty into the system parameters, making negative feedback control an attractive strategy to improve the evolutionary longevity of synthetic gene circuits. Here we propose three classes of controller that improve evolutionary performance in repeated batch culture. Each controller takes a different biological input: (i) the gene product in the cell, (ii) the host growth rate and (iii) the total population output. Using a multi-scale model of host-circuit interactions, we demonstrate that these different modes of action differentially influence the growth dynamics between mutant strains, driving significant differences in the long-term performance of gene circuits. We show that population-based feedback is least effective and that, whilst direct feedback inhibition is effective in the short-term, growth-based feedback can enhance long-term performance to a greater degree. We propose a novel control strategy which combines two of the strategies and improves both short-and long-term performance.