Fig. 8: Designing multi-input controllers.
From: Genetic controllers for enhancing the evolutionary longevity of synthetic gene circuits in bacteria
a–c Schematics of mechanistic multi-input controllers: (a) CLpATXλTL (protein-based inhibition of transcription, growth-based inhibition of translation, teal), (b) CLpATLλTX (protein-based inhibition of translation, growth-based inhibition of transcription, crimson), (c) CLpATLλTL (protein-based inhibition of translation, growth-based inhibition of translation, purple). See key in Fig. 2a for symbol meanings. d–i Optimised outputs for three multi-input controllers compared with CLλTL. (d) τ±10 vs initial output P0, (e) τ50 vs initial output P0, (f) %-change in τ±10 over open-loop vs initial output P0, (g) %-change in τ50 over open-loop vs initial output P0. h–j Robustness analysis for CLpATLλTX. For each of the 100 optimal controllers with nB = 300 aa, 100 further controllers were generated by varying parameters by up to ± 10% (dark grey) and ± 25% (light grey). The percentage changes in three output metrics were calculated versus the original optimal systems: (h) P0, (i) τ±10 and (j) τ50. Plots show the means (for ± 10%) and standard deviations (for both ± 10% and ± 25%) of the percentage changes for each optimal controller. Percentages marked on the plots indicate the standard deviations across the entire Pareto front when parameters were varied by ± 10%(± 25%). Only original optimal controllers where τ±10 = τ90 are considered. Simulation results are provided as a Source Data file.
