Figure 1. Principles of agent-based modelling.

(A) An agent-based simulation consists of a virtual environment where large numbers of autonomous agents can interact. A model of a bacterial colony is shown with agents representing cells. Each cell contains a synthetic genetic circuit that controls its behaviour. In this case, the genetic circuit takes two chemicals as inputs (Q1 and aTc) and produces a single chemical output (Q2) if both inputs are absent (a NOR logic operation). A range of common cellular inputs and outputs are shown. To ensure that simulations faithfully reproduce the biological system, key physical processes encountered or utilized by the agents must be implemented within the virtual environment. Those relevant to bacteria are shown. (B) Interactions between agents implementing specific rules and the shared environment can lead to the emergence of collective behaviours. These include dynamic co-ordination (e.g. synchronization of gene expression; see Figure 2A) and population-level encodings of continuous inputs (e.g. cells are either in an ‘ON’ or ‘OFF’ state and the fraction of the population in an ‘ON’ state corresponds to the continuous concentration of the input, similar to the bimodality of the lactose utilization network in E. coli [10]).