Skip to main content
. 2014 Dec 4;10(12):e1003936. doi: 10.1371/journal.pcbi.1003936

Figure 3. The evolution of division of labor when both replication affinity and metabolic activity of replicators are allowed to evolve separately.

Figure 3

(A) A representative example of simulations resulting in asymmetric strand template reaction averaged over the population of Inline graphic vesicles (Inline graphic: red; Inline graphic: orange; Inline graphic: dark blue; Inline graphic: light blue). Simulations begin from an initially symmetric state, i.e. all strand types are represented in equal numbers (Inline graphic) and equal template replication rates (Inline graphic). We assume low initial metabolic activity of the minus strands (Inline graphic) and a trade-off between the maximum values of the replication affinity and the catalytic activity of the replicators (see red line in C), i.e. no replicator can evolve traits above this boundary, but any rate combination below the curve is accessible (i.e. Inline graphic, see Models Eq. 1b). (B) As metabolic activity gradually evolves towards high values (brown and dark blue lines, Inline graphic) the minus strands trade in replication affinity (red and blue lines, Inline graphic) in order to reach the optimum. When the replication affinity of the plus strand can also evolve, evolution further optimizes the protocell composition in favor of strand asymmetry by evolving the highest possible affinity for the plus strand (grey and dark grey lines, Inline graphic). Here Inline graphic is allowed to evolve without any trade-off (Inline graphic, and the initial condition is Inline graphic). (C) Trajectories from different initial conditions (green: Inline graphic and Inline graphic; purple: Inline graphic and Inline graphic; and blue: Inline graphic and Inline graphic) converge to the same equilibrium. Solid and dotted lines depict molecule types 1 and 2, respectively. Filled circles represent the initial data points, while light shaded circles and rectangles represent the evolutionary endpoints for traits of molecules 1 and 2, respectively. For the above results we employed a continuous-trait model, in which traits were allowed to change continuously between 0 and 1, and mutant traits were drawn from a normal distribution with the resident trait as a mean and with variance Inline graphic. Other parameters: Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic, Inline graphic and Inline graphic.