Fig. 3.—

Evolved quasispecies: Clustering, sequence logo, consensus sequence, and ancestor tree. Top: Clustering, sequence logo, and most abundant sequence with its secondary structure (in dots-and-brackets notation) of the quasispecies at $\mu =0.015$. In total, 1,000 sequences are randomly sampled from the population at one time point, clustered (500 are displayed for clarity) and the sequence logo is generated (sampling at different time points, or larger samples do not produce qualitatively different results. Colors on the leaves correspond to the functional class in which sequences fold (see Materials and Methods). Cyan: most abundant sequence, black: units of replication, green: helpers, red: stallers, grey: junk, blue: hybrids. The tree is visualized with iTOL (Letunic and Bork 2011). Bottom: Ancestors tree over the first $450\times {10}^3$ time steps of a simulation run (the simulation is initialized with a homogenous population of master sequences, it consisted of $\approx 2.5\times {10}^6$ time steps, after which it was interrupted). The rest of the simulation as well as different runs shows qualitatively the same pattern. The tree is built so that the nodes at a given time step are every individual’s ancestor $50\times {10}^3$ time steps later, edges connect lines of descent; yellow nodes are those in the line of descent that persists until the end of the simulation, in red the others. Numbers mark the Hamming Distance of the ancestor from the master sequence. The tree is visualized in Cytoscape (Smoot et al. 2011).