Figure 15. Phylogenetic tree of replicators.

The tree was built from the genotypes (sequences) of 2000 replicators sampled from a simulation with the model explained in the main text (1 − q = 0.004). The leaves of the tree are color-coded based on the sequence composition in the dangling ends of the respective replicators. Each replicator, however, has up to four dangling ends (two ends in each of two complementary strands). Thus, one dangling end was chosen for each replicator as follows. If replicators are a replicase (i.e., if they have the catalytic structure), the 5’ dangling ends that recognize templates are chosen (i.e., the 5’ dangling end of the strands that actually fold into the catalytic structure). (Most replicases have the catalytic structure only in one strand.) The fraction of Cs and As in these dangling ends determine the colors of the corresponding leaves (the color scale is shown in the inset). If replicators are not a replicase, the 3’ dangling ends that have the most extreme sequence composition is chosen for each replicator. In this case, the colors are set to a function of the fraction of Gs and Us in these dangling ends (the color scale is shown in the inset). According to the above coloring scheme, each sequence class tends to appear as follows: C-catalysts tend to be cyan; A-catalysts, magenta; G-parasites, red; U-parasites, green. However, there are exceptions: The red leaves appearing in the clades of C-catalysts are not G-parasites, but are mutants (of C-catalysts) that have lost the catalytic structure. Thus, they belong to the C-catalyst quasi-species. Likewise, the green leaves in the clades of A-catalysts belong to the A-catalyst quasi-species. The tree was constructed with a maximum likelihood method implemented in PhyML [106]. However, owing to great divergence among sequences, the constructed tree does not necessarily depict evolutionary relationships between replicators.