Fig. 3

a Schematic depiction of key genomic elements from the ancestral organism (top) and representative organisms from L0–L4 peaks obtained from simulations at the low mutation rate (see Fig. 1a). The copy-loop (highlighted in green) is the major source of difference between these sequences. Some of the sample sequences from each of these peaks are given in Additional file 1: Fig. S13. The fitness ranges used to isolate these peaks are given in Additional file 1: Table S2. b Box plots depicting merits (speed of the organism’s virtual CPU, a proxy for metabolic activity of the genotype) of 100 genotypes sampled from each of L0–L4 on the fitness distribution of simulations performed at a low mutation rate. Genomes from L0 and L1 are metabolically less active compared to L2, L3, and L4 (statistical significance measured using the two-tailed unpaired Wilcoxon test; ***p < 0.001). c Task complexities for genomes (estimated roughly by the number of “NAND” instructions required to perform a task, increases from left to right) from L0–L4 peaks on the fitness distribution obtained from simulations at a low mutation rate. Higher fitness peaks are found to perform more complex metabolic tasks. The Y axis represents the number of times that a genotype belonging to these peaks performs a task over a single execution of the genome. The data for this figure is also given in Additional file 1: Table S3