Figure 5

E [ t ] for different number of distinct genome s and n = 32, 64 in our adaptive compressed algorithm. For each n, the best and worst population profiles in terms of expected cost are considered. The best case corresponds to roughly uniform n/s cells per distinct genome and the worst corresponds to n − s + 1 cells from one distinct genome and one cell per every other s − 1 distinct genomes. To decouple the effect of genome sizes, the genome size is 4 Mbps for all distinct genomes and τ = 0.2. Capturing probability of all distinct genomes in all cases is 1. Linear interpolated costs are plotted using dashed line for comparison.