Fig. 2.

Diversity of efficient motif sets. Diversity of efficient motif sets under four levels of compression ($q=0,0.1,0.2,0.3$) as more assemblies are incorporated, measured by genome-wide number of efficient motifs. The curve of $q=0$ reflects the number of original motifs. As expected, the number of the original motifs increases as genomes are added, due to the inclusion of rare motifs. The efficient motif set size increases as more assemblies are included, with total diversity in this category showing asymptotic leveling at 30 genomes. As q grows larger, more compression is posed on the original motifs, resulting in less efficient motifs