Fig. 6

G+C content equilibration of random sequences with 60 and 30% initial G+C contents. Here, the in silico genome is equilibrated by using either the Trek \(r_{i,j}^{core}\) constants with up-to 7-mer context dependence (green lines), or only the singleton rates from Trek without any context dependence (red line, shown for only the 30% G+C content start). The simulation accounting for the context dependence results in 38.19 and 38.18% of G+C content (compare to 40.45% for the repeat-masked human genome), starting from random sequences with 60% and 30% G+C contents respectively. The simulation with only the singleton rates converged at lower 37.36% value for the G+C content, where we expected the least disagreement between the “most-ideal” singleton vs. 7-mer description of the substitution rates (see Fig. 5 a). Please note, that in the hypothetic case of the most ideal singleton substitution rate constants, the discrepancy is more pronounced while analysing the genomic contents of the higher k-mers, as presented in Fig. 5