Fig. 5

A The percentage of reads from a simulated coinfection classified by rkmh to each of the HPV16 sublineages, at default settings (k = 16, s = 1000, no pruning, no difference filter). Summing each row of a, with the exception of reads that couldn’t be classified, gives the percent prevalence estimate of each sublineage (b). c The percent of reads classified to each sublineage by rkmh at pruning level M = 100 and I = 1. This significantly improves the prevalence estimates (d)