Fig. 4

Left: average runtime over 10 runs for finding the node paths corresponding to the occurrences of 100 000 Illumina reads of length 101 bp and their reverse complement to the pan-genome graph of 10 human genomes, as a function of the checkpoint sparseness factor \(s_{ cp }\) (8 to \(\infty\)). To find the occurrences, we performed approximate pattern matching with a maximum allowed number of errors of \(K=4\), using the search scheme proposed by Kucherov et al. The pan-genome is built for \(k=25\), \(k=50\) and \(k=75\), and \(s_{\textrm{SA}} = 16\). The 95% confidence intervals for the runtime are also indicated. Right: total memory usage of the \(\textrm{IDmap}\) component as a function of \(s_{ cp }\)