Fig. 1

Runtime as a function of genome size. Note that both the x-axis and y-axis are log-scaled and the results for various genomes are indicated at the corresponding genome size. For all standalone scaffolders, runtimes include the mapping stage. For SOAPdenovo2 and ALLPATHS-LG we report the runtime of the full assembly pipeline (including contig assembly). OPERA-LG (scaffolding only) shows the runtime of the scaffolding algorithm in OPERA-LG, excluding preprocessing, to highlight that it can take a fraction of the overall runtime and is influenced more by the repeat complexity of the genome than the size of the genome. Due to its library-size restrictions, ALLPATHS-LG could be run on only a few of the datasets shown here. SOPRA runs did not complete after 10 days for C. elegans and the three largest genomes