Fig. 4
From: Alignment- and reference-free phylogenomics with colored de Bruijn graphs
Comparison of running time and accuracy of different methods on the Para C dataset [22] comprising assemblies of \(n=220\) genomes. a Running time for computing phylogenies on random subsamples. Times for SANS include DBG construction with \(k=31\), split extraction and agglomeration. b For different values of k, weakly compatible subsets (bullets) and trees (triangles) have been greedily extracted. Each point on a line corresponds to a different threshold to discard low weighting splits. Note that both axes begin at 0.5