Fig. 2
From: Efficient inference of large prokaryotic pangenomes with PanTA
The performance of PanTA and existing tools on the three data collections. a Wall time and memory usage among the competing tools at various dataset sizes. Note that PPanGGOLiN and PIRATE were unable to complete the pangenome construction for some large datasets. b The number of core genes and accessory genes of the pangenomes constructed by all the tools. c The concordance in Adjusted Rand Index between the gene clustering and the grouping by gene annotation