Fig. 2

Simulated Ultraplexing runs with 10–50 S. aureus genomes, in comparison to perfect (True) and random (Random) assignment of long reads. a The proportion of correctly assigned long reads. b The ∆edit distance for random samples of falsely classified long reads. c The distribution of contigs per assembly. d The distribution of assembly lengths. e The distribution of SNPs per assembly. f The distribution of reference recall. SNPs and reference recall were calculated relative to the utilized reference genomes, and all metrics within the same set of genomes are based on the same simulated short-read data