1271 draft assemblies

All 1,271 complete Klebsiella genomes that were available in GenBank as of June 2023 were downloaded, encompassing 1,273 chromosome sequences and 4,796 plasmid sequences. Furthermore, for each complete genome, we simulated reads using wgsim with 55x coverage and no error rate (“-N 1000000 -e 0 -1 150 -2 150”). SPAdes was used to perform de novo assembly of the simulated reads, and the plasmid- or chromosome-associated contigs were identified by aligning all contigs back to the corresponding complete genomes using BLASTn, with ≥98.5% identity and ≥80% coverage. A total of 1674 contigs (~1.3 per genome) that failed to align, or aligned to both chromosome and plasmids were excluded. Furthermore, contigs with a length smaller than 1000 bp were also removed because many of the prediction tools were not able to handle them. The final simulated dataset consists of 662,128 contigs from chromosomes and 64,981 contigs from plasmids.