Seqsphere 4

Based on the MLST results, a selection of 20 strains (Table 1) were Illumina sequenced. DNA was extracted, as for the MLST. Sequencing libraries were prepared using the Illumina Nextera DNA Flex chemistry and sequenced on an Illumina MiniSeq (Illlumina, San Diego, CA, USA) with a minimal coverage of 30 ×. After quality control with FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), the reads were assembled with Spades 3.12.0 [40 (link)]. Core-genome multilocus sequence typing (cgMLST) was performed in the software package SeqSphere 4.1.9 (Ridom, Münster, Germany). Assembled genome sequences were imported and blasted against 1701 genes of the seed genome EGD-e, using the standard settings [10 (link)]. A minimal spanning tree was produced in SeqSphere with the options “ignore missing values pairwise” and “discard genomes with >3% missing genes”. Strains with less than 10 different alleles were considered to belong to the same complex.

FASTA files of each isolate were imported into SeqSphere + 4.1.9 (Ridom GmbH, Münster, Germany) for stable cgMLST analysis to identify cluster types (CT) with default parameters. K. pneumoniae NTUH-K2044 (GenBank accession no. NC_012731.1) was used as a reference with a standard set of 2358 genes for gene-by-gene comparison, and the cgMLST comparison table was established. SeqSphere + 4.1.9 was also utilised to obtain neighbour-joining (N-J) and minimum span trees. The genome sequence used for the construction of phylogenetic trees was searched on the National Center for Biotechnology Information (NCBI) database; the isolate identifier and location origin are listed in Table S2.