Followed by in silico screening of resistance genes using ABRicate [30 ] and the National Center for Biotechnology Information (NCBI) Antimicrobial Resistance Gene Finder Plus database [31 (link)], assemblies passing quality controls were uploaded to the SeqSphere+ software version 7.7.5 (Ridom, Muenster, Germany), which was used for multilocus sequence typing (MLST) following the Achtman scheme and core genome (cg)MLST based on the Enterobase E. coli scheme (2,513 loci). Pairwise allelic differences between isolates were used to construct a neighbour-joining tree with metadata annotated using Interactive Tree Of Life (iTOL) version 6.5.7 [32 (link)]. We further conducted single-nucleotide polymorphism (SNP)-based analyses of the core genome for isolates of specific sequence types (ST). Using the CSI Phylogeny 1.4 server (Call SNPs & Infer Phylogeny; https://cge.cbs.dtu.dk/services/CSIPhylogeny) with default settings, which includes the pruning of SNP within 10 bp, a separate phylogenetic analysis was performed on assemblies of isolates of each ST to construct an SNP distance matrix that was visualised in a heatmap using iTOL. Clusters within the phylogenetic tree corresponding to clonal dissemination were defined based on a pairwise SNP distance of ≤ 100 between isolates, as has recently been suggested for OXA-244-producing E. coli in France [33 (link)]. In this context, it has to be noted that, in the absence of a clear consensus on the maximum cut-off to define clonality, the SNP distance chosen here may appear high compared with thresholds typically used to elucidate local outbreaks. However, to account for our long-term surveillance data, we used this liberal threshold as we expect that identified clusters may reflect patterns on a broader genetic scale, such as clonal lineages. Moreover, we did not correct for recombination which is known to increase genetic diversity. Therefore, we are confident that the SNP distances obtained represent a reliable measure of genetic relatedness between isolates. To illustrate distribution patterns, cluster isolates were mapped based on the first three digits of the postal code.
Free full text: Click here