Comparison of core-genome MLST, coreSNP and PFGE methods for Klebsiella pneumoniae cluster analysis

Abstract

In this work we compared the most frequently used Klebsiella pneumoniae typing methods: PFGE, cgMLST and coreSNP. We evaluated the discriminatory power of the three methods to confirm or exclude nosocomial transmission on K. pneumoniae strains isolated from January to December 2017, in the framework of the routine surveillance for multidrug-resistant organisms at the San Raffaele Hospital, in Milan. We compared the results of the different methods to the results of epidemiological investigation. Our results showed that cgMLST and coreSNP are more discriminant than PFGE, and that both approaches are suitable for transmission analyses. cgMLST appeared to be inferior to coreSNP in the K. pneumoniae CG258 phylogenetic reconstruction. Indeed, we found that the phylogenetic reconstruction based on cgMLST genes wrongly clustered ST258 clade1 and clade2 strains, conversely properly assigned by coreSNP approach. In conclusion, this study provides evidences supporting the reliability of both cgMLST and coreSNP for hospital surveillance programs and highlights the limits of cgMLST scheme genes for phylogenetic reconstructions.

Keywords: CG258; K. pneumoniae; cgMLST; cluster.

Fig. 1.

Clonal relationship of 80 K . pneumoniae isolates in a MST based on core-genome multilocus. Each circle represents a single genotype, i.e. an allelic profile based on up to 2358 target genes present in the isolates with the “pairwise ignoring missing values” option turned on in the SeqSphere⁺ software during comparison. The number on connecting lines represents the number of alleles that differ between the connected genotypes. The clusters identified on the cgMLST MST and absent in the MST computed on coreSNP MST (Fig. 2) are marked with asterisks.

Fig. 2.

Clonal relationship of 80 K . pneumoniae isolates in a MST computed on coreSNP distances. Clusters of strains distant <21 SNPs have been identified, coloured and labelled on the graph. The clusters identified on the coreSNP MST and absent in the MST computed on cgMLST allele distances (Fig. 1) are marked with asterisks.

Fig. 3.

Comparison of the ML phylogenetic trees obtained with coreSNP and cgMLST on the OSR dataset (80 strains isolated during OSR routine surveillance program). On the left, the tree obtained with coreSNP and on the right the tree with SeqSphere+cgMLST. The strains belonging to highly represented MLST profiles (>=10 strains) are connected by coloured lines. Among these strains, those included in the routine surveillance program are highlighted on the trees with coloured dots.

Fig. 4.

Comparison of the ML phylogenetic trees obtained with coreSNP and cgMLST on the Global dataset (486 K. pneumoniae strains: 80 isolated during OSR routine surveillance program and 406 from database). Asterisks are reported on nodes with bootstrap supports below 75. On the left, the tree obtained with coreSNP and on the right the tree with SeqSphere+cgMLST. The strains belonging to highly represented MLST profiles (>=10 strains) are connected by coloured lines. Among these strains, those included in the routine surveillance program are highlighted on the trees with coloured dots.