Establishment and evaluation of a core genome multilocus sequence typing scheme for whole-genome sequence-based typing of Pseudomonas aeruginosa

Tönnies H, Prior K, Harmsen D, Mellmann A

Abstract

Abstract The environmental bacteriumPseudomonas aeruginosa, in particular multidrug resistant clones, is often associated with nosocomial infections and outbreaks. Today, core genome multilocus sequence typing (cgMLST) is frequently applied to delineate sporadic cases from nosocomial transmissions. However, until recently, no cgMLST scheme for a standardized typing ofP. aeruginosa was available.To establish a novel cgMLST scheme forP. aeruginosa, we initially determined the breadth of theP. aeruginosapopulation based on MLST data with a Bayesian approach (BAPS). Using genomic data of representative isolates for the whole population and for all 12 serogroups, we extracted target genes and further refined them using a random dataset of 1,000P. aeruginosagenomes. Subsequently, we investigated reproducibility and discriminatory ability with repeatedly sequenced isolates and isolates from well-defined outbreak scenarios, respectively, and compared clustering applying two recently published cgMLST schemes.BAPS generated sevenP. aeruginosagroups. To cover these and all serogroups, 15 reference strains were used to determine genes common in all strains. After refinement with the dataset of 1,000 genomes, the cgMLST scheme consisted of 3,867 target genes, which are representative for theP. aeruginosapopulation and highly reproducible using biological replicates. We finally evaluated the scheme by reanalyzing two published outbreaks, where the authors used single nucleotide polymorphisms (SNPs) typing. In both cases cgMLST was concordant to the previous SNP results and to the results of the two other cgMLST schemes.In conclusion, the highly-reproducible novelP. aeruginosacgMLST scheme facilitates outbreak investigations due to the publicly available cgMLST nomenclature.