Epidemiological information is key when interpreting whole genome sequence data - lessons learned from a large Legionella pneumophila outbreak in Warstein, Germany, 2013

Research output: Contribution to journalResearch articleContributedpeer-review

Contributors

  • Markus Petzold - , Institute of Medical Microbiology and Virology, The ESCMID Study Group for Legionella infections (ESGLI) (Joint first author)
  • Karola Prior - , University Hospital Münster (Joint first author)
  • Jacob Moran-Gilad - , The ESCMID Study Group for Legionella infections (ESGLI), Ministry of Health, Israel, Ben-Gurion University of the Negev (Author)
  • Dag Harmsen - , University Hospital Münster (Author)
  • Christian Lück - , Institute of Medical Microbiology and Virology, The ESCMID Study Group for Legionella infections (ESGLI) (Author)

Abstract

IntroductionWhole genome sequencing (WGS) is increasingly used in Legionnaires' disease (LD) outbreak investigations, owing to its higher resolution than sequence-based typing, the gold standard typing method for Legionella pneumophila, in the analysis of endemic strains. Recently, a gene-by-gene typing approach based on 1,521 core genes called core genome multilocus sequence typing (cgMLST) was described that enables a robust and standardised typing of L. pneumophila. Methods: We applied this cgMLST scheme to isolates obtained during the largest outbreak of LD reported so far in Germany. In this outbreak, the epidemic clone ST345 had been isolated from patients and four different environmental sources. In total 42 clinical and environmental isolates were retrospectively typed. Results: Epidemiologically unrelated ST345 isolates were clearly distinguishable from the epidemic clone. Remarkably, epidemic isolates split up into two distinct clusters, ST345-A and ST345-B, each respectively containing a mix of clinical and epidemiologically-related environmental samples. Discussion/conclusion: The outbreak was therefore likely caused by both variants of the single sequence type, which pre-existed in the environmental reservoirs. The two clusters differed by 40 alleles located in two neighbouring genomic regions of ca 42 and 26 kb. Additional analysis supported horizontal gene transfer of the two regions as responsible for the difference between the variants. Both regions comprise virulence genes and have previously been reported to be involved in recombination events. This corroborates the notion that genomic outbreak investigations should always take epidemiological information into consideration when making inferences. Overall, cgMLST proved helpful in disentangling the complex genomic epidemiology of the outbreak.

Details

Original languageEnglish
Article number17-00137
JournalEurosurveillance
Volume22
Issue number45
Publication statusPublished - Nov 2017
Peer-reviewedYes

External IDs

Scopus 85033785615
PubMed 29162202
PubMedCentral PMC5718391

Keywords

Keywords

  • Base Sequence, Disease Outbreaks, Germany/epidemiology, Humans, Legionella pneumophila/isolation & purification, Legionnaires' Disease/diagnosis, Molecular Epidemiology/methods, Multilocus Sequence Typing/methods, Retrospective Studies, Whole Genome Sequencing