Abstract
BACKGROUND: Vancomycin-resistant enterococci (VRE) are successful nosocomial pathogens able to cause hospital outbreaks. In the Netherlands, core-genome MLST (cgMLST) based on short-read sequencing is often used for molecular typing. Long-read sequencing is more rapid and provides useful information about the genome's structural composition but lacks the precision required for SNP-based typing and cgMLST. Here we compared prophages among 50 complete E. faecium genomes belonging to different lineages to explore whether a phage signature would be usable for typing and identifying an outbreak caused by VRE. As a proof of principle, we investigated if long-read sequencing data would allow for identifying phage signatures and thereby outbreak-related isolates. RESULTS: Analysis of complete genome sequences of publicly available isolates showed variation in phage content among different lineages defined by MLST. We identified phage present in multiple STs as well as phages uniquely detected within a single lineage. Next, in silico phage typing was applied to twelve MinION sequenced isolates belonging to two different genetic backgrounds, namely ST117/CT24 and ST80/CT16. Genomic comparisons of the long-read-based assemblies allowed us to correctly identify isolates of the same complex type based on global genome architecture and specific phage signature similarity. CONCLUSIONS: For rapid identification of related VRE isolates, phage content analysis in long-read sequencing data is possible. This allows software development for real-time typing analysis of long-read sequencing data, which will generate results within several hours. Future studies are required to assess the discriminatory power of this method in the investigation of ongoing outbreaks over a longer time period.