Application of whole-genome sequencing for bacterial strain typing in molecular epidemiology

Abstract

Nosocomial infections pose a significant threat to patient health; however, the gold standard laboratory method for determining bacterial relatedness (pulsed-field gel electrophoresis [PFGE]) remains essentially unchanged 20 years after its introduction. Here, we explored bacterial whole-genome sequencing (WGS) as an alternative approach for molecular strain typing. We compared WGS to PFGE for investigating presumptive outbreaks involving three important pathogens: vancomycin-resistant Enterococcus faecium (n=19), methicillin-resistant Staphylococcus aureus (n=17), and Acinetobacter baumannii (n=15). WGS was highly reproducible (average≤0.39 differences between technical replicates), which enabled a functional, quantitative definition for determining clonality. Strain relatedness data determined by PFGE and WGS roughly correlated, but the resolution of WGS was superior (P=5.6×10(-8) to 0.016). Several discordant results were noted between the methods. A total of 28.9% of isolates which were indistinguishable by PFGE were nonclonal by WGS. For A. baumannii, a species known to undergo rapid horizontal gene transfer, 16.2% of isolate pairs considered nonidentical by PFGE were clonal by WGS. Sequencing whole bacterial genomes with single-nucleotide resolution demonstrates that PFGE is prone to false-positive and false-negative results and suggests the need for a new gold standard approach for molecular epidemiological strain typing.

FIG 1

Numbers of disparate genomic polymorphisms detected among pairwise comparisons of technical replicates and different PFGE classification categories. Pairwise comparisons of isolates are stratified according to their status as whole-genome sequencing technical replicates (T) or classification by PFGE as “indistinguishable” (I), “closely related” (C), “possibly related” (P), or “different” (D). The numbers of genomic variants identified by whole-genome sequencing which distinguished isolate pairs are plotted along the y axis. Results are separately shown for VRE (A), MRSA (B), and A. baumannii (C). Red shading indicates a pairwise comparison with a count of genomic differences within 1 standard deviation of the mean observed in a different PFGE category for the same organism.