Bacterial whole-genome sequencing revisited: portable, scalable, and standardized analysis for typing and detection of virulence and antibiotic resistance genes

Abstract

Multidrug-resistant nosocomial pathogens present a major burden for hospitals. Rapid cluster identification and pathogen profiling, i.e., of antibiotic resistance and virulence genes, are crucial for effective infection control. Methicillin-resistant Staphylococcus aureus (MRSA), in particular, is now one of the leading causes of nosocomial infections. In this study, whole-genome sequencing (WGS) was applied retrospectively to an unusual spike in MRSA cases in two intensive care units (ICUs) over the course of 4 weeks. While the epidemiological investigation concluded that there were two separate clusters, each associated with one ICU, S. aureus protein A gene (spa) typing data suggested that they belonged to single clonal cluster (all cases shared spa type t001). Standardized gene sets were used to extract an allele-based profile for typing and an antibiotic resistance and toxin gene profile. The WGS results produced high-resolution allelic profiles, which were used to discriminate the MRSA clusters, corroborating the epidemiological investigation and identifying previously unsuspected transmission events. The antibiotic resistance profile was in agreement with the original clinical laboratory susceptibility profile, and the toxin profile provided additional, previously unknown information. WGS coupled with allelic profiling provided a high-resolution method that can be implemented as regular screening for effective infection control.

FIG 1

MRSA t001 cluster schematic. A comparison of the cluster characterization of the 13 t001 isolates collected between 25 August and 20 September 2003, based on epidemiology and on cgMLST, is shown. Each isolate is represented by a circle with the isolate name indicated (P, patient; S, staff member; E, environment), and isolates belonging to the same cluster are indicated by color. Only affected wards are shown with respect to the building in which they were located; however, as this is only a cartoon model, the true structure and geographic relationship between wards occupying the same building is not depicted here. Orange, cluster 1; green, cluster 2; blue, unrelated isolates. (A) Schematic based on epidemiological investigation. While isolate S1 was found in a staff member who was assigned to both ICU-A and ICU-B at the time of the cluster, the epidemiological data suggested that the MRSA isolate was related to ICU-B. (B) Schematic based on cgMLST data. Orange, cluster 1; green, cluster 2; purple, cluster 3; blue, unrelated isolates.

FIG 2

Genotype to phenotype diagram. Species identification, spa type, antibiotic susceptibility profile, and presence of toxins can be rapidly determined by query of the WGS data. The colored squares represent genes potentially present on the chromosome and/or plasmids. The presence of genes in our cluster isolates are indicated by color: red, antibiotic resistance genes; green, toxin genes; blue, catalase-encoding katA; yellow, f spa gene; gray, genes that were queried but not found.

FIG 3

Clonal relationship of t001 isolates. A phylogenetic dendrogram (UPGMA) was generated for all spa type t001 isolates based on the allelic profiles of 1,714 cgMLST target genes. The scale bars indicate the number of differing alleles comprising the calculated distance. (A) Only temporally related isolates (n = 13) are shown. The ward in which each isolate was isolated is noted next to the isolate name. The index patients for clusters 1 and 2 are labeled. Asterisks indicate a ward that is next to and shares the same patients as the ICU within each respective cluster. (B) cgMLST clusters 1, 2, and 3 are collapsed, and isolates differing in time and/or location of collection are included for phylogenetic placement. The approximate time of isolation is shown relative to the August to September 2003 cluster.