全基因组测序与脉冲场凝胶电泳在万古霉素耐药屎肠球菌医院监测中的比较

《Microbiology Spectrum》：Comparing whole-genome sequencing to pulsed-field gel electrophoresis for vancomycin-resistant Enterococcus faecium hospital surveillance

　　

万古霉素耐药肠球菌（Vancomycin-Resistant Enterococci, VREs）已成为重要的公共卫生问题，也是医疗相关感染的主要致病原之一。在院感暴发情境下，准确区分亲缘关系较近的VRE菌株对于感染控制措施至关重要。尽管脉冲场凝胶电泳（Pulsed-Field Gel Electrophoresis, PFGE）被视为VRE菌株分型的金标准，但近期研究已证实全基因组测序（Whole-Genome Sequencing, WGS）是一种具有更高分辨率的替代方案。本研究中，研究人员报道了一套WGS及单核苷酸多态性（Single-Nucleotide Polymorphism, SNP）分析流程在大型学术医学中心进行万古霉素耐药屎肠球菌（Vancomycin-Resistant Enterococcus faecium, VREfm）菌株分型中的验证与应用价值。以PFGE为参照方法，研究人员确定了≤10个SNP的保守阈值，该阈值能够稳健识别相关VREfm分离株，灵敏度为98.5%，特异度为93.1%。比较WGS SNP菌株分型前后VREfm感染控制监测结果发现，WGS SNP组中具有相关菌株型和流行病学关联的分离株比例更高。这些发现表明，与PFGE相比，WGS SNP分析能够提高新兴医院暴发事件和感染控制措施的分辨率。

distinguished ability to differentiate closely related antibiotic-resistant bacteria represents a critical task for implementing effective infection control measures within hospital environments. This study focused specifically on strains of vancomycin-resistant Enterococcus faecium, a high-priority nosocomial pathogen according to the 2024 World Health Organization Bacterial Priority Pathogens List, and systematically evaluated whether a whole-genome sequencing single-nucleotide polymorphism analysis pipeline could provide enhanced discrimination of strain relatedness compared to the established gold standard reference method. The research was conducted at Northwestern Memorial Hospital and Prentice Women's Hospital, encompassing a comprehensive transition from traditional pulsed-field gel electrophoresis to advanced genomic approaches for hospital surveillance. The findings demonstrated that whole-genome sequencing-based strain typing identified a proportionally greater number of transmission and outbreak events compared to the reference method over comparable one-year surveillance periods, with 28% of isolates in the post-whole-genome sequencing cohort exhibiting related strain types and epidemiological links versus 20% in the pre-whole-genome sequencing pulsed-field gel electrophoresis cohort. The clinical and epidemiological significance of these findings suggests that when a whole-genome sequencing pipeline is available and properly validated, it can substantially improve hospital infection control measures by providing more precise and actionable information about nosocomial pathogen transmission dynamics. This work was published in the journal Microbiology Spectrum.

The authors utilized several key technical methodologies in conducting this research. The primary approach involved developing and validating a comprehensive whole-genome sequencing single-nucleotide polymorphism analysis pipeline benchmarked against pulsed-field gel electrophoresis for VREfm strain typing. The study employed clinical VREfm isolates obtained from September 2022 to November 2024 through routine hospital surveillance screening of rectal swabs from patients admitted to high-risk units at Northwestern Memorial Hospital and Prentice Women's Hospital. For molecular characterization, the researchers implemented reference-based genome assembly using the Illumina sequencing platform, specifically utilizing the E. faecium strain NRRL B-2354 as the reference genome based on multi-locus sequence typing results. The bioinformatic analysis was performed using CLC Genomics Workbench software, incorporating quality control metrics for DNA quality, sequencing read quality, reference strain coverage, and process run controls. A critical methodological component involved determining a conservative single-nucleotide polymorphism threshold of ≤10 for defining related strains, which was validated against pulsed-field gel electrophoresis classifications and subsequently applied prospectively. The transition between methods allowed for direct comparison of surveillance outcomes, with epidemiological investigations conducted by the hospital infection prevention team to assess geo-temporal, personnel, and procedural overlaps among patients with matching strain types.

Establishing a Whole-Genome Sequencing Single-Nucleotide Polymorphism Pipeline for VRE Strain Typing

The researchers initially focused on constructing and validating a robust whole-genome sequencing single-nucleotide polymorphism pipeline for VREfm strain typing. A total of 58 unique VREfm isolates with predetermined pulsed-field gel electrophoresis strain type designations, collected during routine hospital VRE surveillance from September 2022 to November 2024, were selected for validation. Given the six-week rolling comparison window used in standard surveillance, the researchers opted for a reference-based genome assembly strategy. To select an appropriate reference genome, a K-mer tree was constructed using an initial sub-sample of three isolates and nine complete genomes from E. faecium reference strains belonging to Clades A1, A2, and B. Inspection of the resulting K-mer tree revealed that the initial set of isolates exhibited closer relatedness to Clades A1 and A2 compared to Clade B, with minimal K-mer distance separating the isolates from Clades A1 and A2. Based on multi-locus sequence typing results obtained via the Bezdicek et al. E. faecium multi-locus sequence typing scheme in the CLC Genomics Workbench software, strain NRRL B-2354 was selected for all subsequent reference-based genome assemblies.

Quality Control and Reproducibility of the WGS SNP Pipeline

To develop a robust whole-genome sequencing single-nucleotide polymorphism pipeline, the researchers identified critical quality control criteria encompassing DNA quality post-library preparation, sequencing read quality, reference strain coverage, and the use of process run controls. To assess reproducibility, a set of 18 unique isolates consisting of both related and distinct strains (as classified by pulsed-field gel electrophoresis) was tested twice by two different technologists following identical sample preparation and sequence analysis procedures. The two independent runs demonstrated highly concordant results with no more than ten single-nucleotide polymorphism deviations for a given comparison. To further assess precision, two different technologists performed separate sequencing runs consisting of six whole-genome sequencing libraries prepared from the same VREfm strain (strain EF7202). Pairwise comparisons between these runs identified no more than one single-nucleotide polymorphism for a given comparison, indicating that the quality control criteria ensured a robust assay with minimal inter-operator and inter-run variability.

WGS SNP Pipeline Comparison to PFGE

After establishing adequate quality control criteria, the whole-genome sequencing single-nucleotide polymorphism pipeline was tested on 58 unique VREfm isolates with known pulsed-field gel electrophoresis classifications from a total of 183 pairwise strain comparisons. The researchers assessed the distribution of single-nucleotide polymorphism differences between two isolates as compared to the pulsed-field gel electrophoresis band differences for the same pairwise comparison. A conservative threshold of ≤10 single-nucleotide polymorphism differences for a "related strain" classification by whole-genome sequencing single-nucleotide polymorphism yielded nearly complete concordance with the same pulsed-field gel electrophoresis classification (i.e., pulsed-field gel electrophoresis strain comparisons with ≤3 band differences). Considering ≤3 pulsed-field gel electrophoresis band differences as the gold standard "true positive" reference and ≥4 band differences as "true negatives," the sensitivity of whole-genome sequencing single-nucleotide polymorphisms was calculated as 98.5% and the specificity as 93.1%. Given the objective of facilitating epidemiological link detection during surveillance, the researchers favored high sensitivity and considered a threshold of ≤10 single-nucleotide polymorphisms as suitable criteria to trigger warranted transmission investigations.

VREfm Outbreak Investigation Outcomes Pre- and Post-WGS SNP Adoption

Upon completing validation, the clinical microbiology laboratory transitioned to strain typing exclusively with the whole-genome sequencing single-nucleotide polymorphism pipeline. The researchers investigated outcomes before and after whole-genome sequencing adoption by assessing hospital transmission and outbreak investigations during a one-year period where pulsed-field gel electrophoresis was exclusively used (November 2023–2024) compared to the following year with whole-genome sequencing single-nucleotide polymorphism (November 2024–2025). Similar numbers of VREfm isolates were analyzed between cohorts (691 and 680 isolates in the pre- and post-whole-genome sequencing single-nucleotide polymorphism cohorts, respectively). The results showed that a significantly greater proportion of isolates with related strain types and epidemiological links were identified in the post-whole-genome sequencing single-nucleotide polymorphism cohort (28%) compared to the pre-whole-genome sequencing single-nucleotide polymorphism (i.e., pulsed-field gel electrophoresis) cohort (20%). Similarly, among strain type clusters with at least one identified epidemiological link, significantly larger clusters were identified via whole-genome sequencing single-nucleotide polymorphism as compared to pulsed-field gel electrophoresis (24 and 29 total strain type clusters, respectively). The researchers interpreted these results as suggesting that the whole-genome sequencing single-nucleotide polymorphism pipeline more readily identifies related strains, facilitating downstream identification of bona fide epidemiological links. However, they noted the inability to distinguish between increased false positives versus decreased true negatives as the chief contributor to pulsed-field gel electrophoresis performance from the current study design.

The discussion section elaborated on several key advantages of the whole-genome sequencing single-nucleotide polymorphism workflow over pulsed-field gel electrophoresis. First, pulsed-field gel electrophoresis band difference assessment is prone to user error regarding gel alignment and relies on potentially inconsistent qualitative interpretation between users, particularly for comparisons between separate gels, whereas whole-genome sequencing single-nucleotide polymorphism allows for high-precision, reproducible strain typing using pre-defined quality control criteria with quantitative classification basis. Second, the whole-genome sequencing single-nucleotide polymorphism threshold yielded larger strain type clusters on average for downstream epidemiological investigation. Third, whole-genome data collection enables additional downstream applications, such as detection of selection pressures and genome sweeps within local VRE isolate clusters. The authors acknowledged limitations including inability to detect large genome rearrangements known to occur among VREfm sub-clades and sensitivity of single-nucleotide polymorphism detection to reference genome choice. The single-nucleotide polymorphism threshold used (≤10) was relatively comparable to recent studies (ranging from ≤6 to ≤16 single-nucleotide polymorphisms) but was specifically derived from their dataset rather than pre-defined, considering potential variable mutation rates among VREfm strains. A previous study estimated a mutation rate of seven single-nucleotide polymorphisms per genome per year for nosocomial E. faecium, suggesting the threshold is sufficiently relaxed for a six-week rolling period.

研究结论部分翻译如下：总而言之，WGS SNP有望成为识别涉及VRE及其他可能的高优先级病原体院内传播的稳健菌株分型方法。