Detection of antibiotic resistance genes in MRSA whole-genome sequencing: database comparison and antimicrobial susceptibility correlation

AIM: This study aims to elucidate the characteristics of various databases for detecting antimicrobial resistance (AMR) genes using whole-genome sequencing (WGS). BACKGROUND: WGS offers a comprehensive detection of AMR genes and has been increasingly utilized to monitor AMR bacteria. Although variou...

Full description

Saved in:
Bibliographic Details
Main Authors: Norihito Kaku, Yasuhide Kawamoto, Fujiko Mitsumoto Kaseida, Kosuke Kosai, Hiroo Hasegawa, Koichi Izumikawa, Hiroshi Mukae, Katsunori Yanagihara
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Journal of Global Antimicrobial Resistance
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2213716524002911
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:AIM: This study aims to elucidate the characteristics of various databases for detecting antimicrobial resistance (AMR) genes using whole-genome sequencing (WGS). BACKGROUND: WGS offers a comprehensive detection of AMR genes and has been increasingly utilized to monitor AMR bacteria. Although various databases are employed for detecting AMR genes, each characteristic and their correlation with conventional antimicrobial susceptibility testing (AST) remains unclear. METHODS: WGS data were obtained from 270 methicillin-resistant Staphylococcus aureus (MRSA) strains collected from bloodstream infections in 45 institutions across Japan. AMR genes were detected using AMRFinderPlus, Comprehensive Antibiotic Resistance Database (CARD), and ResFinder. AST was conducted using the microdilution method. CLSI and EUCAST breakpoint criteria were used for antimicrobial susceptibility interpretations. Diagnostic accuracy was calculated using R version 4.4.1. RESULTS: Number of detected AMR genes in AMRFinderPlus, CARD, and ResFinder was 48, 53, and 42, respectively. The average of detected AMR genes per strain in three AMR genes databases was 8.3±0.2, 18.6±0.1, and 7.1±0.2, respectively. Among the AMR genes, 21 genes were detected in all databeses (Figure 1). In 267 strains determined to be methicillin-resistant by AST, mecA gene was detected by AMRFinderPlus, CARD, and ResFinder in 219 (82.0%), 262 (98.1%), and 267 (100%), respectively. The database with the highest diagnostic accuracy varied depending on the antibiotic and the breakpoint criteria used (Table 1). For clindamycin, the diagnostic accuracy was low, in the 0.5 range, across all databases (Table 1). CONCLUSIONS: AMR genes databases varied in characteristics; further research is needed to determine their optimal use in practice.
ISSN:2213-7165