病院サーベイランスにおけるゲノムの落とし穴：疫学的関連のないエンテロバクター・クロアカの予期せぬクラスター化は、慎重なコアゲノムMLST解析解釈の必要性を浮き彫りにする

Whole-genome sequencing has become a key tool for identifying bacterial cross-transmission in healthcare settings. Among the methods used, core-genome Multi-Locus Sequence Typing (cgMLST) is considered as the reference method owing to its standardization. However, its limitations are often overlooked, particularly for taxonomically complex species such as Enterobacter cloacae. We report a case of false-positive genomic linkage between two E. cloacae isolates recovered from patients hospitalized in different units, four months apart, with no epidemiological overlap. Routine retrospective genomic surveillance using a local cgMLST scheme flagged these isolates as genetically related. However, further analyses based on core-genome Single-Nucleotide Polymorphism (SNPs) calling and molecular phylogeny, including public genomes, revealed that the two isolates belonged to distinct clusters and differed by 99 SNPs. This case illustrates the risk of overinterpreting cgMLST results when applied to species with high genomic heterogeneity and insufficiently robust schemes. These findings highlight the importance of interpreting genomic data within its epidemiological context, and in the case of inconsistency confirming potential links through high-resolution SNP analysis, and developing standardized, species-specific cgMLST schemes.