Whole-genome sequencing and machine learning reveal candidate genomic features associated with Listeria monocytogenes persistence in two ice cream facilities.

Abstract

Listeria monocytogenes is a major foodborne pathogen responsible for listeriosis. Ice cream has emerged as an important vehicle for L. monocytogenes transmission, yet genomic data describing strain persistence in ice cream production environments in China remain limited. In this study, we conducted whole-genome sequencing (WGS) of 57 L. monocytogenes isolates recovered from two ice cream production facilities in China between 2016 and 2018 to characterize population structure, persistence, and associated risk factors. Genomic typing revealed dominance of lineage II serotype 1/2a sequence type (ST) 8, alongside lineage I ST5 and ST59. High-quality single nucleotide polymorphism (hqSNP) analysis identified four facility-specific clusters persisting for up to one year. Persistent strains were primarily associated with floor-related environmental niches, while personnel movement and mobile equipment may facilitate strain dissemination within facilities. Core virulence genes and certain antimicrobial resistance traits were conserved across isolates. Mobile genetic elements were widespread, and Stress Survival Islet 1 (SSI-1) was present in all ST8 and ST5 strains. A random forest model showed strong predictive performance of persistent strains and identified SSI-1 genes, ST, and processing area as the main factors associated with L. monocytogenes persistence. Collectively, these results indicate that persistent and closely related L. monocytogenes populations were established in ice cream production environments in China, underscoring the need for targeted environmental monitoring and control strategies.