J4-4555-2

The project addresses bacterial infectious diseases as a global health threat and, in particular, the foodborne zoonosis listeriosis, which is associated with the highest mortality rate in the EU. It is caused by bacteria Listeria monocytogenes, which is transmitted through the consumption of contaminated food, and its prevalence is increasing. L. monocytogenes is able to survive and grow in acidic, salty and cold conditions and can colonize food processing environments very successfully. It is thus regularly found on ready-to-eat foods, meat and dairy products, raw vegetables and fruits. The incredible persistence of L. monocytogenes, which is evident from the outbreaks in the EU that span several years, is caused by persistent biofilms. L. monocytogenes isolates have been associated with either persistence in the environment or high pathogenicity potential. The features associated with both greater biofilm persistence and higher pathogenicity that lead to outbreaks are unknown. In the project, we address this issue by using machine learning to investigate the association of biofilm phenotype with molecular surface markers and pathogenicity potential. Biofilms are bacterial consortia enclosed in a self-produced extracellular matrix. They allow bacteria to survive under adverse environmental conditions and also promote antimicrobial resistance. The ability to form biofilms varies from isolate to isolate, and no clear link to genetic information has yet been established.

In this project, the characteristics of biofilm phenotypes of different L. monocytogenes strains (WP1) growing on different surfaces and with different nutrients (WP2) will be investigated. Special attention will be paid to the differences between animal and plant nutrient sources and the comparison of pathogenic and non-pathogenic strains. We will then analyze how these nutrients affect the metabolome, surfactome and glycome (WP3) to find molecular markers of distinct biofilm phenotypes. Finally, their effectiveness in mammalian cell adhesion and invasion will be analyzed to evaluate their pathogenicity (WP4). At the same time, an image analysis toolkit will be developed for biofilm image analysis with enriched data (WP1 and WP2) and extended for multimodal learning with omics-level data (WP3). Finally, pathogenicity potential data will be used to assess the potential computational predictability of strain pathogenicity based on previously identified molecular markers (WP4). Based on this deeper understanding of L. monocytogenes biofilms and the features that enable L. monocytogenes persistence in different environments, we will propose new strategies for more efficient surveillance and prevention of listeriosis outbreaks.