Krüger, Gabriel I.Pardo-Esté, CoralÁlvarez, JavieraPacheco, NicolásCastro-Severyn, JuanAlvarez-Thon, LuisSaavedra, Claudia P.2025-04-092025-04-09202500236438https://repositorio.unab.cl/handle/ria/64066Indexación: Scopus.Salmonella, a zoonotic pathogen, is commonly transmitted through contaminated animal products. This bacterium is emerging in poultry production, often exhibiting multidrug resistance (MDR) and high virulence. Understanding the adaptive mechanisms that allow Salmonella to survive in hostile environments and become virulent is crucial for preventing outbreaks that threaten both the industry and public health. This study uses machine learning to identify adaptive genomic signatures in Salmonella isolates from the poultry industry, focusing on responses to environmental stressors. Significant genomic modifications were found in functions like membrane and cell wall biogenesis, amino acid metabolism, and inorganic ion metabolism, including genes related to antibiotic resistance and virulence. The machine learning model demonstrated high precision (0.980) and accuracy (0.954) in classifying isolates based on their genomic characteristics, with an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.98. The model identified Salmonella Infantis as one of the most stress-resistant serovars in the poultry industry. The identification of critical genomic sequences underscores the importance of these traits in understanding the bacterium's adaptive mechanisms. These findings underscore the importance of genomic surveillance and advanced bioinformatics to manage emerging pathogens like Salmonella Infantis. © 2024 The AuthorsenGeneticsMachine learningPoultry farmResistanceSalmonellaArtículoAttribution-NonCommercial-NoDerivatives 4.0 International CC BY-NC-ND 4.0 Deed10.1016/j.lwt.2024.117188