Salmonella comprises more than 2,600 serovars. Very few environmental and uncommon serovars have been characterized for their potential role in virulence and human infections. A complementary in vitro and in vivo systematic high-throughput analysis of virulence was used to elucidate the association between genetic and phenotypic variations across Salmonella isolates. The goal was to develop a strategy for the classification of isolates as a benchmark and predict virulence levels of isolates. Thirty-five phylogenetically distant strains of unknown virulence were selected from the Salmonella Foodborne Syst-OMICS (SalFoS) collection, representing 34 different serovars isolated from various sources. Isolates were evaluated for virulence in 4 complementary models of infection to compare virulence traits with the genomics data, including interactions with human intestinal epithelial cells, human macrophages, and amoeba. In vivo testing was conducted using the mouse model of Salmonella systemic infection. Significant correlations were identified between the different models. We identified a collection of novel hypothetical and conserved proteins associated with isolates that generate a high burden. We also showed that blind prediction of virulence of 33 additional strains based on the pan-genome was high in the mouse model of systemic infection (82% agreement) and in the human epithelial cell model (74% agreement). These complementary approaches enabled us to define virulence potential in different isolates and present a novel strategy for risk assessment of specific strains and for better monitoring and source tracking during outbreaks.

中文翻译：

---

结合全基因组测序和多模型表型鉴定沙门氏菌毒力的遗传预测因子。

沙门氏菌包含2600多个血清型。很少有环境和罕见的血清型病毒在毒力和人类感染中起潜在作用。使用补充性的体外和体内毒性系统性高通量分析来阐明沙门氏菌分离株的遗传变异与表型变异之间的关联。目标是制定一种策略，将分离株分类为基准，并预测分离株的毒力水平。从沙门氏菌中选出了35株毒力未知的系统发育远缘菌株食源性Sys-OMICS（SalFoS）集合，代表了从各种来源中分离出来的34种不同的血清型。在4种互补感染模型中评估了分离株的毒力，以比较毒力性状与基因组数据，包括与人肠上皮细胞，人巨噬细胞和变形虫的相互作用。使用沙门氏菌的小鼠模型进行了体内测试全身感染。在不同模型之间发现了显着的相关性。我们鉴定了与产生高负担的分离物有关的新的假设的和保守的蛋白质的集合。我们还显示，基于泛基因组的33个其他菌株的毒力的盲目预测在全身感染的小鼠模型（一致性为82％）和人上皮细胞模型（一致性为74％）中很高。这些互补的方法使我们能够确定不同菌株中的毒力潜力，并提出了一种新颖的策略，用于对特定菌株进行风险评估以及在暴发期间更好地进行监测和追踪源。