Abundance, relative frequency, and influence of 182 virulence factors used to classify 12,337 S . enterica genomes from human, beef, and bovine sources into two virulence groups.

Points represent the square root mean decrease Gini impurity and natural log relative frequency (Cluster 1/Cluster 2–5) of each virulence factor, with diameter proportional to the number of isolates (min: 11, max: 11,653) carrying at least one copy of the virulence gene. Color designates virulence mechanism categories, as derived from the VFDB [ 32 ] and BV-BRC [ 33 ] databases. Vertical dashed line represents equal frequency of virulence factors between the clusters, while points to the right of this line represent factors more frequently found in cluster 1. Points above the horizontal dashed line (square root of the mean decrease Gini impurity = 12.5) represent virulence factors that were more influential differentiators in classifying isolates.