FIG 1.

(a) Heat map. We selected the top 10 discriminative peaks by chi-square testing of the occurrence frequency of peaks in VREfm and VSEfm (see Table S2 in the supplemental material). The heat map was plotted based on hierarchical clustering of all of the VREfm and VSEfm isolates from the CGMH Linkou branch. Rows represent the isolates, and columns represent the top 10 discriminative peaks. The values in the heat map represent the MS spectral intensity, which was log₁₀ normalized and Z-score standardized. Red indicates relatively higher peak intensity, while blue indicates lower peak intensity. The isolates are grouped into five clusters. VREfm and VSEfm isolates can be visually differentiated by using the top 10 discriminative peaks. (b) Intensity of the top 10 important predictors. The logarithms to base 10 of the peak intensities are plotted for VREfm and VSEfm. (c) Occurrence frequency of the top 10 important predictors. The occurrence frequency of the 10 peaks in VREfm and VSEfm is plotted.