Fig. 6

The abundance of the risk bacteria was negatively associated with immune infiltration. A Sankey plots show the significantly differential pathways according to malignant property-, metabolism-, and immune-related categories between NPC tumours with high (high-risk group) or low (low-risk group) abundances of risk bacteria based on GSEA. In the bubble plots, the size of the bubbles represents gene numbers in the labelled pathway. The colour of the bubbles represents that the labelled pathway is upregulated in the high-risk (red) or low-risk (blue) groups. The permutation-based P value shows the statistical significance of the NES. B Network plot shows the Spearman correlations (P<0.05, |r|>0.30) among four risk bacterial genera and significantly differentially expressed immune-related genes. C Violin plots show the immune infiltration in the high- or low-risk groups estimated by the MCP-counter algorithm. The comparison was performed with the Wilcoxon rank-sum test. The dots represent the mean scores, and the error bars represent the standard deviation. D Representative images of immune cell infiltration in NPC tissues in the high- or low-risk group determined by multiplex immunohistochemistry (mIHC). The 6-colour mIHC images (1 mm), high magnification (100 μm) of a region of interest, and individual images of DAPI (purple), panCK (pink), CD3 (blue), CD8 (red), CD20 (yellow), and CD56 (green) are shown. (E) Box plots show the comparison of immune cell infiltration in the intratumoural (left) and stromal (right) tissues in the high- and low-risk groups. The comparison was performed with the Wilcoxon rank-sum test. The centreline represents the median, and the bounds represent the 25th and 75th percentiles. The upper whisker extends from the hinge to the largest value no further than 1.5*IQR from the hinge. The lower whisker extends from the hinge to the smallest value of at most 1.5*IQR of the hinge