Characterization of oral virome and microbiome revealed distinctive microbiome disruptions in paediatric patients with hand, foot and mouth disease

Fig. 1. Comparison of salivary virome in healthy, symptomatic and asymptomatic for HFMD samples (n = 38). a Heatmap of the top 40 viruses with the highest relative number of reads detected in saliva samples. b Phylogenetic tree of enterovirus A detected in symptomatic and asymptomatic samples based on the viral protein 1 gene sequences. Symptomatic samples are coloured in green and asymptomatic samples are coloured in orange. Bootstrap values are shown at the branch nodes and the scale bar represents the genetic distance. KKH103 was found to have dual infection of both CV-A5 and CV-A6, labelled as KKH103 1 and KKH103 2. c NMDS ordination plot using Bray–Curtis distance metric of saliva virome with vectors representing correlated viruses. Vectors represent viruses that influence each axis, only significant viruses with p value <0.05 were shown d Boxplots showing the distribution of the relative number of viral reads for each virus detected in the salivary virome. Only the top 40 viruses that were detected with the significance of p value <0.05 between at least two of the cohorts, and viruses that were detected exclusively in one cohort were shown. Boxplots display the median distribution…

Fig. 2. Comparison of salivary prokaryotic microbiome in healthy (n = 17), symptomatic (n = 13) and asymptomatic HFMD samples (n = 13). a Alpha-diversity analysis using Chao1 index, Shannon–Wiener and Simpson index analysis. Boxplots display the median distribution (centre line) and boxes show the interquartile range. Whiskers extending out of the boxes mark the minimum and maximum observed values for each dataset. Kruskal–Wallis test followed by Dwass–Steel–Critchlow–Fligner post hoc test were used to generate p values and is represented as a red asterisk above the boxplot. (*p value < 0.05, **p value < 0.01) b NMDS ordination plot using Bray–Curtis distance metric of saliva prokaryotic microbiome with vectors representing correlated bacteria. Vectors represent bacteria that influence each axis, only significant bacteria with p value < 0.01 were shown.

Fig. 3. Correlation analysis of prokaryotic microbiome. a Pairwise correlation matrix among all bacterial species in (i) healthy, (ii) symptomatic and (iii) asymptomatic populations. Spearman’s correlation was generated using corrplot R package. Only significant correlation with p value < 0.05 were shown. b Correlation analysis between enterovirus A and bacteria in both symptomatic (n = 11) and asymptomatic (n = 6) cohort. Only bacteria with at least Spearman’s correlation >0.4 with enterovirus A were included. The distribution of each sample is represented as a histogram. Bivariate scatter plots of a relative number of reads with a fitted line are displayed with each point representing an enterovirus-infected sample (both asymptomatic and symptomatic). The value of the correlation between the two members of the microbiome was indicated with significance level in red symbols. Significant correlations were marked with either red asterisk or red dot (**p value ≤ 0.001; *p value ≤ 0.01; •p value ≤ 0.05).