Fig. 4.

Fig. 4. Correlation analysis of differential metabolite and microbiota data after VM treatment. (A) Map of factors affecting differential metabolic pathways after VM treatment. (B) The top 20 metabolites changed most significantly before and after VM treatment. (C) Heatmap of correlations between genus-level microbial abundance and metabolome data. Calculation of the Bray-Curtis distance matrix for the two data sets ‘metabolome’ and ‘microbial composition’ utilizing the R package ‘vegan’, followed by the Mantel test statistical test utilizing the QIIME2 software and the permutation test for the samples (999 times). The statistical significance of the similarity between the metabolomics data and the microbial composition data was assessed (p-value < 0.05) and a p-value = 0.001 was determined, which indicates significance. Using Mothur software, Spearman rank correlation coefficients were calculated between metabolomics data and microbial abundance, and heatmaps were plotted based on the results of the correlation coefficient matrix (rho correlation coefficients are values between -0.6 and 0.6; when -0.6<rho<0, the two are negatively correlated; when 0<rho<0.6, the two are positively correlated; and when rho=0, the two are not correlated). If the correlation between the two is positive, it will be shown in red, and vice versa, if it is negative, it will be shown in blue; the color indicates the strength of the correlation.

© J. Microbiol. Biotechnol.