Fig. 3
From: The influence of the gut microbiome on BCG-induced trained immunity
Immunomodulatory species strongly influence the serum metabolome. A Canonical correlation analysis (CCA) between abundances of the 345 MSPs and 1607 annotated metabolites across all 321 samples. Twenty-five canonical components with significant increase in explained variance were projected in two dimensions using t-SNE. Positions of three immunomodulatory MSPs (left) and metabolite classes (right) are highlighted. Joint deviation of points from the origin (coordinate 0, 0) reflects increased covariance and grouping between metabolites and MSPs. B MSPs showing effects on cytokine responses have larger effects on metabolite intensities. Distribution of maximum absolute Spearman correlations per MSP across all metabolites. MSPs present in at least 20% of the samples and significant subject to P < 0.05 (left), < 0.1 (center), and < 0.2 (right) in any of the two log-linear models are shown in blue and the remaining non-significant MSPs are in white. Dotted vertical line shows threshold for significantly correlated species and metabolites (PBonferroni < 0.05). C All 32 significantly associated metabolites by 34 MSPs. Colored heat map cells show Spearman correlations with PBonferroni < 0.05. Columns (MSPs) are sorted by minimum P-value from any of the two log-linear models. Asterisks indicate MSPs associated in cytokine responses shown in Fig. 2