Fig. 4

Correlations between luminal and circulating metabolome. A) Hierarchical clustering heatmap of gut microbiota using genera. Sub-correlation network of B) luminal metabolome and C) serum metabolome. See Supplemental Fig. 8 for the rest of the network. Debiased Sparse Partial Correlation (DSPC) network was used [51]. Nodes denote taxonomic groups or metabolites; edges represent association measures. Default cutoff value was used for degree filter and betweenness. Correlation significance value < 0.01 used. Sparse partial least squares (or Projection to Latent Space, PLS) [52, 53] to integrate pairwise datasets of D) gut microbiome and luminal metabolome, E) gut microbiome and serum metabolome, and F) luminal metabolome and serum metabolome from paired samples of the same mouse. Samples represented in the latent space from multiple coordinates to demonstrate the level of agreement between the two paired data sets. Arrows connecting paired samples of the same mouse from the indicated two datasets. Cluster Image Map of the Pearson correlation coefficients between two matched datasets [54] of G) gut microbiota and serum metabolome and H) lumen metabolome and serum metabolome. Hierarchical clustering applied on the rows and columns of the similarity matrix simultaneously. The color represents the values of the similarity matrix when performing two dataset integration. Ward linkage to cluster both samples and metabolites based on their Euclidean distance. Color bar indicates the scaled z-score of each feature