Fig. 5

Distance-based redundancy analysis of fecal microbiota samples according to genotypes with diet, genotype, and NASH status as explanatory variables. Samples are colored according to feeding type (blue: standard diet, green: high-fat diet). A EBI2^−/− (dark/light) versus wildtype. B CH25H^−/− (dark/light) versus wildtype. C CYP7B1^−/− (dark/light) versus wildtype. In these plots global centroids for genotype (`KO` vs `WT`) and diet (`STD` vs `HFD`) are shown; a larger distance of the centroids from `[0, 0]` indicate a stronger effect on community composition. The (constrained) vectors of continuous variable effects are shown (for NAS, q.0, q.1, q.2); the direction of the vectors relative to each other indicates if effects are correlated (same direction) or anti-correlated (different direction). q.0, q.1, and q.2 are Hill alpha-diversities; q.0 is (rarefied) taxa richness, q.1 and q.2 are effective taxa numbers weighted by taxa abundances