Fig. 5
Within-study case vs. control logistic regression results at the genus-level. A Per-study taxonomic enrichment in cases is denoted in red, and enrichment in controls is denoted in blue (N = 30 data sets, 4706 samples). Blank/gray spaces indicate no significant association. Only taxa with at least one significant association are shown. Significant associations are defined as having FDR-corrected two-tailed p-value < 0.05. Black boxes are shown around consistent enrichments within a disease, in which taxa are enriched in the same direction in at least two studies within a disease. Overall disease associations are shown in the last heatmap row, in which enrichment in the same direction in three or more studies than in the opposite direction (Nsame_direction − Nopposite_direction ≥ 3) are considered across-disease significant. B Mean relative abundance across all samples of each taxon shown in A. C Prevalence across all samples for each taxon shown in A