Fig. 2

Compositional diversity of the gut microbiome. a Histogram of alpha diversity. Richness of each sample is shown as the number of OTUs detected per sample. b, c Non-metric multidimensional scaling (NMDS) analysis of Bray-Curtis distances calculated from OTU-level relative abundances. NMDS 1 and 2 are shown on the x- and y-axis, respectively. Each dot represents an individual and is colored by the proportion of Bacteroidetes (b) or Firmicutes (c) abundance. d Overall phylum-level relative abundance composition across all samples binned by cardiometabolic disease status of CVD, CVD plus diabetes (denoted C+D), diabetes, and no CVD or diabetes. Each sample is represented by one stacked bar colored by phylum. e A 10-year CVD risk score calculated on all participants without a CVD diagnosis. Each dot represents an individual that corresponds to a bar in d and indicates risk for developing CVD as a probability from 0 to 100%. f Boxplots of Shannon diversity across cardiometabolic disease status. Wilcoxon test comparing each disease status to no CVD or diabetes found no significant (ns) difference between Shannon diversity across these categories. g Boxplots of Shannon diversity across binned 10-year CVD risk scores. Wilcoxon test comparing 0–4 to 5–14% and 15–19% CVD risk are significant (* indicates p<0.05). h Boxplots of Shannon diversity across BMI classification. Wilcox test comparing normal BMI to all other BMI categories are significant (* indicates p<0.05). i NMDS analysis of Bray-Curtis distances calculated from OTU-level relative abundances. NMDS 1 and 2 are shown on the x- and y-axis, respectively. Each dot represents an individual and is colored by Shannon diversity. Factors were fit onto this ordination by the envfit function in R. Factors that significantly correlated with vector projections in the ordination space are shown. The strength of the correlation is depicted by the length of the arrow, which points in the direction of the variable that changes most rapidly and with maximum correlation with the ordination configuration. j A bar plot of percent variance explained in Shannon diversity (x-axis) explained by each variable (y-axis). Each variable is colored by its corresponding category. Each variable was tested for association with Shannon diversity by fitting a linear model, and the percent variance explained represents the (sign of the estimate) (r²)*100. Asterisks indicate significance at corrected p<0.05