Gut microbiome dynamics and cardiometabolic disease progression: evidence from a secondary data analysis of pooled clinical datasets

Figures

• 

  Figure 1. Box plots of Shannon’s diversity revealed statistically significant differences in phyla-level alpha diversity for sex (A) and genus-level alpha diversity for age (B) and study (C). PCoA plots of Bray-Curtis diversity revealed statistically significant differences in phylum-level beta diversity for age (D) and study (E), and genus-level beta diversity for sex (F), age (G), and study (H). Shannon’s diversity was analyzed utilizing nonparametric Mann–Whitney/Kruskal-Wallis statistical tests. Bray-Curtis distances were analyzed using PERMANOVA statistics. Statistical significance was assigned as p < 0.05 and a false discovery rate <0.05.

• 

  Figure 2. Box plots of genus level Shannon’s and Chao1 diversity indices for clinical stratifications of BP (A,B), RHI (C,D), LDL (E,F), TC (G,H), and BMI (I,J) revealed no statistically significant differences in alpha diversity. Shannon’s diversity indices were analyzed utilizing nonparametric Mann–Whitney/Kruskal-Wallis statistical tests. Statistical significance was assigned as p < 0.05 and a false discovery rate <0.05. BP, blood pressure; RHI, reactive hyperemia index; LDL, low density lipoprotein; TC, total cholesterol; BMI, body mass index.

• 

  Figure 3. PCoA of genus level Bray-Curtis distances showed no significant separation of the gut microbiota across clinical stratifications of BP (A), RHI (B), LDL (C), TC (D), or BMI (E), as calculated by PERMANOVA. Statistical significance was assigned as p < 0.05 and a false discovery rate <0.05. BP, blood pressure; RHI, reactive hyperemia index; LDL, low-density lipoprotein; TC, total cholesterol; BMI, body mass index.

• 

  Figure 4. Heatmaps of all differentially abundant taxa (A), and differentially abundant taxa identified in ≥2 comparisons (B), at the ASV level for cardiometabolic parameters. Statistical significance was assigned as p < 0.05 and a false discovery rate <0.1. Direction of fold change is indicated by – or +. BP, blood pressure; RHI, reactive hyperemia index; LDL, low-density lipoprotein; TC, total cholesterol; BMI, body mass index.

• 

  Figure 5. Co-occurrence networks of stratifications in cardiometabolic and anthropometric parameters. Color of nodes denotes modules in each network. Line color denotes positive (green) and negative (red) correlation coefficients. Networks were constructed using the SpiecEasi network method with SPARCC correlations, using a correlation coefficient threshold of 0.6, a p-value cutoff of 0.01, and FDR p-value adjustment. BP, blood pressure; RHI, reactive hyperemia index; LDL, low density lipoprotein; TC, total cholesterol; BMI, body mass index.

• 

  Figure 6. Positive and negative cohesion analyses revealed significant differences in microbial network connectivity for BP (A,B), RHI (C,D), LDL (E,F), TC (G,H), and BMI (I,J). Robustness analyses revealed differences in microbial network ability to resist collapse upon targeted node and edge removal for BP (K), RHI (L), LDL (M), TC (N), and BMI (O). BP, blood pressure; RHI, reactive hyperemia index; LDL, low density lipoprotein; TC, total cholesterol; BMI, body mass index.