Table 3.
Pairwise comparisons of saliva and fecal microbial community beta diversities (Bray-Curtis, unweighted Unifrac, weighted Unifrac and Jaccard distances)2 (N=63)
| Saliva microbiome | Fecal microbiome | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Pairwise Comparisons1 | Bray-Curtis | Unweighted UniFrac | Weighted Unifrac | Jaccard distances | Omnibus test | Bray-Curtis | Unweighted UniFrac | Weighted Unifrac | Jaccard distances | Omnibus test |
| T2 vs. T1 | 0.005 | 0.734 | 0.115 | 0.135 | 0.017 | 0.588 | 0.264 | 0.574 | 0.142 | 0.347 |
| T3 vs. T1 | 0.016 | 0.286 | 0.014 | 0.002 | 0.007 | 0.411 | 0.661 | 0.513 | 0.501 | 0.777 |
| T3 vs. T2 | 0.806 | 0.935 | 0.594 | 0.702 | 0.927 | 0.164 | 0.044 | 0.230 | 0.039 | 0.119 |
| T4 vs. T1 | 0.123 | 0.990 | 0.271 | 0.076 | 0.227 | 0.831 | 0.826 | 0.704 | 0.308 | 0.665 |
| T4 vs. T2 | 0.801 | 0.748 | 0.896 | 0.190 | 0.477 | 0.863 | 0.154 | 0.806 | 0.522 | 0.377 |
| T4 vs. T3 | 0.557 | 0.586 | 0.656 | 0.930 | 0.910 | 0.843 | 0.731 | 0.344 | 0.848 | 0.700 |
| T5 vs. T1 | <0.001 | 0.008 | <0.001 | <0.001 | <0.001 | 0.508 | 0.491 | 0.285 | 0.048 | 0.136 |
| T5 vs. T2 | 0.002 | 0.225 | 0.008 | 0.228 | 0.006 | 0.099 | 0.240 | 0.057 | 0.105 | 0.180 |
| T5 vs. T3 | 0.005 | 0.084 | 0.033 | 0.121 | 0.018 | 0.693 | 0.902 | 0.453 | 0.263 | 0.586 |
| T5 vs. T4 | 0.001 | 0.042 | 0.022 | 0.016 | 0.002 | 0.180 | 0.580 | 0.045 | 0.098 | 0.145 |
Pairwise comparisons of interventions (T1=baseline; T2=end of Diet A phase, conventional processed meat for Groups 1 and 2; T3=end of Diet B phase, poultry wash out for Groups 1 and 2; T4=end of Diets C or D phase, conventional processed meats with phytochemicals or low nitrite processed meat with phytochemicals, respectively; T5=end of high nitrate-enriched water + either Diet A, B, C, or D phase
P-values of pairwise comparisons estimated using a Generalized Linear Mixed Model - Microbiome Regression-based Kernel Association Test (GLMM-MiRKAT) adjusted for age (years, continuous), sex, and body mass index (kg/m2, continuous). All statistical tests were two-sided.