Table 2. Impact of various factors on the microbiome across all samples (n = 292).
Factor | Strata | Unweighted UniFrac | Weighted UniFrac | ||
---|---|---|---|---|---|
R2 | P | R2 | P | ||
Utility & Rig & Pipe & Stagnation | 0.339 | 0.001 | 0.702 | 0.001 | |
Utility.rig | 0.220 | 0.001 | 0.520 | 0.001 | |
Utility | 0.157 | 0.001 | 0.387 | 0.001 | |
Rig | Utility | 0.013 | 0.001 | 0.034 | 0.001 |
Pipe Material | Utility.rig | 0.014 | 0.001 | 0.030 | 0.001 |
Stagnation | Utility.rig | 0.005 | 0.001 | 0.015 | 0.001 |
Batch | Utility.rig.pipe | 0.005 | 0.001 | 0.003 | 0.008 |
Disinfectant type | 0.032 | 0.001 | 0.066 | 0.001 |
Adonis analysis was applied using package “vegan” from R, with permutation = 999. Unweighted UniFrac considers presence/absence of each OTU, while weighted UniFrac also considers relative abundance of each OTU. “Strata” were defined based on sampling design as the overarching factor, with the subsequent hierarchy of factors derived according to relative magnitudes of impact on the microbiome. Further permutations were constrained within a given stratum. For example, when Utility.rig (utility and rig location) was set as the stratum for the purpose of examining the impact of stagnation, samples from the same Utility.rig were pooled and randomized, but samples across different Utility.rig combinations were not.