Figure 3
Impact of experimental procedures on the variation of mock oral microbiota. Bray-Curtis distance was calculated and compared between all pairs of samples collected from the expected microbiota and the 54 mock ones. (A) The perceived microbiome structure within and between DNA extraction methods and choice of hypervariable regions. DNA extraction method (R2 = 0.764, p < 0.001, Adonis) dominated the effect on oral microbiota structure over choices of primer sets (R2 = 0.21, p < 0.001, Adonis). (B) PCoA plot of the expected and all the 54 mock communities with the Bray-Curtis distance. For the DNA extraction methods, apart from the similar performance between M5 and M6 (p = 0.513, Kruskal-Wallis test), the other extraction methods produced significantly different oral microbiota structures (all p < 0.01, Kruskal-Wallis test). As for the choice of primer sets, results for oral microbiota from P2 (V3-V4) and P3 (V4-V5) were similar (p = 0.825, Kruskal-Wallis test), and both of them were distinct from P1 (V1-V3) samples (p < 0.01, Kruskal-Wallis test).
