Table 1.
Percent of variation in X explained and R2 for prediction of B⋅1 through B⋅9 for the Tobacco data.
Fig 1.
Scree plots for variance explained by each component.
For each method, the contribution of the first 20 (of 44) components plotted. Note the change of scale for unconstrained regression. Components are ordered by decreasing eigenvalue of the Unifrac distance matrix.
Fig 2.
Scree plots for variance explained by each OTU.
For each method, the contribution of the 20 (of 271) OTUs having the largest contributions are plotted. Note the change of scale for unconstrained regression.
Table 2.
Correlation between values for each method.
Table 3.
Taxonomic assignment for OTUs selected as a top 5 OTU in explaining variability, and the variability explained by each OTU (d = 3 components).
Fig 3.
Biplot for second and third component of tobacco bacterial data.
Points are colored by type (blue = dry, red = moist, green = toombak) and samples corresponding to replicates of the same product are plotted with the same symbol. The taxonomic families corresponding to the OTUs shown are Staphylococcaceae (4312974), Aerococcaceae (52399), Lactobacillaceae (4379247), Enterococcaceae (29012) and Corynebacteriaceae (810425). The scale on bottom and left corresponds to coordinates of samples, scale on right and top to coordinates of OTUs.