Figure 3.
(A) The sums of squares and K-value used to identify the number of groups that the samples should be split into. No clear elbow was evident; therefore silhouette plots were used to examine the data. (B) A silhouette plot showing how it creates metagenomic groups in the data. The most favorable grouping number is where the average silhouette width is nearest to one. (C) The variation of average silhouette width and K. There is a peak at K = 6 and an uptick at K = 10.