Skip to main content
. 2019 Jul 30;10(4):e01054-19. doi: 10.1128/mBio.01054-19

FIG 4.

FIG 4

(A) A graph generated using random forest model proximity scores, trained to distinguish individuals. It is thresholded by proximity less than 0.076. It is colored by Top Module, the highest-level clustering produced by Infomap. Module 1 is mostly composed of shoe and floor samples, similarly to that shown in Fig. 3. (B) Significant Spearman correlations (P < 0.05) between each module and various metadata categories.