Fig. 2.

Variables predicting population structure and cross-species sharing of bacterial MGEs at taxonomic animal-animal and animal-human interfaces across Nairobi. (A) Results of variance partitioning, showing the variance in MGE community structure explained by variables representing evolutionary, host ecology, and environmental factors. These values were derived from a distance-based redundancy analysis (dbRDA) performed on a distance matrix of MGE sharing between hosts. (B) % deviance explained by environmental variables included within generalized additive models (GAMs) exploring the impact of biotic and anthropological environmental variables that are hypothesized to influence host dynamics and contact rates on centrality within multilayer TPNs (mTPNs). The results from three different categories of models are presented, each representing a different measure of centrality across ten pairwise taxonomic networks. In EC, ODC, and OCC models, eigenvector centrality, Opsahl degree centrality, and Opsahl closeness centrality were measured as the response variable, respectively.