Fig. 7

Effects of the major factors on the pathogen resistance as determined by the PLS-PM analysis. a PLS-PM showing the cascading relationships of different factors. An observable variable or a latent variable is represented by a box. The loading for bacterial diversity, the potential keystone taxa, the network complexity, and infection level that create the latent variables are shown in the dashed rectangles. After 1000 bootstraps, path coefficients are calculated and represented by the width of the arrow (red stands for positive relationship, green stands for negative relationship). The dashed arrow indicates a coefficient that did not differ significantly from 0 (P > 0.05). The GoF statistic was used to evaluate the model, and the GoF value was 0.74. b Standardized effects of each factor on zebrafish pathogen resistance profiles calculated from the results of partial least squares path modeling. The direct and indirect impacts are added together to form the total effects