Fig. 4: Function and disease association of microbially targeted human proteins.

a, Odds ratios of representative functional annotations enriched among effector-targeted human proteins (FDR < 0.05, Fisher’s exact test with Bonferroni FDR correction). Terms (#) show the number of represented terms. The lowest and highest odds ratios observed for the represented group are indicated by light shaded areas of bars. Black line indicates odds ratios for shown representative terms. White triangles indicate functions also enriched in pathogen targets. b, Genetic predisposition for traits and diseases enriched among human genes encoding effector-interacting proteins in HuRI (α = 0.05, Fisher’s exact test; n = 349). The odds ratio in a and b estimates the effect size of significant function/trait (two-sided Fisher’s exact test FDR < 0.05) and is calculated as the odds of function-annotated/trait-associated human genes encoding effector targets to function-unannotated/trait-unassociated human genes encoding effector targets in the target set, divided by the same ratio in the HuRI set (see Methods). c, Disease groups for which genetic predisposition proteins are enriched in network neighbourhoods of effectors of the indicated strains. Trait node size corresponds to number of significantly targeted traits in that group as indicated in the legend. Thickness of strain–group edges reflects the number of underlying significant effector–trait links (α < 0.01 and odds ratio > 3, two-sided Fisher’s exact test). d, Specific diseases underlying the ‘immunological’ group in c. Node size reflects the number of underlying effector–trait associations as indicated in the legend. Precise P values and n for all tests are provided in Supplementary Data 23.