Pattern-triggered immunity(PTI)and effector-triggered immunity(ETI)are two main forms of plant immune response to counter pathogen invasion.Up to now,genome-wide regulatory network organization principles leading to quantitative differences between PTI and ETI remain elusive.We combined an advanced machine learning method and modular network analysis to systematically characterize the organization principles of Arabidopsis PTI and ETI.We report our major findings from three network resolutions.At a single network node/edge level,we ranked important genes and gene interactions for immune response and successfully identified many known immune regulators for PTI and ETI,respectively.Topological analysis showed that important gene interactions tend to link network modules.At a subnetwork level,we identified a subnetwork shared by PTI and ETI,which covers 1159 genes and 1289 interactions.In addition to being enriched with interactions linking network modules,it is also a hotspot attacked by pathogen effectors.The subnetwork likely represents a core component to coordinate multiple biological processes in the transition from development to defense.Finally,we constructed modular network models for PTI and ETI to explain the quantitative differences from the global network architecture.Our results show defense modules appeared to be interdependently connected in PTI,but independently connected in ETI,providing an explanation for the robustness of ETI to genetic mutations and effector attacks.Taken together,the multiscale comparisons between PTI and ETI provide a systems biology point of view to understand plant immunity,and highlight the coordination among network modules to establish a robust immune response.