30Gene regulatory networks (GRNs) have been widely used as a fundamental tool to reveal the 31 genomic mechanisms that underlie the organism's response to environmental and developmental 32 cues. Standard approaches infer GRNs as holistic graphs of gene co-expression, but such 33 graphs cannot quantify how gene-gene interactions differentiate among organisms and how 34 they alter structurally across spatiotemporal gradients. Here, we develop a generalized 35 framework for inferring informative, dynamic, omnidirectional, and personalized GRNs 36 (idopGRNs) from routine transcriptional experiments. This framework is constructed by a 37 system of quasi-dynamic ordinary differential equations (qdODEs) derived from the combination 38 of ecological and evolutionary theories. We reconstruct idopGRNs from a clinical genomic study 39 and illustrate how network structure and organization affect surgical response to infrainguinal 40 vein bypass grafting and the outcome of grafting. idopGNRs may shed light on genotype-41 phenotype relationships and provide valuable information for personalized medicine. 42 43 Key words: gene regulatory network, evolutionary game theory, niche biodiversity theory, 44 community ecology, ordinary differential equation, variable selection 45 46 47 49that guide the organism's response to changes in their environment 1,2 . One promising subject 50 of research in modern biology and translational medicine is how to infer biologically realistic 51 and statistically robust GRNs from increasingly available transcriptional data and link them to 52 physiological, pathological, and clinical characteristics [3][4][5] . A number of statistical approaches, 53 such as Boolean networks 6 , Bayesian networks 7 , mutual information theory 8,9 , and graphical 54 models 10 , have been developed for network inference, and these approaches visualize GRNs as 55 probabilistic, undirected or unidirectional graphs, where each node represents a gene and edges 56 depict relationships between genes. However, such graphs may not be sufficiently informative 57 for charting the topological structure of a GRN because genes may regulate and also be regulated 58 by other genes, with regulations in various signs and strengths and varying across time and space 59 scales 3,11 .As the time generalization of Bayesian networks, dynamic Bayesian networks (DBNs) can code 61 cyclic, causally directed, and probabilistic interactions into networks through temporal 62 interdependence, but they are often puzzled by the choice of granularity when time spaces vary 12-63 14 . When gene networks are modeled by a system of time-derivative ordinary differential 64 equations (ODEs), all these issues can be mostly addressed [15][16][17][18] . The successful use of such ODE-65 based networks is, however, impaired by two factors: (1) parametric dynamic modeling, which is 66 difficult to justify, given that gene expression is often stochastically fluctuated 19,20 and alters 67 across discrete regimes, such as cell/tissue types and medical tre...