DPA 框架通过模块重连解析胃癌分期进展并识别治疗脆弱点

Disease progression is driven not only by transcriptional alterations but also by dynamic reorganization of molecular interaction networks. However, most existing approaches focus on differential gene expression and fail to capture modular-level reorganization. Here, we present dynamic perturbation analysis (DPA), a modular rewiring framework that quantifies dynamic perturbations across disease stages using a novel metric termed the degree of perturbation. Using gastric cancer (GC) progression as a model system, DPA revealed that extensive modular rewiring persists despite limited stage-specific transcriptional changes after tumor initiation. DPA identified progression-associated perturbation programs characterized by extracellular matrix remodeling, oncogenic signaling activation, stemness acquisition, and late-stage microenvironmental adaptation. Perturbation analysis revealed two major disease-driving programs and prioritized previously unrecognized regulators, including protein tyrosine phosphatase non-receptor type 12 (PTPN12) and leucine rich repeat and Ig domain containing 1 (LINGO1), whose functional roles in GC progression were experimentally validated. Furthermore, modular rewiring-guided drug repurposing uncovered PJ-34 as a promising therapeutic candidate targeting late-stage disease-associated perturbations. Collectively, DPA provides a generalizable framework for decoding disease progression through dynamic modular rewiring, enabling the identification of driver programs, candidate regulators, and therapeutic vulnerabilities. This framework offers a scalable strategy for network-guided mechanistic discovery and therapeutic prioritization in complex diseases.