PCA-Enhanced Deep Features for Alzheimer's Disease Stage Classification with EFMM.

Hepatic inflammation serves as a central pathological mechanism in the progression of various liver diseases, with its molecular basis involving abnormal activation of inflammation-related genes and disruption of regulatory networks [–]. In chronic hepatitis B, persistent hepatitis B virus (HBV) infection can induce host immune response disorders, driving aberrant expression of key inflammatory regulatory genes such as and , thereby establishing a chronic inflammatory microenvironment that significantly increases the risk of hepatic fibrosis, cirrhosis, and hepatocellular carcinoma [,]. Although nucleos(t)ide analogs and pegylated interferon can effectively suppress viral replication, their efficacy in reversing established inflammatory damage and immune exhaustion remains limited [,]. Therefore, targeted synergistic regulation of multiple inflammatory gene nodes has emerged as a critical strategy for developing novel anti-hepatic inflammation approaches. Fort. (Banlangen), a classic antipyretic and detoxifying herbal medicine in Traditional Chinese Medicine (TCM), has been demonstrated by modern research to possess significant antiviral, anti-inflammatory, and immunomodulatory activities [,]. However, due to the chemical complexity of , the traditional “single-component–single-target” research paradigm inadequately explains its synergistic regulatory mechanisms on inflammatory gene networks, thereby constraining deeper mechanistic understanding and clinical translation. Network pharmacology, by constructing multi-dimensional regulatory networks of “component–target gene–pathway,” aligns with the holistic perspective and systemic regulatory characteristics of TCM, providing a novel strategy for deciphering gene-level mechanisms of complex herbal systems []. This integrated strategy has been successfully applied to the study of Food and medicine homology substances. For instance, Liu et al. employed network pharmacology to identify key targets of mulberry leaf extract in metabolic dysfunction-associated fatty liver disease (MAFLD), and then validated their predictions via animal experiments []. While several pioneers have mapped the basic network pharmacology landscape of against HBV, their findings were limited to static network topologies and generic inflammatory hubs []. Such topology-driven approaches are intrinsically biased toward highly interconnected nodes, often yielding recurrent generic signaling hubs across disparate disease models rather than context-specific therapeutic targets []. To overcome the limitations of static network inference, molecular docking and molecular dynamics (MD) simulations have emerged as gold-standard computational tools for validating ligand–target binding modes and conformational stability under physiological conditions []. Integrating network pharmacology with molecular docking and MD simulations constitutes a powerful in silico pipeline that bridges systems-level target prediction with atomic-level structural evidence []. The present work uniquely advances this field by combining network pharmacology with molecular docking and MD simulation, which enables validation of binding modes and stability between bioactive components and target gene products at the protein structural level, offering structural biological evidence for network predictions. Although several network pharmacology studies on have reported similar inflammatory pathways, the present study provides three distinct advances: (1) exclusive application of stringent ‘highest confidence’ (score >0.9) PPI filtering to reduce false positives; (2) 100 ns all-atom MD simulation to validate binding stability, which is rarely performed in previous studies; and (3) identification of IQ as a potential AKT1 stabilizer, a hypothesis not previously proposed.