Integrating Network Pharmacology, Machine Learning, and Structural Simulation Evidence
Graphical abstract
Summary
Target, pathway, model and structural signals are integrated into a traceable mechanism hypothesis with explicit evidence grades. The graphical abstract connects network, functional and structural results across the evidence layers. The cross-layer results establish focused priorities for targets, pathways and mechanistic investigation. The result-focused presentation supports efficient review of the main evidence and research priorities.
Computational results
PPI protein interaction network
Figure 2 STRING Database derived PPI networks. The nodes represent candidate proteins, the color of the connections corresponding to different sources of evidence; the central region connections are dense, and there is a wide crossover between cell proliferation, death, inflammation and growth factor signaling modules.
The shared targets form a highly connected core in the STRING PPI network, linking functional modules associated with proliferation, stress responses, and inflammation.
Figure 3 Cytoscape redrawing of the PPI network. The color of the nodes increases from light to deep, the position increases with up-to-down reflectance; the bottom deep red nodes are high-connectivity hubs, and the high density on the edges further shows the common target constituting the interconnected core control network.
Cytoscape topology ranking highlights several highly connected hub nodes, establishing a priority target set for machine-learning feature selection and structural review.
Compound–disease–target network
Figure 4 Percentage of the zinc disease target network. The orange is the common target, the purple is the active ingredient, and the green is the node for bladder cancer; the gray line represents the link between the components of zinc targets and the disease targets. Many components in the diagram are connected to multiple targets at the same time, and many targets are also covered by different components.
The compound–disease–target network maps the many-to-many relationships between candidate compounds and targets, supporting layered tracing toward core mechanism modules.
Docking-score comparison
Chart 10 of the molecular docking classification of five candidate components and five machine learning key targets. Figure 10 The molecular docking of the five candidate components and the five key targets of machine learning is shown in the heat chart. The more pale blue the color indicates the lower the combined power, the stronger the predictive affinity; each unit is marked with a Vina rating (kcal/mol).
The docking heatmap creates a unified comparison matrix between machine-learning-prioritised targets and candidate compounds, highlighting combinations for structural review.
Full case PDF
Complete results, high-resolution figures and analysis are provided in the client PDF. Add our service account on WeChat and send HS-CASE-0009 to request it.