Mechanistic elucidation of Panax ginseng in improving atherosclerosis via multidimensional artificial intelligence computing coupled with experimental validation

Figures

• 

  Fig. 1. Detailed work-flowchart of the overall research idea.

• 

  Fig. 2. Identification of the main active components of Panax ginseng and their potential targets. (A) Base peak chromatogram in positive ion mode. (B) Base peak chromatogram in negative ion mode. (C) Structural formulas of the compounds corresponding to the chromatographic peaks in (A). (D) Structural formulas of the compounds corresponding to the chromatographic peaks in (B). (E) Potential target protein predicted based on the main active components of Panax ginseng.

• 

  Fig. 3. Identification of AS-related target genes.(A) Heatmap of DEGs. (B) Volcano plot of DEGs. (C) Venn diagram showing the union between DEGs and WGCNA module genes. (D) Sample dendrogram and trait heatmap. (E) Scale-free topology model fit index (left axis) and mean connectivity (right axis) across different soft-thresholding powers. (F) Hierarchical clustering dendrogram of DEGs based on topological overlap matrix dissimilarity (1-TOM), with color bands indicating the results of automatic single-module analysis. (G) Correlation heatmap between module eigengenes and clinical traits, showing correlation coefficients (cor) and significance (P-values) in control/treat groups. (H) Scatter plot of GS versus MM for the MEturquoise module across groups.

• 

  Fig. 4. Screening and validation of core targets. (A) The intersection between Panax ginseng-related targets and AS-related targets of venn diagram. (B) PPI network of the potential targets. (C) GO and KEGG enrichment analysis of the potential targets. (D) Reverse cumulative distribution of |residuals|. (E) Box plot of |residuals|. (F) Histogram of residuals. (G) Precision recall curve. (H) Variable importance plot (mean importance over 50 permutations). (I) ROC curve. (J) Partial dependence plot.

• 

  Fig. 5. Virtual screening and validation of Panax ginseng-derived compounds targeting AS. (A) Heatmap of molecular docking scores for candidate compounds. (B) 3D binding pose of Adenine within the RXRA binding pocket. (C) 2D interaction diagram of Adenine with key RXRA residues. (D) RMSD of the RXRA-Adenine complex during molecular dynamics simulation. (E) RMSF per residue of RXRA in complex with Adenine. (F) Rg of the RXRA backbone. (G) Number of hydrogen bonds between RXRA and Adenine over time. (H) 3D FEL of the RXRA-Adenine system. (I) 2D contour map of the FEL.(J)CETSA quantitative detection of Adenine -mediated thermal stabilization of RXRA in ox-LDL-induced RAW264.7 foam cells lysates.

• 

  Fig. 6. In vivo validation of the core target RXRA and the efficacy of Adenine. (A) Schematic of the mouse model grouping. (B) The chart of atherosclerosis index. (C) The chart of vulnerability index. (D) Representative H&E staining and immunofluorescence images of RXRA protein expression after treatment with different concentrations of Adenine. (E) The chart of RAP. (F) Statistical analysis of RXRA protein expression levels after treatment with different concentrations of Adenine. (G) Normalized stacked bar chart of various indicators relative to the MG. (H) Clustered heatmap of normalized distributions of different indicators relative to the MG. (I) Heatmap of Pearson correlation analysis among different indicators across groups. (J) Mantel test correlation heatmap of Adenine effects on protein expression, AI, VI, and RAP. Data are expressed as mean ± SD. Symbols # and ## were denoted relative to CG P < 0.05 and P < 0.01, respectively.