*2.1. Identification of the Chemical Components of DGS*

HPLC-qTOF-MS/MS was used to analyze the composition of DGS. Based on chromatographic retention times and fragment ion information of the molecules, 37 compounds were identified from the DGS samples (Figure 1). The compound names, retention times (min), and molecular formulae are listed in Table 1. The *m/z* 359.0761 in the negative ion mode corresponds to the [M-H]− ion peak of Rosmarinic acid. The primary and secondary mass spectra of Rosmarinic acid are consistent with the cleavage of primary and secondary given in the MoNA database. Furthermore, a molecular ion peak was seen at *m/z* 609.1415, which corresponded to the [M-H]− ion peak of rutin [20]. The secondary mass spectrum of rutin showed a mass spectral fragment of *m/z* 301, a [M-H-C12H22O10] − mass spectral fragment formed by the loss of two sugar rings of rutin. Salvianolic acid B [21], tanshinone IIA, cryptotanshinone, and hyperoside were respectively identified by the [M-H]− ion peak at m/z 717.1220, *m/z* 293.1192, *m/z* 295.1362, and [M-H]− ion peak at *m/z* 463.0851 in the negative ion mode. Tanshinone IIA, cryptotanshinone, and hyperoside were identified through the NoMB database. Based on mass spectrometry information of primary and second ion fractions, nearly half of the 34 compounds identified from the DGS samples were found to be flavonoids and their glycoside derivatives and the rest were phenolic acids and tanshinones.




35

**Figure 1.** The HPLC-QTOF/MS total ion chromatogram of DGS in the negative ion modes.

*2.2. Key Component and Target Screening and Prediction Using Network Pharmacology Analysis*

A total of 35 compounds were successfully predicted to be the targets of action in the SwissTargetPrediction database. The compounds Piscidic acid and Schaftoside could not be predicted in the SwissTargetPrediction database based on two-dimensional (2D) and threedimensional (3D) similarities. Hence, they were excluded. A total of 442 biological targets were obtained by combining these 35 compounds and removing duplicates. Moreover, 859 CAD-related targets were obtained from the Genecards database using the correlation score of >20 as the screening criterion. A Venn diagram analysis of the compound's target library against the disease's target library yielded 112 incorporative targets (Figure 2A). The top five enriched target categories were protease, enzyme, kinase, nuclear receptor, and secreted protein. As shown in Figure 2B, the top five targets in the protein–protein interaction (PPI) network in terms of degree-value were ALB, TNF, VEGFA, EGFR, and CASP3. Cytoscape 3.9.1 was employed to calculate the network topology data, and the MCODE plug-in was used to calculate the most closely linked clusters in the PPI network. A total of 4 clusters were obtained, with cluster scores of 23.667, 8.147, 3.5, and 3 (Figure 2B'). Cluster 1 with the highest clustering score was used for subsequent analysis.

In the biological process enrichment analysis, Cluster1 was imported into the David database for GO enrichment analysis. Annotations of genes were obtained by using GO analysis. The screening was performed on a *p* < 0.05 basis, and the top 30 biological processes were listed according to the magnitude of the *p*-value (Figure 2C). It chiefly involved cytokine-mediated signaling pathway, positive regulation of MAPK cascade, positive regulation of cell migration, positive regulation of kinase activity, protein autophosphorylation, positive regulation of MAP kinase activity, MAPK cascade, positive regulation of protein phosphorylation, positive regulation of nitric oxide biosynthetic process, positive regulation of ERK1 and ERK2 cascade, positive regulation of cell proliferation, positive regulation of phosphatidylinositol 3-kinase signaling, and negative regulation of endothelial cell apoptotic. These biological processes were mostly involved in the regulation of cytokines, cascade activation of the MAPK pathway, positive regulation of protein phosphorylation, positive regulation of NO synthesis, positive regulation of cell proliferation, anti-endothelial cell apoptosis, etc. The top 30 KEGG signaling pathways at *p* < 0.05 and

sorted according to count value size. As shown in Figure 2D, KEGG pathway enrichment analysis revealed that most of the DGS targets were concerned with the regulation of cell proliferation, anti-apoptosis, cell migration, and angiogenesis.

**Figure 2.** Network pharmacology analysis: (**A**) The Venn diagram analysis for DGS and HCD targets. (**B**) PPI network of the DGS compounds targets against CAD. (**B'**) MCODE analysis of PPI network. (**C**) Biological process analysis of PPI networks with a clustering score of 23.667. \* Represents a potentially important biological process. (**D**) KEGG enrichment analysis of PPI networks with a clustering score of 23.667. \* Represents a potentially important signaling pathway.
