2.7. Network Pharmacology

Firstly, the target genes of HT were acquired from the Encyclopedia of Traditional Chinese Medicine (ETCM) [36], Comparative Toxicogenomics Database (CTD) [37], and SwissTargetPrediction (STP) [38], respectively (Table 1). The term “hesperetin” was applied to search for the target genes of HT in ETCM and CTD. A simplified molecular input line entry system (SMILES) form of HT (Compound CID: 72281) obtained from Pubchem [39] was keyed in STP for the prediction of target genes. Repeated target genes of HT from the three databases were removed and the others were regarded as target genes of HT (HTTGs).

The list of websites of databases used in this study.

Secondly, to screen the target genes of UC, retrieval was performed in DisGeNET [40], GeneCards [41], and Gene Expression Omnibus (GEO) [42] databases, respectively, with “Ulcerative colitis” as the search term. For target genes of UC from GeneCards, those with relevance scores ≥ 3.0 were chosen. To obtain the UC target genes in the GEO database, two datasets ({"type":"entrez-geo","attrs":{"text":"GSE65114","term_id":"65114"}}GSE65114 and {"type":"entrez-geo","attrs":{"text":"GSE87466","term_id":"87466"}}GSE87466) were employed, followed by normalization with “limma” R package and differential expression gene (DEG) analysis with the criteria |fold change| ≥ 1.5 and adjusted. p < 0.05. Duplicates of DEGs from the two datasets were removed and target genes for UC in GEO were obtained. The genes that existed in more than one database were defined as target genes in UC (UCTGs). Then, the intersection of HTTGs and UCTGs was recognized as the target genes of HT in UC (UCHTTGs), visualized by jvenn [43].

The list of UCHTTGs was input for the construction of the PPI network with the reassessment of protein names in the STRING 12.0 database [44]. “Homo sapiens” was selected and the minimum required interaction score was set as 0.7, which indicated a high confidence in protein interactions. Then, the PPI network was exported and visualized via Cytoscape 3.9.1 software [45]. The node’s score was calculated by the plug-in cytoHubba [45], and the top ten targets ranked by closeness, degree, Maximal Clique Centrality (MCC), Maximum Neighborhood Component (MNC), and radiality value were obtained. Furthermore, an Upset diagram drawn by Xiantao Academic Tools [46] showed that the intersection of the top-ten targets of these five ranking methods (closeness, degree, MCC, MNC, and radiality) were identified as the hub genes of HT in UC.

The GO and KEGG enrichment analysis of UCHTTGs was performed by “clusterProfiler” and “org.Hs.eg.dbo” R package with pvalueCutoff = 0.05 and the visualization of the results was completed with the usage of “ggplot2” R package in R 4.3.1 software.

For molecular docking, firstly, the HT 3D structure in sdf file format was acquired from Pubchem [39] and the core target proteins in pdb or cif file format were downloaded from the Protein Data Bank (PDB) [47]. The target protein in cif file format was transformed to the pdb file format via OpenBabel 3.1.1 software [48]. Molecular docking between HT and its target proteins was performed via CB-DOCK2 [49]. Specifically, the pdb files of HT and its target protein were uploaded to the CB-DOCK2 website, followed by searching cavities and docking within the selected CurPockets. The molecular docking results included binding cavity volume, center of cavities, docking size, 3D view of docking, and the Vina Score (kcal/mol).