4.2. Network Pharmacology Analysis

The TCMSP database (https://old.tcmsp-e.com/tcmsp.php, accessed on 10 July 2023), the ETCM database (http://www.tcmip.cn/ETCM/index.php, accessed on 12 July 2023), and the ITCM database (http://itcm.biotcm.net, accessed on 12 July 2023) were utilized to predict all targets associated with the main compounds identified by mass spectrometry. SwissTargetPrediction (http://www.swisstargetprediction.ch/result.php, accessed on 15 July 2023) was employed to predict the compounds that lacked target information, with a focus on targets with a confidence level of 0.8 or higher. These targets were then standardized into gene names, using the UniProt database (https://www.uniprot.org/, accessed on 20 July 2023). Ultimately, the prediction results from all these databases were compiled and organized.

The DisGeNET database (https://www.disgenet.org/, accessed on 10 September 2023), the OMIM database (https://www.omim.org/, accessed on 10 September 2023), and the GeneCards database (https://www.genecards.org/, accessed on 10 September 2023) were accessed to screen targets related to “arthritis”. Subsequently, the data obtained from these sources were merged and converted into standard gene names in Uniprot.

The online tool Venny 2.1 (https://bioinngp.cnb.csic.es/tools/venny/index.html, accessed on 15 September 2023) was employed to intersect the disease targets with the drug targets to identify the possible targets of CF against RA.

The STRING database (https://cn.string-db.org, accessed on 25 September 2023) was used to analyze the CF and RA targets that overlapped. The organism was set to Homo sapiens and the minimum required interaction score was >0.7. Cytoscape 3.7.1 software was used for network topology analysis. Subsequently, the core compounds within this network were analyzed using the CytoNCA 2.1.6 plugin in Cytoscape.

STRING was used to perform KEGG pathway enrichment analysis. We prioritized pathways based on their FDR, from lowest to highest. Among the top 30 pathways, 9 pathways that were related to inflammation were selected for primary analysis. The targets within these 9 inflammation-related pathways, which are directly involved in treating RA, were designated as the primary targets. They were imported into Cytoscape for the construction of the CPS network. Core targets were screened through the CytoNCA plugin for Cytoscape. The DisGeNET database was used to collect targets related to RA. These targets were imported into STRING to construct an RA-related gene network, which was analyzed using the CytoNCA 2.1.6 plugin in Cytoscape. The core targets were screened and ranked based on their degree value to confirm their importance in the pathogenesis of diseases.

The predicted targets were analyzed for GO enrichment using R 4.4.4 software (https://www.r-project.org, accessed on 28 September 2023). ClusterProfiler, GOplot, and Pathview are the major visualization packages included in the R package.

Molecular docking was conducted to validate the interaction between core compounds and their corresponding targets. As a positive control, the known inhibitors of the targets of interest were also selected to perform molecular docking. The SDF format files of the core compounds were downloaded from the PubChem database (https://pubchem.ncbi.nlm.nih.gov, accessed on 10 December 2023) and converted into MOL2 format files. The PDB structure files of the target proteins were acquired from the RCSB PDB database (https://www.rcsb.org, accessed on 11 December 2023). These targets were prepared by removing water molecules and adding hydrogen atoms using AutoDock Tools 1.5.7. Molecular docking simulations were performed with AutoDock Vina 1.2.3, and the results were visualized using PyMOL 2.5.7 software.