Computational Prediction of Candidate Targets and Pathways of Wild Ginseng-Derived Adventitious Root Constituents Against Triple-Negative Breast Cancer: An In Silico Study

Background: Triple-negative breast cancer (TNBC) is an aggressive breast cancer subtype that lacks validated molecular targets and effective targeted therapies. Wild ginseng-derived adventitious root (WGAR) extract was previously shown experimentally to reduce MDA-MB-231 TNBC cell viability (half-maximal inhibitory concentration 79 µg/mL), induce caspase-9/caspase-3-mediated apoptosis, and suppress epithelial-mesenchymal transition (EMT) markers and interleukin (IL)-6 secretion, but that study examined only two docking targets and left the constituent-level targets undefined. The present work is a purely computational (in silico) follow-up study predicting, at the systems level, the candidate targets and pathways underlying this validated phenotype; no new experimental data were generated. Methods: The 15 WGAR constituents previously identified in that study were subjected to in silico target prediction, and the predicted targets were intersected with TNBC-associated disease genes. Shared targets were analyzed by protein-protein interaction network construction and hub (centrality) analysis with Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Reactome enrichment. Four aglycones were docked against six hub targets, protein-ligand interactions were profiled computationally, and absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties were predicted with a machine learning-based predictor. Results: Disease-gene intersection yielded 128 predicted candidate targets forming a network of 128 nodes and 1623 edges centered on signal transducer and activator of transcription 3 (STAT3) and epidermal growth factor receptor (EGFR), with an inflammatory nuclear factor κB-IL-6 module and hypoxia-inducible factor-1 signaling. Enrichment analysis returned 182, 951, and 389 KEGG, GO, and Reactome terms, including cancer, hypoxia, and immune checkpoint pathways. In docking, the aglycones did not separate from the glycosides in predicted binding energy but did so in predicted drug-likeness; the predicted energies at EGFR are not interpreted, since native-ligand re-docking did not reproduce the crystallographic pose at that receptor. The predicted IL-6 module and glycogen synthase kinase-3β hub corresponded to inflammatory and EMT endpoints measured in the previous study, whereas predicted EGFR and STAT3 centrality remains experimentally untested. Conclusions: This in silico analysis nominates STAT3, EGFR, and associated inflammatory and hypoxic pathways as candidate targets of the WGAR aglycones, directionally consistent with the previously validated phenotype. All findings are computational predictions, not experimental observations; they are hypothesis-generating and prioritize the target-level experimental validation to follow.