Exploratory transcriptomic and network pharmacology prioritization of candidate targets and phytochemicals in vascular cognitive impairment

Vascular cognitive impairment (VCI) is closely associated with cerebral small vessel disease and is linked to increased disability, institutionalization and mortality, highlighting the need to identify biologically relevant therapeutic candidates. This study applied an exploratory transcriptomic and network pharmacology framework to prioritize candidate targets and phytochemicals relevant to VCI-associated molecular dysfunction. Analysis of the GSE201482 whole-blood dataset identified 1163 nominally differentially expressed genes between VCI and SCD/SVD comparator samples. Weighted gene co-expression network analysis identified trend-associated turquoise and green modules, yielding 352 overlapping DEG--WGCNA candidate genes. Integration with medicinal plant-derived compound targets identified 52 common candidate genes. Functional enrichment linked these genes to membrane-potential regulation, synaptic compartments, G-protein-coupled receptor signalling, oxidative stress response, hypoxia response, neuroactive ligand signalling and amino-acid metabolism. Immune deconvolution provided exploratory immune-context information on macrophage, monocyte, NK cell and T cell populations. Machine-learning feature prioritization using LASSO, Random Forest and SVM-RFE identified ACP1, CDC25A and MAO-B as candidate genes selected by all three algorithms. External assessment using the independent GSE303449 human brain RNA-seq dataset added a validation layer for candidate interpretation, although FDR-level support for the 52-gene signature and ML-derived panels was limited. MAO-B was selected for docking because it is a druggable neuroactive enzyme with established inhibitor biology and an experimentally resolved ligand-bound structure, enabling reference-based comparison with rasagiline. Docking analysis highlighted isorhamnetin, quercetin and oroxylin A as top-ranked phytochemicals against MAO-B. Overall, this study provides a reproducible candidate-prioritization workflow for guiding future experimental evaluation of multi-target phytochemical approaches in VCI.