Network pharmacology as a systems approach to asthma.

Network pharmacology integrates polypharmacology and network biology to decode drug-disease interactions at a systems level, moving beyond single-target paradigms. It maps bioactive compounds to molecular targets and pathways, revealing how multi-component herbal formulations rich in bioactive compounds modulate complex diseases like asthma through synergistic effects. This approach suits asthma research, where pathways like mitogen-activated protein kinase (MAPK) and Janus kinases - signal transducers/ /activators of transcription (JAK-STAT) drive inflammation, by identifying multi-target interventions informed by ADME feasibility. In our study, we applied a structured workflow (Figure 1) combining systems biology with pharmacokinetics to screen herbal compounds for asthma [14,25].

Figure 1.: General workflow of network pharmacology

Common network pharmacology workflow in asthma research

Compound identification and ADME screening

To align with the network pharmacology objective of identifying biologically relevant multicomponent agents, candidate herbal compounds were first retrieved from the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database, PubChem and relevant literature on antiasthmatic herbal formulations [26,27]. These compounds were then subjected to oral bioavailability (OB) ≥ 30 % and druglikeness (DL) ≥ 0.18 filtering using the inbuilt pharmacokinetic predictors in TCMSP, which prioritized constituents with a higher likelihood of systemic exposure. This ADMEdriven triage ensured that only pharmacokinetically feasible compounds were carried forward into downstream target and network analyses [7].

Target identification

For the ADMEfiltered compounds, potential molecular targets were predicted using SwissTargetPrediction, PharmMapper and the Search Tool for Interacting Chemicals (STITCH) database [28,29,30], supplemented by manual extraction of reported targets from published literature. Concurrently, asthmaassociated genes were compiled from GeneCards, DisGeNET, Online Mendelian Inheritance in Man (OMIM), DrugBank and the Therapeutic Target Database (TTD) [31-35]. The intersection of compound-related targets and asthma-related genes was identified using Venn diagram analysis, yielding a core set of candidates, “compound-disease” targets, that were used as the basis for subsequent network construction [36].

Building and evaluating networks

Using the intersected target list, a compound-target-path Joi network was constructed in Cytoscape and PPI network was obtained from the STRING database [37]. The resulting networks were analysed using network topology metrics, such as degree, betweenness centrality and closeness centrality, to identify hub nodes. These hubs were considered putative key regulators within the asthma-herbal interaction network and were carried forward into functional enrichment and docking analyses [38].

Functional enrichment analyses

The hub targets were subjected to Gene Ontology (GO) and KEGG pathway enrichment analysis using an online bioinformatics tool (e.g. Database for Annotation, Visualization, and Integrated Discovery (DAVID)) [39]. GO terms were categorized into biological processes, molecular functions and cellular components, whereas KEGG analysis highlighted specific signalling pathways potentially modulated by the herbal compounds. This step explicitly linked the topological features of the network (hub genes) to functional mechanisms underlying asthma pathophysiology [40].

Docking-based validation

To validate the predicted compound-target interactions at a molecular level, we performed molecular docking simulations between the top bioactive compounds and prioritized hub proteins using AutoDock Vina, Python Prescription (PyRx) or Molecular Operating Environment (MOE) as represented in Figure 2 [41].

Figure 2.: Workflow of molecular docking

Docking scores and interaction patterns (e.g. hydrogen bonds, hydrophobic contacts) were analysed to assess binding affinity and stability. In selected cases, these in silico findings were further supported by experimental validation using cell-based assays or in vivo models, thereby bridging network pharmacology predictions to pharmacological relevance [42].

By integrating computational predictions with biological data, network pharmacology effectively guides experimental validation by prioritizing key compounds, predicting relevant targets and suggesting appropriate biological readouts for testing. Furthermore, it facilitates novel drug discovery by uncovering new therapeutic molecules and signalling pathways that may extend beyond asthma.