AI upgrades network pharmacology for TCM research

Network pharmacology is now a mainstream paradigm in drug discovery and is particularly well matched to traditional Chinese medicine (TCM), where formulas are inherently multi-component, multi-target, and multi-pathway. By embedding network biology principles into pharmacological research, TCM-oriented network pharmacology aims to systematically integrate chemical, target, and pathway information to evaluate therapeutic potential and generate testable mechanistic hypotheses, supporting the modernization of TCM research.

In a new review in the Chinese Journal of Natural Medicines, researchers outline how rapid progress in artificial intelligence (AI), including deep learning, can further advance network pharmacology by strengthening network construction, network inference/analysis, and experimental prioritization/validation in TCM studies. The authors frame AI as a catalyst for shifting from predominantly rule-based or database-dependent workflows toward more data-driven, generalizable, and scalable computational frameworks. The review summarizes a canonical TCM network pharmacology pipeline comprising four stages: ingredient identification, network construction, network analysis, and experimental validation. Within this pipeline, the authors summarize AI-enabled strategies assembling diverse biological and pharmacological networks and for extracting interpretable structure from them. These approaches can help identify influential nodes, functional modules, and pathway-level interactions that may underpin observed therapeutic effects.

Looking beyond established pipelines, the authors emphasize cell–cell communication (CCC)-informed network construction and analysis as a promising next frontier. They outline technical bottlenecks and methodological needs across CCC network building, statistical/algorithmic analysis, and validation, underscoring that robust experimental confirmation and high-quality data standards will be critical to ensure reliability as AI adoption in network pharmacology accelerates.