Abstract:Traditional Chinese Medicine (TCM) involves complex compatibility mechanisms characterized by multi-component and multi-target interactions, which are challenging to quantify. To address this challenge, we applied graph artificial intelligence to develop a TCM multi-dimensional knowledge graph that bridges traditional TCM theory and modern biomedical science (<a class="link-external link-https" href="https://zenodo.org/records/13763953" rel="external noopener nofollow">this https URL</a> ). Using feature engineering and embedding, we processed key TCM terminology and Chinese herbal pieces (CHP), introducing medicinal properties as virtual nodes and employing graph neural networks with attention mechanisms to model and analyze 6,080 Chinese herbal formulas (CHF). Our method quantitatively assessed the roles of CHP within CHF and was validated using 215 CHF designed for COVID-19 management. With interpretable models, open-source data, and code (<a class="link-external link-https" href="https://github.com/ZENGJingqi/GraphAI-for-TCM" rel="external noopener nofollow">this https URL</a> ), this study provides robust tools for advancing TCM theory and drug discovery.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to quantify and explain the complex compatibility mechanisms in traditional Chinese medicine (TCM), especially the interactions of multi - components and multi - targets. Specifically, the paper aims to combine traditional TCM theory with modern biomedical science through Graph Neural Networks (GNNs) and Knowledge Graph technology, so as to realize the quantitative analysis of the compatibility mechanisms of Chinese medicine. ### Main problems and challenges: 1. **Quantification of complex interaction relationships**: Herbal formulas in TCM (CHF) involve complex interactions among multiple components, and these interactions are difficult to quantify. 2. **Interpretability of the model**: Existing models are insufficient in explaining the action mechanisms among herbs and cannot provide clear insights into the mechanisms. 3. **Combination of theory and computational framework**: Current model designs often fail to fully incorporate the core theories of TCM, resulting in a disconnection between the theoretical basis and the computational framework. 4. **Lack of high - quality data**: The lack of high - quality, multi - dimensional and publicly shared TCM data sets limits interdisciplinary cooperation and innovation. ### Solutions: - **Construct a multi - dimensional TCM Knowledge Graph (TCM - MKG)**: Integrate traditional TCM concepts with modern biomedical data to form a knowledge graph with a multi - layer information structure. - **Introduce virtual nodes**: Introduce virtual nodes in graph encoding to represent drug properties and improve the modeling of the compatibility mechanisms of herbal formulas. - **Use graph neural networks and attention mechanisms**: Model and analyze 6,080 herbal formulas through GNNs and attention mechanisms to quantify the effects of key herbs therein. - **Open - source code and data**: Provide open - source data, models and code to support TCM research and practical applications. ### Verification methods: - Use 215 herbal formulas for COVID - 19 management for verification to prove the effectiveness of the method. - Through feature ablation and node masking analysis, reveal the key factors in the prediction mechanism. ### Conclusion: This research has successfully quantified the compatibility mechanisms in TCM through graph neural networks and attention mechanisms, and provided new ways to promote the modernization of TCM theory and drug discovery. In particular, the research has revealed the core role of Radix Astragali as the "monarch drug" in the COVID - 19 treatment formula, providing a scientific basis for further optimizing existing formulas and exploring the material basis.

Graph Neural Networks for Quantifying Compatibility Mechanisms in Traditional Chinese Medicine

A method for finding groups of related herbs in traditional chinese medicine

A Knowledge Graph Enhanced Topic Modeling Approach For Herb Recommendation

Database of traditional Chinese medicine and its application to studies of mechanism and to prescription validation.

Modern Bioinformatics Meets Traditional Chinese Medicine.

Traditional Chinese Medicine Prescription Recommendation Model Based on Large Language Models and Graph Neural Networks

Application of a Complex Network Modeling Approach to Explore the Material Basis and Mechanisms of Traditional Chinese Medicine: A Case Study of Xuefu Zhuyu Decoction for the Treatment of Two Types of Angina Pectoris

A KG-Enhanced Multi-Graph Neural Network for Attentive Herb Recommendation

Network-based modeling of herb combinations in traditional Chinese medicine

Multi-target meridians classification based on the topological structure of anti-cancer phytochemicals using deep learning

Semantic-Aware Graph Convolutional Networks for Clinical Auxiliary Diagnosis and Treatment of Traditional Chinese Medicine

AI-assisted literature exploration of innovative Chinese medicine formulas

Model informed Precision Medicine of Chinese Herbal Medicines Formulas- A Multi-scale Mechanistic Intelligent Model

Studies of consomic mice bearing the Y chromosome of the BXSB mouse.

Leveraging Representation Learning for the Construction and Application of a Knowledge Graph for Traditional Chinese Medicine: Framework Development Study

Advances in the Application of Network Analysis Methods in Traditional Chinese Medicine Research

Research and implementation of intelligent question answering system based on knowledge Graph of traditional Chinese medicine

Meta-path guided graph attention network for explainable herb recommendation

TCMNPAS: a comprehensive analysis platform integrating network formulaology and network pharmacology for exploring traditional Chinese medicine

A Knowledge-Guided and Traditional Chinese Medicine Informed Approach for Herb Recommendation

Network Systems Underlying Traditional Chinese Medicine Syndrome and Herb Formula