Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
. 2024 Sep 23.
doi: 10.1039/d4sc04107k. Online ahead of print.

AI empowering traditional Chinese medicine?

Zhilin Song  1   2 , Guanxing Chen  3 , Calvin Yu-Chian Chen  1   2   4   5   6
Affiliations
Review

AI empowering traditional Chinese medicine?

Zhilin Song et al. Chem Sci. .

Abstract

For centuries, Traditional Chinese Medicine (TCM) has been a prominent treatment method in China, incorporating acupuncture, herbal remedies, massage, and dietary therapy to promote holistic health and healing. TCM has played a major role in drug discovery, with over 60% of small-molecule drugs approved by the FDA from 1981 to 2019 being derived from natural products. However, TCM modernization faces challenges such as data standardization and the complexity of TCM formulations. The establishment of comprehensive TCM databases has significantly improved the efficiency and accuracy of TCM research, enabling easier access to information on TCM ingredients and encouraging interdisciplinary collaborations. These databases have revolutionized TCM research, facilitating advancements in TCM modernization and patient care. In addition, advancements in AI algorithms and database data quality have accelerated progress in AI for TCM. The application of AI in TCM encompasses a wide range of areas, including herbal screening and new drug discovery, diagnostic and treatment principles, pharmacological mechanisms, network pharmacology, and the incorporation of innovative AI technologies. AI also has the potential to enable personalized medicine by identifying patterns and correlations in patient data, leading to more accurate diagnoses and tailored treatments. The potential benefits of AI for TCM are vast and diverse, promising continued progress and innovation in the field.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Workflow of the integration of AI in TCM research and applications. The process begins with data mining of TCM databases, which collate historical and contemporary data on various herbs. These herbs are then analysed to identify their key ingredients, which are subsequently subjected to AI-driven analysis for various applications including chemical component identification, drug discovery, personalized treatment planning, network pharmacology, and elucidating pharmacological actions. The outcome of these analyses informs various TCM applications such as drug compatibility, diagnostic strategies, and personalized treatment plans.
Fig. 2
Fig. 2. Comparison of the TCM prescriptions, herbs, chemical ingredients, targets, and diseases that the databases contain.
Fig. 3
Fig. 3. The road map of TCM modernization. This diagram outlines the integration of computational methods and databases in modernizing TCM. Starting with disease analysis, it utilizes gene ontology and signaling pathways to explore metabolic and gene targets. These inform the regulation of TCM-linked metabolic pathways, analyzed through TCMbank and metabolome databases. Virtual drug screenings via iScreen and big data analysis refine TCM formulas, integrating modern technology to advance TCM applications.
Fig. 4
Fig. 4. Overview of TCM prescription network design and optimization approaches. (A) Five basic elements that prescription networks are composed of. (B) Three fundamental natural processes that prescriptions are composed of. (C) Four basic models for operating on targets. (D) Prescriptions can be simplified into three elements in order to facilitate design: highly contributed nodes (HCN), lowly contributed nodes (LCN), and intermediate (which fall in between the previous two); two types of conditions appear in non-ideal networks: one is contributions being capable of balancing the network, and the other is trending towards HCN and not being able to become balanced; non-ideal networks can be sub-divided into various sub-networks. Prescription design is divided into (E) bottom up formulas design and (F) top down formulas design. Top down formulas design can be further divided into five basic design models: replacement, addition, enhancement, reduction, and removal. Both top down and bottom up prescription design has the potential to (G) be able to take different Chinese medicine theories into consideration when engaging the adjustment and control of different disease networks and abnormal nodes. (H) Natural product categorization and organization is capable of expanding the designable range of most TCM prescription optimizations. Biological data, providing a holistic view of patient health and enhancing the precision of TCM treatments.
Fig. 5
Fig. 5. Case study of AI integration in TCM formula classification using LLM. (A) Illustrates the data collection from textbooks of TCM formulas, input from TCM experts, and compilation of experimental data into training, test, and validation datasets. (B) Depicts the utilization of various LLMs, including pre-trained models such as ChatGLM-130b and InternLM-20b, and fine-tuning specific models like ChatGLM-6b and ChatGLM2-6b for TCM application. The workflow demonstrates how experimental data feeds into model training and validation, with TCM experts evaluating the final outcomes.
Fig. 6
Fig. 6. AI analysis of TCM chemical data and analysis of application examples. (A) AI-driven TCM chemical data analysis workflow. For the complexity and diversity of herbal chemical components, machine learning and neural networks have been applied to the preprocessing, feature extraction and analysis of herbal chemical data, ultimately predicting biological activities and identifying new drug candidates. (B) Case study. Predicting associations between meridians and TCM using cost-sensitive graph convolutional neural networks.
Fig. 7
Fig. 7. AI analysis and application cases of TCM chemical composition identification. (A) AI-driven pattern recognition workflow for TCM compound analysis. From data collection to integration and analysis of complex chemical structures and their biological activities. (B) Case analysis. An example of computer identification of hepatotoxic components of TCM, including the process of data collection from the TCM database to model validation and virtual screening of hepatotoxic herbs.
Fig. 8
Fig. 8. AI analysis and application cases of TCM herbal ingredients. (A) AI-based TCM herbal medicine analysis process. AI starts from the initial herbal input to complex chemical profile analysis and finally the intrinsic mechanism of the herbal medicine through AI and data mining. (B) Case study of Qiang Huoshengshi decoction – intelligent chemical analysis strategy guides the accurate quantification of multiple components of TCM compound. Apply UHPLC-Q-TOF/MS analysis to detailed steps for identifying herbal species-specific chemical classes using deep learning models, enabling precise chemical analysis quantitation. (C) Case analysis of using a data-driven approach to identify potential treatments for COVID-19 from TCM. Through molecular docking and network pharmacology research, the utilization process of TCM prescription database to high-frequency TCM and main drugs is demonstrated.
Fig. 9
Fig. 9. AI analysis of TCM drug discovery and its application cases. (A) AI-based TCM drug discovery process. Apply AI technology to perform virtual screening from Chinese medicine databases to target and identify active chemical ingredients, followed by new compound identification and in vitro validation. (B) Case study of effective clues to gain insights into Alzheimer's disease through the use of multiple AI algorithms. Screening is carried out using structure-based and ligand-based methods, then AI and QSAR models are developed to continue screening, and finally validated through molecular dynamics simulations.
Fig. 10
Fig. 10. Application of AI in analysing HDIs. (A) General workflow for analysing drug incompatibilities in TCM using AI. It illustrates the steps from TCM formulations to the prediction of herb–drug interactions (HDI) through AI-driven models, emphasizing the identification of antagonistic effects and the integration of diverse data. (B) Case study of AI application in HDI analysis. It outlines the process from data collection to molecular docking and verification, focusing on the application of graph embedding techniques like node2vec to predict new potential targets within HDIs.
Fig. 11
Fig. 11. A comprehensive proposal for automatic clinical study system. The framework is divided into three main modules: clinical study system, integrated system and fundamental study system. The clinical study system module includes stages such as surgical sample extraction, chemical analysis, and data mining. This workflow aims to build a validation system to rigorously test and confirm observed interactions between herbal medicines and pharmacological substances. The integrated system serves as a link between clinical and basic science data, facilitating the integration of disparate data sets through platforms such as HAZA@home. It supports comprehensive analysis at home, allowing users to perform preliminary disease assessments using integrated network pharmacology tools. The fundamental study system focuses on the theory and mechanism of drug interactions. The module employs a variety of computational and experimental methods, including target prediction, ADMET simulations, and construction of molecular databases. Aims to elucidate the molecular basis of drug interactions through in-depth chemical analysis and simulation modeling.
Fig. 12
Fig. 12. Application of AI in TCM diagnosis. (A) General workflow for AI applications in TCM diagnosis. It illustrates the process of utilizing AI to enhance TCM diagnosis through disease prediction, risk assessment, early detection, intervention strategies, and precise diagnostics, leveraging chemical data from TCM sources. (B) Case study on lung cancer diagnosis and therapeutic insights. It shows the integration of an optimized SVM for serological data analysis, network pharmacology, and experimental techniques to uncover the mechanisms of wogonin in lung cancer.
Fig. 13
Fig. 13. AI-enhanced personalized TCM treatment planning. (A) General process for AI applications in personalized treatment planning. This workflow illustrates how AI integrates patient-specific and TCM chemical data to model disease progression and predict treatment outcomes, ultimately guiding the development of personalized treatment plans. (B) Case study on optimizing herbal formulas using AI for personalized treatment planning. It depicts the use of genetic algorithms to analyse and derive optimal herbal formulas from TCM datasets, validating their effectiveness and functional impact through systematic analysis.
Fig. 14
Fig. 14. Applications of AI in network pharmacology. (A) General process for AI applications in network pharmacology. This workflow demonstrates how AI integrates disease, genetic, and TCM chemical data to identify node importance, network modules, and ultimately construct a comprehensive network model. (B) Case study on network medicine framework. It illustrates the use of a network medicine framework to analyse the relationships between herbal prescriptions, patient symptoms, and genetic data to localize symptoms, identify key patterns, and establish a network-based medicinal strategy.
Fig. 15
Fig. 15. Utilization of AI in elucidating pharmacological actions. (A) General process for AI applications in elucidating pharmacological actions. This showcases the workflow where AI integrates TCM chemical data to analyse multi-component and multi-target interactions, employ data mining and pattern recognition techniques, and ultimately elucidate pharmacological mechanisms. (B) Case study on elucidating pharmacological actions through network pharmacology analysis. This diagram illustrates the detailed steps in network pharmacology analysis starting from the TCMS database to the integration of various networks including herb-compound, compound-protein, gene ontology, and pathway enrichment, contributing to the understanding of pharmacological actions.
See this image and copyright information in PMC

References

    1. Ma H. D. Deng Y. R. Tian Z. Lian Z. X. Traditional Chinese medicine and immune regulation. Clin. Rev. Allergy Immunol. 2013;44:229–241. - PubMed
    1. Yan Z. Lai Z. Lin J. Anticancer Properties of Traditional Chinese Medicine. Comb. Chem. High Throughput Screening. 2017;20:423–429. - PubMed
    1. Cyranoski D. Why Chinese medicine is heading for clinics around the world. Nature. 2018;561:448–450. - PubMed
    1. Yuan H. Ma Q. Ye L. Piao G. The Traditional Medicine and Modern Medicine from Natural Products. Molecules. 2016;21:559. - PMC - PubMed
    1. Cai Y. Luo Q. Sun M. Corke H. Antioxidant activity and phenolic compounds of 112 traditional Chinese medicinal plants associated with anticancer. Life Sci. 2004;74:2157–2184. - PMC - PubMed

LinkOut - more resources