Lee Won-Yung, Park Kwang-Il, Bak Seon-Been, Lee Seungho, Bae Su-Jin, Kim Min-Jin, Park Sun-Dong, Kim Choon Ok, Kim Ji-Hwan, Kim Young Woo, Kim Chang-Eop
School of Korean Medicine, Wonkwang University, Iksan 54538, Republic of Korea; Research Center of Traditional Korean Medicine, Wonkwang University, Iksan 54538, Republic of Korea; School of Korean Medicine, Woosuk University, Jeonju 54986, Republic of Korea.
Department of Veterinary Medicine, Research Institute of Life Science, Gyeongsang National University, Jinju 52828, Republic of Korea.
J Adv Res. 2024 Dec 25. doi: 10.1016/j.jare.2024.12.040.
Network pharmacology has gained significant traction as a tool for identifying the mechanisms and therapeutic effects of herbal medicines. However, despite the usefulness of these approaches, their diversity underscores the critical need for a systematic evaluation to ensure consistency and reliability.
We aimed to evaluate the network pharmacological analyses, focusing on identifying the mechanisms and therapeutic effects of herbal medicines.
We employed a comprehensive approach involving systematic data retrieval, network construction, and analysis. Herbal compounds and their targets were meticulously extracted from five distinct network pharmacology databases to ensure extensive coverage and high data reliability. Advanced network-based methods were used to identify key herbal targets and predict therapeutic effects, thereby enriching the depth and breadth of the analysis. Experimental validation was performed on prostate cancer models to substantiate the computational predictions.
The results of the recapitulating task for known herbal ingredient targets revealed distinct patterns in performance and coverage based on network construction and aggregation methods. We performed the same analysis to identify herbal targets and found that network centrality, path counts, and downweighted path counts had their own pros and cons. By comparing network-based methods, we found that considering the impact on the multiscale interactome yielded the highest accuracy in discriminating known therapeutic effects. Using optimal conditions, we successfully identified new indications for herbal medicines and validated these findings through follow-up in vitro and in vivo experiments.
This study presents the first comprehensive and critical evaluation of the current network pharmacology analyses in the field of herbal medicine and provides valuable guidance for continued advances in the elucidation of the mechanisms and therapeutic effects.
网络药理学作为一种识别草药作用机制和治疗效果的工具已获得显著关注。然而，尽管这些方法有用，但其多样性凸显了进行系统评估以确保一致性和可靠性的迫切需求。
我们旨在评估网络药理学分析，重点是识别草药的作用机制和治疗效果。
我们采用了一种综合方法，包括系统的数据检索、网络构建和分析。从五个不同的网络药理学数据库中精心提取草药化合物及其靶点，以确保广泛覆盖和高数据可靠性。使用先进的基于网络的方法识别关键草药靶点并预测治疗效果，从而丰富分析的深度和广度。在前列腺癌模型上进行实验验证以证实计算预测。
对已知草药成分靶点的重现任务结果显示，基于网络构建和聚合方法，在性能和覆盖范围上存在不同模式。我们进行了相同的分析以识别草药靶点，发现网络中心性、路径计数和加权路径计数各有优缺点。通过比较基于网络的方法，我们发现考虑对多尺度相互作用组的影响在区分已知治疗效果方面具有最高的准确性。在最佳条件下，我们成功识别了草药的新适应症，并通过后续的体外和体内实验验证了这些发现。
本研究首次对草药领域当前的网络药理学分析进行了全面且关键的评估，并为阐明作用机制和治疗效果的持续进展提供了有价值的指导。
