Yuyang Sha, Meiting Jiang, Gang Luo, Weiyu Meng, Xiaobing Zhai, Hongxin Pan, Junrong Li, Yan Yan, Yongkang Qiao, Wenzhi Yang, Kefeng Li
{"title":"HerbMet: Enhancing metabolomics data analysis for accurate identification of Chinese herbal medicines using deep learning.","authors":"Yuyang Sha, Meiting Jiang, Gang Luo, Weiyu Meng, Xiaobing Zhai, Hongxin Pan, Junrong Li, Yan Yan, Yongkang Qiao, Wenzhi Yang, Kefeng Li","doi":"10.1002/pca.3437","DOIUrl":null,"url":null,"abstract":"<p><strong>Introduction: </strong>Chinese herbal medicines have been utilized for thousands of years to prevent and treat diseases. Accurate identification is crucial since their medicinal effects vary between species and varieties. Metabolomics is a promising approach to distinguish herbs. However, current metabolomics data analysis and modeling in Chinese herbal medicines are limited by small sample sizes, high dimensionality, and overfitting.</p><p><strong>Objectives: </strong>This study aims to use metabolomics data to develop HerbMet, a high-performance artificial intelligence system for accurately identifying Chinese herbal medicines, particularly those from different species of the same genus.</p><p><strong>Methods: </strong>We propose HerbMet, an AI-based system for accurately identifying Chinese herbal medicines. HerbMet employs a 1D-ResNet architecture to extract discriminative features from input samples and uses a multilayer perceptron for classification. Additionally, we design the double dropout regularization module to alleviate overfitting and improve model's performance.</p><p><strong>Results: </strong>Compared to 10 commonly used machine learning and deep learning methods, HerbMet achieves superior accuracy and robustness, with an accuracy of 0.9571 and an F1-score of 0.9542 for distinguishing seven similar Panax ginseng species. After feature selection by 25 different feature ranking techniques in combination with prior knowledge, we obtained 100% accuracy and an F1-score for discriminating P. ginseng species. Furthermore, HerbMet exhibits acceptable inference speed and computational costs compared to existing approaches on both CPU and GPU.</p><p><strong>Conclusions: </strong>HerbMet surpasses existing solutions for identifying Chinese herbal medicines species. It is simple to use in real-world scenarios, eliminating the need for feature ranking and selection in classical machine learning-based methods.</p>","PeriodicalId":20095,"journal":{"name":"Phytochemical Analysis","volume":" ","pages":""},"PeriodicalIF":3.0000,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Phytochemical Analysis","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.1002/pca.3437","RegionNum":3,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"BIOCHEMICAL RESEARCH METHODS","Score":null,"Total":0}
引用次数: 0
Abstract
Introduction: Chinese herbal medicines have been utilized for thousands of years to prevent and treat diseases. Accurate identification is crucial since their medicinal effects vary between species and varieties. Metabolomics is a promising approach to distinguish herbs. However, current metabolomics data analysis and modeling in Chinese herbal medicines are limited by small sample sizes, high dimensionality, and overfitting.
Objectives: This study aims to use metabolomics data to develop HerbMet, a high-performance artificial intelligence system for accurately identifying Chinese herbal medicines, particularly those from different species of the same genus.
Methods: We propose HerbMet, an AI-based system for accurately identifying Chinese herbal medicines. HerbMet employs a 1D-ResNet architecture to extract discriminative features from input samples and uses a multilayer perceptron for classification. Additionally, we design the double dropout regularization module to alleviate overfitting and improve model's performance.
Results: Compared to 10 commonly used machine learning and deep learning methods, HerbMet achieves superior accuracy and robustness, with an accuracy of 0.9571 and an F1-score of 0.9542 for distinguishing seven similar Panax ginseng species. After feature selection by 25 different feature ranking techniques in combination with prior knowledge, we obtained 100% accuracy and an F1-score for discriminating P. ginseng species. Furthermore, HerbMet exhibits acceptable inference speed and computational costs compared to existing approaches on both CPU and GPU.
Conclusions: HerbMet surpasses existing solutions for identifying Chinese herbal medicines species. It is simple to use in real-world scenarios, eliminating the need for feature ranking and selection in classical machine learning-based methods.
期刊介绍:
Phytochemical Analysis is devoted to the publication of original articles concerning the development, improvement, validation and/or extension of application of analytical methodology in the plant sciences. The spectrum of coverage is broad, encompassing methods and techniques relevant to the detection (including bio-screening), extraction, separation, purification, identification and quantification of compounds in plant biochemistry, plant cellular and molecular biology, plant biotechnology, the food sciences, agriculture and horticulture. The Journal publishes papers describing significant novelty in the analysis of whole plants (including algae), plant cells, tissues and organs, plant-derived extracts and plant products (including those which have been partially or completely refined for use in the food, agrochemical, pharmaceutical and related industries). All forms of physical, chemical, biochemical, spectroscopic, radiometric, electrometric, chromatographic, metabolomic and chemometric investigations of plant products (monomeric species as well as polymeric molecules such as nucleic acids, proteins, lipids and carbohydrates) are included within the remit of the Journal. Papers dealing with novel methods relating to areas such as data handling/ data mining in plant sciences will also be welcomed.