Herbal medicines are popular natural medicines that have been used for decades.The use of alternative medicines continues to expand rapidly across the world.The World Health Organization suggests that quality assessme...Herbal medicines are popular natural medicines that have been used for decades.The use of alternative medicines continues to expand rapidly across the world.The World Health Organization suggests that quality assessment of natural medicines is essential for any therapeutic or health care applications,as their therapeutic potential varies between different geographic origins,plant species,and varieties.Classification of herbal medicines based on a limited number of secondary metabolites is not an ideal approach.Their quality should be considered based on a complete metabolic profile,as their pharmacological activity is not due to a few specific secondary metabolites but rather a larger group of bioactive compounds.A holistic and integrative approach using rapid and nondestructive analytical strategies for the screening of herbal medicines is required for robust characterization.In this study,a rapid and effective quality assessment system for geographical traceability,species,and variety-specific authenticity of the widely used natural medicines turmeric,Ocimum,and Withania somnifera was investigated using Fourier transform near-infrared(FT-NIR)spectroscopy-based metabolic fingerprinting.Four different geographical origins of turmeric,five different Ocimum species,and three different varieties of roots and leaves of Withania somnifera were studied with the aid of machine learning approaches.Extremely good discrimination(R^(2)>0.98,Q^(2)>0.97,and accuracy=1.0)with sensitivity and specificity of 100%was achieved using this metabolic fingerprinting strategy.Our study demonstrated that FT-NIR-based rapid metabolic fingerprinting can be used as a robust analytical method to authenticate several important medicinal herbs.展开更多
Understanding the evolution of physicochemical properties during solid-state fermentation is essential for quality control in complex fermented foods such as Chinese Baijiu.In this study,we propose an NIR-based chemom...Understanding the evolution of physicochemical properties during solid-state fermentation is essential for quality control in complex fermented foods such as Chinese Baijiu.In this study,we propose an NIR-based chemometric framework that explicitly incorporates fermentation-round information to enhance the monitoring of physicochemical changes in the stacking fermentation of sauce-flavor Baijiu.A total of 671 fermented grain(Zaopei)samples were analyzed across seven fermentation rounds.Variable selection strategies—including Competitive Adaptive Reweighted Sampling(CARS),Genetic Algorithm(GA),and Random Forest-based Selection(RFS)—were combined with nonlinear models such as Support Vector Machine(SVM)and eXtreme Gradient Boosting(XGBoost)to construct classification and regression models.The GA+XGBoost model achieved a classification accuracy of 99.26%for fermentation round identification.The RFS+SVR(Support Vector Regression)model accurately predicted acidity(R^(2)=0.9674)and starch(R^(2)=0.9610).Importantly,incorporating fermentation-round information as a categorical covariate improved model generalization and interpretability,highlighting its methodological importance for managing stage-dependent variability in solid-state systems.By combining fermentation-round covariates with interpretable spectral features,the modeling framework demonstrated enhanced robustness and applicability.Overall,this study demonstrates the potential of NIR-based chemometrics as a rapid,non-destructive approach for fermentation stage monitoring and quality control in traditional Baijiu production and other complex fermented foods.展开更多
As traditional Chinese medicines,Fritillaria from different origins are very similar and it is difficult to distinguish them.In this study,the laser-induced breakdown spectroscopy combined with learning vector quantiz...As traditional Chinese medicines,Fritillaria from different origins are very similar and it is difficult to distinguish them.In this study,the laser-induced breakdown spectroscopy combined with learning vector quantization(LIBS-LVQ)was proposed to distinguish the powdered samples of Fritillaria cirrhosa and non-Fritillaria cirrhosa.We also studied the performance of linear discriminant analysis,and support vector machine on the same data set.Among these three classifiers,LVQ had the highest correct classification rate of 99.17%.The experimental results demonstrated that the LIBS-LVQ model could be used to differentiate the powdered samples of Fritillaria cirrhosa and non-Fritillaria cirrhosa.展开更多
基金Department of Science and Technology-SERB-SRG research grant(Grant No.:SRG/2021/000750-G)and Department of Biotechnology for Ramalingaswami grant(Grant No.:BT/RLF/Re-entry/21/2020)Director,Prabodh Kumar Trivedi,of CSIR-CIMAP for providing infrastructure,facility,and funding support from CSIR,India(Grant Nos.:FC2020-23/NMITLI/TLP0001&TLP0002)We acknowledge Dr.Ritu Trivedi(CSIR-CDRI Lucknow,India)for support and Dr.Abolie Girme and Dr.Lal Hingorani(Pharmanza herbal Pvt.Ltd,India)for providing Withania somnifera samples.We acknowledge Dr.Neerja Tiwari for FT-NIR access,Ms.Manju Yadav and Ms.Namita Gupta for HPLC access,and Ms.Anju Yadav for GC-MS access.Authors would like to thank Aroma mission HCP-0007,India for funding support.Prof.Christopher T.Elliott would like to thank Bualuang ASEAN Chair Professor Fund,UK and Queen's University Belfast Fund,UK.
文摘Herbal medicines are popular natural medicines that have been used for decades.The use of alternative medicines continues to expand rapidly across the world.The World Health Organization suggests that quality assessment of natural medicines is essential for any therapeutic or health care applications,as their therapeutic potential varies between different geographic origins,plant species,and varieties.Classification of herbal medicines based on a limited number of secondary metabolites is not an ideal approach.Their quality should be considered based on a complete metabolic profile,as their pharmacological activity is not due to a few specific secondary metabolites but rather a larger group of bioactive compounds.A holistic and integrative approach using rapid and nondestructive analytical strategies for the screening of herbal medicines is required for robust characterization.In this study,a rapid and effective quality assessment system for geographical traceability,species,and variety-specific authenticity of the widely used natural medicines turmeric,Ocimum,and Withania somnifera was investigated using Fourier transform near-infrared(FT-NIR)spectroscopy-based metabolic fingerprinting.Four different geographical origins of turmeric,five different Ocimum species,and three different varieties of roots and leaves of Withania somnifera were studied with the aid of machine learning approaches.Extremely good discrimination(R^(2)>0.98,Q^(2)>0.97,and accuracy=1.0)with sensitivity and specificity of 100%was achieved using this metabolic fingerprinting strategy.Our study demonstrated that FT-NIR-based rapid metabolic fingerprinting can be used as a robust analytical method to authenticate several important medicinal herbs.
基金supported by the National Key Research and Devel-opment Program of China(2023YFE0105500)the National Natural Science Foundation of China(32272407,32372465)+2 种基金the China Post-doctoral Science Foundation(2020M683372)the Cooperation Project of Luzhou Laojiao Co.,Ltd.Jiangsu University,China.
文摘Understanding the evolution of physicochemical properties during solid-state fermentation is essential for quality control in complex fermented foods such as Chinese Baijiu.In this study,we propose an NIR-based chemometric framework that explicitly incorporates fermentation-round information to enhance the monitoring of physicochemical changes in the stacking fermentation of sauce-flavor Baijiu.A total of 671 fermented grain(Zaopei)samples were analyzed across seven fermentation rounds.Variable selection strategies—including Competitive Adaptive Reweighted Sampling(CARS),Genetic Algorithm(GA),and Random Forest-based Selection(RFS)—were combined with nonlinear models such as Support Vector Machine(SVM)and eXtreme Gradient Boosting(XGBoost)to construct classification and regression models.The GA+XGBoost model achieved a classification accuracy of 99.26%for fermentation round identification.The RFS+SVR(Support Vector Regression)model accurately predicted acidity(R^(2)=0.9674)and starch(R^(2)=0.9610).Importantly,incorporating fermentation-round information as a categorical covariate improved model generalization and interpretability,highlighting its methodological importance for managing stage-dependent variability in solid-state systems.By combining fermentation-round covariates with interpretable spectral features,the modeling framework demonstrated enhanced robustness and applicability.Overall,this study demonstrates the potential of NIR-based chemometrics as a rapid,non-destructive approach for fermentation stage monitoring and quality control in traditional Baijiu production and other complex fermented foods.
基金supported by National Natural Science Foundation of China(No.62075011)Graduate Technological Innovation Project of Beijing Institute of Technology(No.2019CX20026)。
文摘As traditional Chinese medicines,Fritillaria from different origins are very similar and it is difficult to distinguish them.In this study,the laser-induced breakdown spectroscopy combined with learning vector quantization(LIBS-LVQ)was proposed to distinguish the powdered samples of Fritillaria cirrhosa and non-Fritillaria cirrhosa.We also studied the performance of linear discriminant analysis,and support vector machine on the same data set.Among these three classifiers,LVQ had the highest correct classification rate of 99.17%.The experimental results demonstrated that the LIBS-LVQ model could be used to differentiate the powdered samples of Fritillaria cirrhosa and non-Fritillaria cirrhosa.
