浏览全部资源
扫码关注微信
1.天津中医药大学 中药制药工程学院,天津 301617
2.天津市中药智能制药与绿色制药重点实验室,天津 301617
3.中药制药过程控制与智能制造技术全国重点实验室,江苏康缘药业股份有限公司,江苏 连云港 222001
4.现代中医药海河实验室,天津 301617
王海霞,博士,副研究员,研究方向:微生物快速检测,E-mail:whxtcm@tjutcm.edu.cn
纸质出版日期:2024-11-15,
收稿日期:2024-03-22,
修回日期:2024-05-11,
移动端阅览
赵堉文,李芷瑶,刘艺丹,李正,王海霞.基于机器学习的中药制剂中间体微生物限度快速预判研究[J].分析测试学报,2024,43(11):1725-1734.
ZHAO Yu-wen,LI Zhi-yao,LIU Yi-dan,LI Zheng,WANG Hai-xia.Research on Rapid Prediction of Microbial Limit of Intermediates in Traditional Chinese Medicine Formulations Based on Machine Learning[J].Journal of Instrumental Analysis,2024,43(11):1725-1734.
赵堉文,李芷瑶,刘艺丹,李正,王海霞.基于机器学习的中药制剂中间体微生物限度快速预判研究[J].分析测试学报,2024,43(11):1725-1734. DOI: 10.12452/j.fxcsxb.24032201.
ZHAO Yu-wen,LI Zhi-yao,LIU Yi-dan,LI Zheng,WANG Hai-xia.Research on Rapid Prediction of Microbial Limit of Intermediates in Traditional Chinese Medicine Formulations Based on Machine Learning[J].Journal of Instrumental Analysis,2024,43(11):1725-1734. DOI: 10.12452/j.fxcsxb.24032201.
通过比较随机森林(RF)、支持向量机(SVM)、主成分分析-支持向量机(PCA-SVM)与卷积神经网络(CNN)的模型性能,获得了基于表面增强拉曼光谱(SERS)技术的微生物限度快速预判最优模型,为中药制剂中间体的微生物限度快速预判提供了新方法。首先合成Au@Ag@SiO
2
复合纳米材料作为SERS增强基底,随后使用双层膜过滤法制备中药制剂中间体待测样本,并对样本抗菌活性进行考察。最后采集30批中药制剂中间体样本的SERS光谱,并分别建立RF、SVM、PCA-SVM与基于ResNet架构的CNN快速预判模型。结果表明,所建立的CNN模型的准确度、精确度、召回率均为100.0%,F1分数为1.0,受试者操作特征曲线(ROC)显示CNN模型对中药制剂中间体需氧菌总数(TAMC)、霉菌和酵母菌总数(TYMC)的快速预判能力均高于其他3种算法,能对待测样品微生物限度进行有效预判,对不合格样品进行有效风险预警,从而提高对中药生产过程中间体微生物的质量控制水平。
By comparing the model performance of random forest(RF),support vector machine(SVM),principal component analysis support vector machine(PCA-SVM),and convolutional neural network(CNN),the optimal model for rapid microbial limit prediction based on surface enhanced Raman spectroscopy(SERS) technology was obtained,providing a new method for rapid microbial limit prediction of intermediates in traditional chinese medicine(TCM) formulations. First,Au@Ag@SiO
2
composite nanomaterials were synthesized as SERS reinforcement substrates. Secondly,a double-layer membrane filtration method was used to prepare the intermediate test sample of TCM preparations,and the antibacterial activity of the sample was investigated. Finally,SERS spectra of 30 batches of intermediate samples from TCM preparations were collected,and RF,SVM,PCA-SVM,and CNN fast prediction models based on ResNet architecture were established,respectively. The results showed that the accuracy,precision,and recall of the established CNN model were all 100.0%,with an F1 score of 1.0. The receiver operating characteristic(ROC) curve showed that the CNN model had higher rapid prediction ability for the total aerobic bacterial count(TAMC),total mold and yeast count(TYMC) of TCM intermediates than the other three algorithms. It can effectively predict the microbial limit of the test sample,provide effective risk warning for
unqualified samples,and improve the quality control level of intermediate microorganisms in the production process of TCM.
表面增强拉曼散射(SERS)机器学习微生物限度检测快速预判
surface enhanced Raman scattering(SERS)machine learningmicrobial limit testingrapid prediction
Pharmacopoeia Committee. Pharmacopoeia of the People's Republic of China(2020). 1107 Microbial Limit Standards for Non Sterile Drugs. Beijing:China Medical Science and Technology Press(国家药典委员会. 中国药典. 1107 非无菌药品微生物限度标准. 北京:中国医药科技出版社),2020.
Cui Z,Ojaghian M R,Tao Z,Kakar K U,Zeng J,Zhao W,Duan Y,Vera Cruz C M,Li B,Zhu B,Xie G. J. Appl. Microbiol.,2016,120(5):1357-1367.
Marjan M,Akhtar H,Louis M J. TrAC,2018,107:60-77.
Zhu Y T,Li Z Y,Xie M M,Yan Y L,Zhang T,Wang H X. J. Instrum. Anal.(朱怡亭,李芷瑶,谢茂梅,颜月玲,张桐,王海霞. 分析测试学报),2024,43(1):138-146.
Gao W C,Li B,Yao R Z,Li Z P,Wang X W,Dong X L,Qu H,Li Q X,Li N,Chi H,Zhou B,Xia Z P. Anal. Chem.,2017,89(18):9836-9842.
Prakash O,Sil S,Verma T,Umapathy S. J. Phys. Chem. C,2019,124(1):861-869.
Cui L,Chen P,Chen S,Yuan Z,Yu C,Ren B,Zhang K. Anal. Chem.,2013,85(11):5436-5443.
Zhu X Y,Zhao Y W,Zhang Z S,Wang H X,Liu B S,Li Z,Wang M F. Microchim. Acta,2021,188:1-11.
Zhou S Y,Guo X J,Huang H Q,Huang X Q,Zhou X,Zhang Z B,Sun G D,Cai H H,Zhou H B,Sun P H. Anal. Chem.,2022,94(15):5785-5796.
Liu H B,Du X J,Zang Y X,Li P,Wang S. J. Agric. Food Chem.,2017,65:10290-10299.
Zhu X Y,Ning Y,Zhang Z,Wen Y,Zhao Y W,Wang H X. Anal. Bioanal. Chem.,2023,415(8):1529-1543.
Zhao Y W,Zhang Z S,Ning Y,Miao P Q,Li Z,Wang H X. Spectrochim. Acta A,2023,293:122510.
Wang Q,Zeng W D,Xia Z P,Li Z P,Qu H. Chin. J. Laser(王其,曾万聃,夏志平,李志萍,曲晗. 中国激光),2021,48(3):136-144.
Seju K,Inyoung K,Peter J V. Anal. Chem.,2021,93(27):9319-9328.
Cheng N T,Chen D J,Lou B,Fu J,Wang H Y. Biosens. Bioelectron.,2021,186:113246.
Cheng J L,Ma J,Xiao A H,Zhu Z L,Sang Z Y,Ma L. J. Beijing Fores.Univ.(程嘉莉,马江,肖爱华,朱仲龙,桑子阳,马履. 北京林业大学学报),2020,42(2):96-105.
Lee P,Meisel D. J. Phys. Chem. C,1982,86(17):3391-3395.
Xie S N,Girshick R,Dollar P,Tu Z W,He K M. IEEE CVPR,2017,7:1492-1500.
0
浏览量
3
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构