摘要
目的 采用近红外光谱(near-infrared spectroscopy,NIRS)、中红外光谱(mid-infrared spectroscopy,MIRS)技术,实现对金振口服液(Jinzhen Oral Liquid,JOL)矿、植物药浸膏中黄芩苷、汉黄芩苷、甘草酸、没食子酸和固含量5个指标和人工牛黄浸膏中猪去氧胆酸、胆酸和固含量3个指标的含量预测,建立预测模型。方法 收集矿、植物药、人工牛黄2种浸膏样本,采集NIRS、MIRS并测定各指标含量。优选最佳光谱预处理方法和特征波段,融合光谱数据,采用偏最小二乘(partial least square,PLS)法建立8个指标的含量预测模型,并比较其性能,选出最优的含量预测模型。结果 NIRS技术对黄芩苷、汉黄芩苷、甘草酸、猪去氧胆酸、人工牛黄固含量的预测效果较好,预测相对偏差(relative standard error of prediction,RSEP)均低于8%,故选用NIRS模型作为这5个指标的最佳模型;没食子酸、胆酸和矿、植物药固含量的融合光谱数据模型预测效果较好,RSEP值均低于6%,故选用融合模型作为这3个指标的最佳模型。结论 基于NIRS和MIRS技术建立的模型,可用于JOL浸膏中8个指标的快速检测,方法操作简单、结果可靠。
Objective The near-infrared spectroscopy(NIRS)and mid-infrared spectroscopy(MIRS)technologies were used to predict the content of five indicators of baicalin,wogonoside,glycyrrhizic acid,gallic acid and solid content in Jinzhen Oral Liquid(JOL)mineral plant extract and three indicators of hyodeoxycholic acid,cholic acid and solid content in artificial bezoar extract,and a prediction model was established.Methods Collecting samples of mineral plants and artificial bezoar extracts,NIRS and MIRS were collected and the content of each index was determined.Optimizing the optimal spectral preprocessing method and feature bands,fusing spectral data,and using partial least squares(PLS)method to establish content prediction models for eight indicators,compare their performance and select the optimal content prediction model.Results NIRS technology has a better prediction effect on the content of baicalin,wogonoside,glycyrrhizic acid,hyodeoxycholic acid and artificial bezoar solid,and the relative standard error of prediction(RSEP)is less than 8%,so the NIRS model is selected as the best model for these five indicators;The fusion spectral data model of gallic acid,cholic acid,and mineral plant solid content has good prediction performance,with RSEP values all below 6%.Therefore,the fusion model is selected as the best model for these three indicators.Conclusion The model established based on NIRS and MIRS technology can be used for rapid detection of eight indicators in JOL extract,with simple operation and reliable results.
作者
李秀梅
徐芳芳
张欣
刘佳丽
张永超
樊成
王振中
LI Xiu-mei;XU Fang-fang;ZHANG Xin;LIU Jia-li;ZHANG Yong-chao;FAN cheng;WANG Zhen-zhong(Nanjing University of Chinese Medicine,Nanjing 210023,China;National Key Laboratory on Technologies for Chinese Medicine Pharmaceutical Process Control and Intelligent Manufacture,Lianyungang 222001,China;Jiangsu Kanion Pharmaceutical Co.,Ltd.,Lianyungang 222001,China;The Key Laboratory for the New Technique Research of TCM Extraction and Purification,Lianyungang 222047,China)
出处
《中草药》
CAS
CSCD
北大核心
2023年第24期8007-8017,共11页
Chinese Traditional and Herbal Drugs
基金
国家中医药管理局基于重点研究室研究领域的中医药多学科研究能力提升项目——中药提取精制技术(国中医药办科技函[2021]315号)
重点研发计划(社会发展):中药制造产生的废活性炭无害化处理及再生关键技术研究(SF2231)。
关键词
近红外光谱
中红外光谱
金振口服液
中间体含量
光谱融合
偏最小二乘法
黄芩苷
汉黄芩苷
甘草酸
没食子酸
固含量
人工牛黄
猪去氧胆酸
胆酸
预测相对偏差
near-infrared spectroscopy
mid-infrared spectroscopy
Jinzhen Oral Liquid
intermediate content
spectral fusion
partial least squares
baicalin
wogonoside
glycyrrhizic acid
gallic acid
solid content
artificial bezoar
hyodeoxycholic acid
cholic acid
relative standard error of prediction
