Quick Evaluation Of The Quality Of Mildew-prone Traditional Chinese Medicine Based On Electronic Nose

[Research Background]Traditional Chinese medicines and their decoction pieces are complex in composition and rich in nutrients.If the storage conditions are improper,they will produce varying degrees of deterioration,such as insects,mildew,oil and so on,which will reduce the quality and efficacy.Among them,mildew is one of the most prone deterioration phenomena.Some nutrient-rich Chinese medicines are prone to mildew at appropriate temperature and humidity,which can reduce or lose their efficacy.Even mycotoxins can be produced,so there is a saying that"mildew does not cure disease,or even kill people".However,some traditional Chinese medicines are often mildewed from inside to outside.It is difficult for the identification of such Chinese medicines to deteriorate and whether they can be used.Myristica fragrans Houtt.(so-calles Rou-Dou-Kou in Chinese),Da-Huang,Pang-Da-Hai are the representatives of this kind of traditional Chinese medicines.Due to the inability to detect such Chinese medicine,many adverse reactions have occurred in medicinal and edible foods.Rou-Dou-Kou is a medicinal and edible homologous variety prescribed by the Health Committee.Because it is rich in fatty oil and volatile oil,it is easy to produce "mildew"during storage.However,it is difficult to detect when nutmeg is mildewed,and the degree of mildew is very serious when the appearance produces visible mildew spots.At present,the main methods for judging the mildew of myristate are trait identification and physical and chemical determination.The former is the most commonly used,but both of them have some limitations:the particularity of Rou-Dou-Kou mildew makes it difficult to distinguish the characters,the boundary between mildew degree and medicinal use is difficult to distinguish,and the experiential requirement for human is high;the physical and chemical detection is too complicated,time-consuming and high cost.Moreover,the components of traditional Chinese medicine are complex,so it is inevitable that there are limitations in determining the quality of traditional Chinese medicine only by determining some indicators.In the early stage of mildew,Rou-Dou-Kou was identified as qualified,and the pharmacopoeia index was still within the standard range,but the odor of medicinal materials has changed slightly.In addition,mildew also produces some musty,rancid and other odors,which are difficult to perceive through the nose.Therefore,the electronic nose bionic olfactory system is introduced to analyze the odor produced during the moulding process of Rou-Dou-Kou in order to realize the early identification of moulding.[Purpose and significance]In this project,Rou-Dou-Kou was used as the carrier to explore the introduction of electronic nose technology in the identification of mildew in traditional Chinese medicine,and the changes of intrinsic and exogenous substances in the process of mildew in traditional Chinese medicine were comprehensively analyzed by character identification,physical and chemical analysis and electronic nose technology.From the overall perspective,explore the relationship between volatile and non-volatile substances in the process of mildew;understand the changes in the nature of mildew through the phenomenon,systematically evaluate the law of mildew of traditional Chinese medicine,establish a rapid and simple evaluation system,and realize the mildew of special Chinese medicine.Early judgment,quasi-judgment,easy explanation,and safety.[Contents and methods]1.Preparation of samples and establishment of sensory evaluation formThe samples of different degree of mildew were prepared based on the sample retentionmethod and the accelerated test method,and the sensory evaluation table was established by trait identification.2.Physical and chemical analysisAccording to the 2015 edition of the Chinese Pharmacopoeia and literature,the contentof aflatoxin,dehydroabietyl eugenol and chemical fingerprints in the samples were determined by HPLC.The volatile oil content was determined by steam distillation.The volatile oil and volatile components were determined by GC-MS.And the total number of aerobic bacteria,fungi and yeast was determined by microbial counting.Revealing the changes of intrinsic and exogenous substances in the process of mildew of Rou-Dou-Kou.3.Electronic nose analysisTaking the maximum response value of the electronic nose as the index,the optimal parameters of the determination of the Rou-Dou-Kou were determined by single factor investigation.The three discriminant analysis methods were used to establish the best discriminant model of different mildew degree of the Rou-Dou-Kou.A real-time monitoring system for the degree of mildew of Rou-Dou-Kou,was established by synthesizing sensory evaluation,physical and chemical analysis and electronic nose analysis results,and the early warning value of mildew was set up.[Result]1.Character identification resultsThe Rou-Dou-Kou samples with different degrees of mildew were prepared by the sample retention method and the accelerated test method.The sensory evaluation table was established based on the trait identification method.2.Physicochemical results(1)By analyzing the contents of dehydrodiisoeugenol and volatile oil in the Pharmacopoeia index,it was found that the content of dehydrodiisoeugenol decreased rapidly during the mildew of myristate,and the content of dehydrodiisoeugenol dropped below 0.1%on the third day of acceleration.And the contents of volatile oil in both the retention group and the acceleration group met the pharmacopoeia standard.(2)By analyzing the chemical composition fingerprint of the sample,it was found that the similarity between the fingerprints of the samples gradually decreased with the extension of the sample retention and acceleration time,indicating that the chemical composition of Rou-Dou-Kou samples changed greatly before and after moulding,no matter in observation or accelerated test.(3)Through GC-MS analysis of volatile oil and volatile odor of samples,and analysis of the difference products,it was found that the main difference products of volatile oil and volatile odor components were terpenes,whether retained or accelerated.The content of myristic ether and elemene in the volatile oil components of the sample and accelerated groups increased.(4)By measuring the total number of colonies and aflatoxin content,it was found that the total number of colonies in the sample group was small,and the aflatoxin B1 content did not exceed the lower limit of the Chinese version of the 2015 Chinese Pharmacopoeia.5μg·kg-1.The total number of colonies in the accelerated group showed an increasing trend and increased rapidly;the aflatoxin B,content exceeded the limit standard at 12 days of acceleration.3.Measurement results of electronic nose(1)The single-factor test method was used to investigate the electronic nose detection parameters,and the optimal conditions for the electronic nose were determined.Discriminant analysis and three machine learning methods were used to establish an electronic nose discrimination model for different degrees of mildew of Rou-Dou-Kou.The rationality of the electronic nose discriminant model is verified by physical and chemical analysis.(2)Correlation analysis between physical and chemical indicators and electronic nose sensor response showed that there was a significant correlation between the content of dehydroabietyl eugenol,the total number of colonies,aflatoxin B1 and the response of multiple electronic nose sensors.Furthermore,multivariate linear regression and BP neural network were used to establish these prediction models.It was found that BP neural network can predict component content more accurately than multiple linear regression,and the relative error is small.It indicates that the electronic nose has a certain predictive ability for aflatoxin and the total number of colonies.(3)A real-time monitoring system for mycotic degree of Rou-Dou-Kou was established based on physical and chemical analysis and electronic nose data.Good prediction results were obtained by using three pattern recognition algorithms.The prediction model was analyzed by Random Forest.The results showed that the model can effectively predict the different degree of mildew of Rou-Dou-Kou.Based on the electronic nose discriminant model,combined with the content of dehydrodiisoeugenol,aerobic bacteria,mold,yeast total number of colonies,AFB1 content established a pre-warning value of Rou-Dou-Kou.When the AFB1 content reaches 4.702ug·Kg-1,the dehydroisoisoeugenol content is less than 0.1%,the aerobic bacteria reach 4×106 cfu·g-1,and the mold and yeast colonies reach 5×106 cfu·g-1.It indicates that the medicinal materials have been slightly mildewed and cannot be used for medicinal purposes,providing a reference for clinical guidance for safe medication.Combining trait identification,physical and chemical analysis and electronic nose discriminant model provide the basis for clinical guidance medication.Based on previous studies and literature reports,it is suggested that the storage conditions of Rou-Dou-Kou and other mildew-prone medicinal materials should be less than 80%humidity.[Conclusion]The results of electronic nose discriminant model are consistent with those of physical and chemical analysis,which can make up for the shortcomings of various analytical methods.Based on the physical and chemical analysis,the change rule of substance basis in the process of Rou-Dou-Kou mildew was clarified.The established monitoring system for the degree of mildew can quickly and objectively distinguish the mildew of Rou-Dou-Kou from the medicinal material’s medicinal value and toxicity value,and has important reference significance for the establishment of early warning value.