CN104502307A - Method for quickly detecting content of glycogen and protein of crassostrea gigas - Google Patents

Abstract

The invention discloses a method for quickly detecting the content of glycogen and protein of crassostrea gigas. The method comprises the following steps: firstly, homogenizing a to-be-detected sample; then, collecting near infrared spectrum data; and finally comparing with a pre-established crassostrea gigas glycogen and protein content near infrared model to acquire a measurement value of the to-be-detected sample. The method has the characteristics of high detection speed, no chemical reagent, low experiment cost and no environmental pollution in the process of analyzing the content of glycogen and protein of crassostrea gigas. Establishment of the method for quickly detecting content of glycogen and protein of crassostrea gigas has important meaning to crassostrea gigas meat quality analysis, meat property breeding, breeding generation identification and genetic resource evaluation.

Description

The method of the long oyster glycogen of a kind of quick detection and protein content

Technical field

The present invention relates to a kind of method of shellfish nutritional labeling, specifically, relate to the method for the long oyster glycogen of a kind of quick detection and protein content.

Background technology

Long oyster ( crassostreagigas) also known as Pacific oyster, have environmental suitability strong, grow the advantages such as rapid, nutritious, be the large economic shellfish of blazoning property of the world that output is the highest.Within 2012, China's oyster total production reaches 3,950,000 tons, occupies first place in the world, for people provide abundant protein sources.

The local flavor that oyster is different and nutritional quality affect the consumption propensity of consumer, and then decide its commodity value.Along with improving constantly of living standard, people, for the consumption of oyster, more focus on local flavor and the nutrition of meat, and in oyster, glycogen content and protein content affect its local flavor and nutritional quality, therefore, the new varieties that seed selection mouthfeel is good, nutrition is good are active demands of oyster high-end market.But traditional glycogen and protein chemistry detection method have length consuming time, efficiency is low, testing cost is high shortcoming.Thus, setting up a kind of method of fireballing long oyster glycogen and protein content of checking is the basis of carrying out the seed selection of oyster Meat Quality new lines.

The present invention to long oyster glycogen and protein content analysis process, have inspection speed fast, without the need to using chemical reagent, experimental cost low and the feature of environmentally safe.The foundation of long oyster glycogen and rapid detection method for protein content, splitting length oyster meat quality parameters, Meat Quality seed selection, the qualification of breeding generation and germplasm resource evaluation has very important meaning.

Summary of the invention

The object of this invention is to provide the method for the long oyster glycogen of a kind of quick detection and protein content, the method has that inspection speed is fast, without the need to using chemical reagent, experimental cost low and the feature of environmentally safe, application prospect is extensive.

For solving the problem, the technical solution adopted in the present invention is:

The method of the long oyster glycogen of a kind of quick detection and protein content, first homogenized is carried out to testing sample, then near infrared spectrum data collection is carried out, finally compare with the long oyster glycogen set up in advance and protein content near-infrared model, thus obtain the measured value of testing sample, it is characterized in that: described long oyster glycogen and the foundation of protein content near-infrared model comprise the steps:

1. carry out anniversary sampling to long oyster main producing region, the long oyster sample number of acquisition all ages and classes, Various Seasonal is more than 50;

2. gather the near infrared spectrum of each sample, and sample is divided into modeling collection sample and checking collection sample;

3. the method using conventional chemical to detect detects the content of glycogen and protein in sample, obtains the chemical actual value of each sample;

4. by partial least-squares regression method, spectral information and chemical actual value information are carried out matching, set up glycogen and protein near-infrared analysis model, and the accuracy of testing model and predictive ability.

Compared with prior art, the present invention to long oyster glycogen and protein content analysis process, has that inspection speed is fast, without the need to using chemical reagent, experimental cost low and the feature of environmentally safe, its application prospect is extensive.The foundation of long oyster glycogen and rapid detection method for protein content, splitting length oyster meat quality parameters, Meat Quality seed selection, the qualification of breeding generation and germplasm resource evaluation has very important meaning.

Accompanying drawing explanation

Fig. 1 is the NIR diffuse reflection original spectrum that in embodiment, all samples organized by the fresh sample of long oyster.

Fig. 2 is the collection of illustrative plates in embodiment after the process of long oyster Xian Yang tissue glycogen content sample NIR light spectrum first derivation.

Fig. 3 A is the parameter index (degree of accuracy is high) of moisture NIR model in modeling process in embodiment.

Fig. 3 B is the parameter index (degree of accuracy is high) of glycogen content NIR model in modeling process in embodiment.

Fig. 3 C is the parameter index (degree of accuracy is high) of protein content NIR model in modeling process in embodiment.

Fig. 3 D is the parameter index (degree of accuracy is low) of content of taurine NIR model in modeling process in embodiment.

Embodiment

materials and methods

1.1 key instrument equipment

Fourier transform NIR light spectrometer (Antaris MX, USA), is equipped with RESULT ^tMthe integrated software of sample spectra acquisition and data processing software TQ analyst (Thermo Fisher, USA), microplate reader (BIO-RAD, USA), atomic fluorescence spectrometer (PA-10) and microwave digestion system (Mars Xpress, USA), high performance liquid chromatograph (Waters Inc., USA), flame atomic absorption spectrophotometer (AA-800, USA), Protein Analyzer (FOSS kjeltec ^tM2300, Sweden), apparatus,Soxhlet's, muffle furnace and large-scale vacuum freeze drier (ZDGX5), refiner (IKA ^?t18 basic ULTRA-TURRAX ^?, Germany).

sample collection

During year October in November, 2012 to 2013, respectively from Rushan, Shandong Province, Zhifu Island, Kongdong Island and Liugong Island, Liaoning Donggang City and hill island, 7 places of isometric oyster main producing region, Ganyu, Jiangsu, acquire 54 batches of long oyster samples that are wild and cultivation and amount to 94 parts, the fresh soft body meat of sample heavy (not comprising closed shell flesh) is 0.51g – 44.69g, sample contain 1 age shellfish, 2 age shellfish and 3 age shellfish.The season of male and female can not distinguished, maximum individuality and minimum individuality will be divided into be divided into two groups, as maximum value sample and minimal value sample with batch sample; In May, 2013-July, different according to male and female, will with batch sample be divided into male and female two increment this.

sample pre-treatments

Dissected by long oyster, get soft tissue, every increment originally gets 20-60g, is placed in the refrigeration centrifugation pipe of 50ml.First with scissors, soft tissue is shredded, then use refiner at maximum (top) speed homogenate 30-50s, in homogenization process, centrifuge tube is placed in ice chest.The sample that homogenate is good is divided into two parts: a part is used for NIR light spectrum and gathers, and another part is used for the chemical assay analysis of 8 kinds of component contents.

near infrared spectra collection

Before spectral scan, application RESULT integrated software compiles and edits the workflow that sample spectra gathers, and makes spectrometer start preheating at least 0.5 h.Highly good for 1.5cm homogenate sample is added in the quartz curette of diameter 1cm.Adopt diffuse reflection spectrum, spectral scan scope 10000 ﹣ 4000 cm ^-1, scanning times is 32 times, and resolution is 8 cm ^-1, during measurement, environment temperature is 20 DEG C, and relative humidity is 10 %.Gather the impact that background spectrum eliminates background before each collected specimens, Measuring Time is less than 1 min.

chemistry actual value measures

Centralab of oceanic resources and environment research institute of Shandong Province measuring the chemical actual value of 94 increments, 8 kinds of component contents originally.The mensuration of protein, total fat, zinc, selenium and ash content is carried out (GB 50095-2010, GB/T 14772-2008, GB/T 9695.20-2008, GB 500993-2010 and GB/T 9695.18-2008) according to industry standard; The mensuration of content of taurine is with reference to Chen Shenru (2013); Glycogen content measures and uses EnzyChrom ^tMglycogen kit (BioAssay Systems, USA).

model is set up and checking

Utilize TQ Analyst (version 9; USA) spectroscopic data of software process collection; select partial least square method (Partial Least Squares; PLC) as the stoechiometric process setting up calibration model; select the spectral range that software is recommended automatically; and model is optimized and checks, the preprocessing procedures that screening is best, to guarantee to obtain the optimal mathematical model of mathematics index.In this experiment, the modeling collection of 94 increment product and the sample number of checking collection are in table 1.Testing model adopts internal chiasma inspection, and namely each concentrating from modeling sample rejects 1 sample successively, predicts 1 disallowable sample, make all samples all disallowable and predicted with remaining sample Modling model.Mainly through comparison prediction value and chemical analysis value related coefficient ( r), validation-cross residual mean square (RMS) root (Root mean square error of cross-validation, rMSECV), prediction residual root mean square (Root mean square error of external prediction, rMSEP) and rPD(checking with the standard deviation (SD) of sample actual value with rMSECVor rMSEPratio, namely rPD _cV =SD/ rMSECVor rPD _eV =SD/ rMSEP) etc. index weigh the accuracy of calibration model.Due to reasons such as the systematic error of spectrometer, the drifts of spectral signal, may exception be there is in the NIR of measured sample, make model prediction accuracy decline (old snow English etc., 2009), this experiment adopts mahalanobis distance in TQ Analyst software to differentiate abnormity point, carries out the rejecting of sample exceptional value.

result

the descriptive statistic of 8 kinds of component content indexs organized by the fresh sample of 2.1 long oyster

This experiment chooses 94 increment product of different developmental phases, different breeding mode and all ages and classes in the multiple long oyster place of production of China, and the minimum sample number proposed more than Windham (1989) is the requirement of 50 parts.Table 1 is that 8 kinds of heterogeneity modeling collection and checking collect the number of sample and the analysis result of chemical actual value.The various component content scopes analyzed in experiment are larger; the ratio of content maxima and minima is respectively: moisture (1.23); glycogen (80.63); protein (3.20); total fat (33.74), taurine (3.28), zinc (5.55); selenium (4.97) and ash content (2.51), meet the requirement had a very wide distribution to sample size in near-infrared analysis model process of establishing.

The long oyster of table 1. fresh sample organization modeling collection and checking collection sample 8 kinds of component contents

Note: nthe quantity of sample; SD _cit is the standard deviation of modeling collection; SD _vit is the standard deviation of checking collection.

spectroscopic data pre-service

The NIR diffuse reflection original spectrum of the fresh sample tissue samples of long oyster as shown in Figure 1.Carry out different smoothing processing to the spectroscopic data of heterogeneity in modeling process, the final optimal parameter being applicable to respective Component Model of selecting combines.According to the optimal parameter combination filtered out, determine that the fresh sample of long oyster organizes moisture, glycogen, protein, always fat, taurine, zinc, selenium, the forecast model of ash content.The major parameters such as the main cause subnumber of the spectral range of each component content model, spectral manipulation method and modeling are in table 2.

model is set up and is optimized

According to the recommendation of TQ Analyst software, by heterogeneity content sample abnormality value removing.In the result of each component content modeling of sample, the related coefficient of moisture, glycogen and protein content ( r _c ) higher, be all greater than 0.96, its rMSECbe worth all less (Fig. 3. A, B, C); Cross validation related coefficient ( r _cV ) and external certificate related coefficient ( r _eV ) higher, be all greater than 0.93, its rMSECVvalue and rMSEPbe worth also less; The important parameter of modelling verification rPDvalue variation range is 2.80-7.04, illustrates that the model accuracy of moisture, glycogen and protein content is high, can be used for the ingredient prediction of the fresh sample tissue of long oyster.On the other hand, 5 compositions such as total fat, taurine, zinc, selenium, ash content, in modeling process r _c value and rMSECthe parameters such as value all undesirable (Fig. 3. D); The related coefficient of cross validation and external certificate ( r _cV with r _eV ) all lower than 0.85, rPDbe worth except Zn content all lower than 2.5.Consider multiple measurement index, the grade NIR model of 5 component contents of total fat, taurine, zinc, selenium, ash is not suitable for the accurate quantitative analyses of long oyster fresh meat tissue.

Table 2. long oyster component content modeling collection and checking light harvesting Spectrum data processing parameter

^afirst order derivative

^bnorris derivative filtrator

^cwave band is long

^dgap between wave band

^efiltrator

^fdata point

^gpolynomial expression power

^hsecond derivative

3 discuss

3.1 fresh sample patterns compare with dry sample pattern

This experiment is being set up while the fresh sample of long oyster organizes NIR model, and establish the NIR model of long oyster freeze drying sample, the modeling effect of fresh sample tissue and powdered sample two kinds of different disposal forms there are differences.Wherein, powdered sample model result shows; glycogen and protein content model can predict the content of unknown sample accurately; total fat, zinc, selenium, ash content can predict the content of unknown sample accurately, only have the prediction effect of content of taurine model bad (not delivering data).In the fresh sample tissue of long oyster and powdered sample two experimental results, glycogen is consistent with protein content modeling effect, all can predict unknown sample accurately; Content of taurine model all can not accurately predict unknown sample in two experiments; And the model of the fat of fresh sample tissue, zinc, selenium and ash content can not Accurate Prediction.Viljoen et al. (2005,2007) finds in ostrich meat and the fresh sample tissue of mutton and freeze drying example research process, and the prediction effect of freeze drying example model is better.When carrying out spectral analysis to freeze drying example, sample temperature change is not obvious; And freeze drying example can avoid very high absorption peak (the Murray & Williams in IR regions, 1987), these noises may reduce the accuracy (Pedersen et al. 2003) of model prediction.Due to moisture high (about 80%) in long oyster fresh sample tissue, the uncertainty of temperature variation in derivative spectomstry gatherer process, this may be cause the unsuccessful main cause of modeling.

major parameter standard in model process of establishing

In the model built, modeling related coefficient ( r _c ), validation-cross related coefficient ( r _c ) and external certificate related coefficient ( r _c ) prediction effect of value more close to 1 model be better, in this experiment, each related coefficient is all more than 0.93, and mxm. is 0.99, illustrates that the degree of accuracy of model is very high.Simultaneously model rPDwhether accurately value weighs model another important indicator, and good model has high rPDvalue.When rPDwhen value is greater than 2.5, model can carry out Accurate Prediction (Zhou et al., 2012; Guy et al. 2011).Although the RPD value of Zn content forecast model is more than 10 in this experiment, no matter be it r _cvalue, r _cVvalue still r _eVvalue is all less than 0.52, and consider each and every one index each, the NIR model of Zn content does not possess the ability of Accurate Prediction.

Taurine is a kind of free amino acid, belongs to organism, in theory by setting up NIR model, can predict the content of unknown sample accurately.But content of taurine forecast result of model is poor in this experiment, analyze the SD value little (Zhou et al., 2012) that reason may be taurine chemistry actual value, caused by mobility scale narrow (Prieto et al., 2009).

the meaning to Meat Quality seed selection set up by model

This experimental result will be used for the research such as qualification and germplasm resource evaluation of the seed selection of long oyster Meat Quality, breeding generation.The research of these aspects needs the analysis several thousand individualities, multiple index being carried out to component content, therefore conventional chemical analysis method is difficult to fast, efficiently in batches complete these analytical works.Seminar carries out component content analysis by using this experimental result to the long oyster family built, and carrying out genetic parameter estimation and breeding value estimation, laying the foundation for carrying out Meat Quality seed selection.

conclusion

NIR technology can measure moisture, glycogen and the protein 3 kinds of component contents in long oyster fresh meat tissue quickly and accurately, but can not Measurement accuracy to total fat, zinc, selenium, taurine and ash content.The above 3 kinds of component content NIR models set up in this research, can have very important meaning to the qualification of long oyster meat quality parameters, Meat Quality seed selection and breeding generation and germplasm resource evaluation fast, accurately, non-environmental-pollution.

Claims (1)

1. one kind is detected the method for long oyster glycogen and protein content fast, first homogenized is carried out to testing sample, then near infrared spectrum data collection is carried out, finally compare with the long oyster glycogen set up in advance and protein content near-infrared model, thus obtain the measured value of testing sample, it is characterized in that: described long oyster glycogen and the foundation of protein content near-infrared model comprise the steps:

1. carry out anniversary sampling to long oyster main producing region, obtain the long oyster sample of all ages and classes, Various Seasonal, sample number is more than 50 parts;

2. gather the near infrared spectrum of each sample, and sample is divided into modeling collection sample and checking collection sample;

3. the method using conventional chemical to detect detects the content of glycogen and protein in sample, obtains the chemical actual value of each sample;

4. by partial least-squares regression method, spectral information and chemical actual value information are carried out matching, set up glycogen and protein near-infrared analysis model, and the accuracy of testing model and predictive ability.