A gc-ims-based rapid evaluation method for the sensory digitization of pure milk flavor

Abstract

The application discloses a kind of based on GC-IMS pure milk flavor sensory digitalization fast evaluation method, belong to food analysis and evaluation technical field.The pure milk sample to be measured is detected by headspace sampling using GC-IMS, and a fingerprint spectrum is obtained.The fingerprint spectrum data or characteristic volatile component data is input into the OPLS-DA model, and the preference degree of pure milk is evaluated according to the clustering result.The construction method of OPLS-DA model is as follows: different pure milk samples are selected, and CATA sensory evaluation method is used for sensory evaluation of pure milk samples.Different pure milk samples are detected by headspace sampling using GC-IMS, and a fingerprint spectrum is obtained.Partial least squares-discriminant analysis method is used for regression modeling using fingerprint spectrum data or characteristic volatile component data to obtain OPLS-DA model.The application uses GC-IMS to digitally and quickly evaluate the flavor of pure milk, which has high sensitivity, does not require sample enrichment and concentration, saves detection time, and avoids tedious pretreatment process.

Description

Technical Field

[0001] This invention belongs to the field of food analysis and evaluation technology, specifically relating to a rapid digital evaluation method for the sensory flavor of pure milk based on GC-IMS. Background Technology

[0002] As early as the 19th century, scholars began studying the influence of measurement psychology on human perception and decision-making. In the 1940s, systematic methods such as the nine-point preference scale analysis were used to collect people's acceptance of food and to use this as a basis for food selection. This is considered the earliest application of food sensory evaluation technology. In the 1960s, with the rapid development of the food industry, sensory analysis technology gained attention and developed rapidly. In 1965, Amerine et al. proposed that sensory analysis is a method of sensory judgment based on sensory physiology and cognitive psychology, through reasonable statistical design and data analysis, which is also the prototype of the sensory analysis concept. Currently, sensory analysis technology is generally considered a scientific method for evoking, measuring, analyzing, and interpreting the responses of products through vision, smell, touch, taste, and hearing. Sensory science is relatively mature in developed countries, with many companies establishing dedicated sensory evaluation departments and many companies providing product sensory survey services. In China, research on sensory analysis began in the 1980s, but sensory evaluation was gradually applied to the food industry only in the 1990s. Sensory analysis technology, as an effective tool for evaluating product quality, plays a practical role in all aspects of product manufacturing and has a function that chemical instruments cannot replace. The development of sensory analysis technology in my country started relatively late, and most existing methods are based on foreign standards, lacking a complete system and lagging behind international research. To date, research on sensory analysis technology for dairy products has been developing for nearly a century abroad, while in China it is mainly applied to yogurt and cheese products, with less application and depth in other dairy products. Dairy companies need to learn from international and domestic applications in other products and innovate sensory analysis methods to provide a basis for the research and development and production of dairy products.

[0003] In the 1990s, the concept of digital sensory systems simulating human senses was proposed. This was primarily based on simulating the human sensory perception process. Sensors generate corresponding signals based on the properties of the sample being tested, acting as analogous to human sensory organs. Signal acquisition devices transmit and process these signals, and then computers analyze and identify the data to make judgments. With the development of intelligent systems, more and more digital intelligent sensory instruments have been developed. Currently applied intelligent sensory analysis technologies include electronic noses, electronic tongues, machine vision, taste meters, and texture analysis. Compared to time-consuming and methodologically demanding manual sensory analysis methods, digital intelligent sensory analysis offers advantages such as objectivity, stability, and speed, making it a current research hotspot and development trend.

[0004] Therefore, there is currently limited research on artificial sensory evaluation methods for pure milk, and intelligent digital sensory evaluation methods based on analytical instruments are also immature. There is an urgent need for a sensitive, reliable, and rapid digital evaluation method to conduct objective digital sensory evaluation and even prediction of pure milk. This could help dairy companies improve their products, understand market and consumer preferences, and enhance product quality. It would also provide researchers with methods to conduct in-depth studies on the relationship between the characteristic flavors and sensory properties of dairy products. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a rapid digital evaluation method for the flavor sensory characteristics of pure milk based on GC-IMS. GC-IMS is characterized by high sensitivity, does not require any pretreatment processes such as enrichment and concentration, and directly injects samples into the headspace. It is fast, efficient, simple to operate, and preserves the true flavor composition of the samples, making it suitable for rapid screening of batch samples.

[0006] This invention is achieved through the following technical solution:

[0007] This invention provides a rapid digital evaluation method for the sensory flavor of pure milk based on GC-IMS. The pure milk sample to be tested is detected by gas chromatography-ion mobility spectrometry via headspace injection to obtain a fingerprint spectrum. The fingerprint spectrum data or the characteristic volatile component data in the fingerprint spectrum are input into the OPLS-DA model, and the preference for pure milk is evaluated based on the clustering results.

[0008] The method for constructing the OPLS-DA model is as follows:

[0009] (1) Select different pure milk samples and use the CATA sensory evaluation method to evaluate the pure milk samples;

[0010] (2) Different pure milk samples were detected by headspace injection using gas chromatography-ion mobility spectrometry to obtain fingerprint spectra. Partial least squares-discriminant analysis was used to perform regression modeling based on fingerprint spectra data or characteristic volatile component data in fingerprint spectra to construct the OPLS-DA model.

[0011] Furthermore, before gas chromatography-ion mobility spectrometry (GC-IMS) detection, the pretreatment method for pure milk samples is as follows: heating at 60℃ for 15 min followed by headspace injection.

[0012] Furthermore, the detection conditions of the gas chromatography-ion mobility spectrometer are as follows: column: MXT-wax, 15m, ID: 0.53mm, df: 1.0um; column temperature: 60℃; carrier gas / drift gas: N2; IMS temperature: 45℃; injection volume in the automatic headspace sampling unit: 1000uL; incubation time: 15min; incubation temperature: 60℃; injection needle temperature: 85℃; incubation speed: 500rpm; drift gas flow rate: 75mL / min; carrier gas flow rate: 0~2min 2mL / min, 2~10min 2~10mL / min, 10~20min 10~100mL / min, 20~30min 100mL / min.

[0013] Furthermore, the characteristic volatile components are nonanone, octanone, heptanone, pentanone, 2,3-butanedione, acetaldehyde, and 2-pentylfuran.

[0014] Furthermore, the contents of nonanone, octanone, heptanone, pentanone, 2,3-butanedione, acetaldehyde, and 2-pentylfuran in pure milk samples with high sensory evaluation scores were significantly different from those in pure milk samples with low sensory evaluation scores.

[0015] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0016] (1) This invention uses gas chromatography-ion mobility spectrometry (GC-IMS) to digitally and rapidly evaluate the flavor and sensory characteristics of pure milk. It has extremely high sensitivity and does not require enrichment and concentration of pure milk samples, which can save sample detection time, avoid cumbersome pretreatment process, and retain the true flavor composition of pure milk to the greatest extent. It is currently the instrumental analysis method that is closest to the sensory characteristics and can be better combined with the sensory evaluation results.

[0017] (2) This invention utilizes the CATA sensory evaluation method, which reduces the reliance on professional sensory personnel and obtains better sensory results of pure milk from the consumer's perspective. Combined with GC-IMS, relevant VOC components are screened and an OPLS-DA model is established. The method has high reliability and provides a fast and objective evaluation method. Attached Figure Description

[0018] Figure 1 Here are the GC-IMS fingerprints of different pure milk samples from Example 1;

[0019] Figure 2 The OPLS-DA model constructed from the fingerprint map of Example 1;

[0020] Figure 3 The fingerprint spectra of volatile components characteristic of different pure milk samples in Example 1;

[0021] Figure 4 The OPLS-DA model constructed for the fingerprint spectrum of characteristic volatile components in Example 1;

[0022] Figure 5 This is a Y-prediction value graph for 9 pure milk samples in Example 2. Detailed Implementation

[0023] The present invention is further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Experimental methods in the following embodiments, unless otherwise specified, are generally performed under conventional conditions or as recommended by the manufacturer.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of skill in the art. All reagents and materials used in this invention are readily available through conventional means, and unless otherwise specified, they shall be used in accordance with conventional methods in the art or as per the product instructions.

[0025] Example 1

[0026] Construction of the OPLS-DA model:

[0027] (1) Take 10 samples of pure milk from different brands and number them 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 respectively. Use the CATA sensory evaluation method to evaluate the pure milk samples. The specific operation is as follows:

[0028] A CATA sensory evaluation form was designed (as shown in Table 1), and 15 consumers were recruited to complete the sensory evaluation. The 15 consumers were all between 30 and 40 years old, 7 women and 8 men, and each of them drank pure milk more than 4 times a week on average. Before the sensory evaluation began, they were given standardized training and introduction.

[0029] Table 1 CATA Sensory Evaluation Table

[0030]

[0031] After all participants completed the CATA sensory evaluation, the results were compiled and summarized. First, each person's score was individually digitized to obtain a new data table. Then, the scores for the same product were tallied, the highest and lowest scores were removed to eliminate outliers, and the average was taken to obtain the final numerical matrix of scores from 15 people for 10 products, as shown in Table 2.

[0032] Table 2. Numerical matrix of final scores for different pure milk samples.

[0033]

[0034] Table 2 shows that the preference scores for pure milk samples 1-6 are greater than 0, especially samples 1, 5, and 6, which have preference scores greater than 1, indicating they are the more popular products. Samples 7, 8, 9, and 10 have preference scores less than 0, indicating lower preference for these products. This is related to other sensory characteristics of pure milk; for example, samples 1, 5, and 6 have a strong overall aroma, rich milky flavor and aftertaste, and low levels of off-flavors and fishy smells, making them more popular. Conversely, samples 7-10 have lower sensory scores for aroma and aftertaste, and stronger off-flavors and fishy smells, resulting in lower preference scores. This illustrates the correlation between aroma intensity, milky flavor, aftertaste, milky smell, and other off-flavors and preference levels, and also demonstrates the objectivity and reliability of these sensory evaluations.

[0035] (2) The pure milk samples (1~10) of the above different brands were detected by headspace injection using gas chromatography-ion mobility spectrometry to obtain fingerprint spectra. Partial least squares-discriminant analysis was used to perform regression modeling based on the fingerprint spectra data (peak height) to construct an OPLS-DA model to study the correlation between volatile flavor components in milk and consumers' preference for milk. The specific method is as follows:

[0036] Take 5 mL of each of the 10 pure milk samples from step (1) above and transfer them into 20 mL headspace vials. Add 0.2 mg / kg of 2-octanol to each sample as an internal standard for quantification (adding 2-octanol is only for later quantification; it can be omitted during modeling). After heating at 60°C for 15 min, inject 1 mL of headspace sample (each sample is injected 3 times, numbered as follows: 1 (011, 012, 013), 2 (014, 015, 016, 017, 018, 019, 010, 011, 012, 013, ... 21, 022, 023), 2 (021, 022, 023), 3 (031, 032, 033), 4 (041, 042, 043), 5 (051, 052, 053), 6 (061, 062, 063), 7 (071, 072, 073), 8 (081, 082, 083), 9 (091, 092, 093), 10 (101, 102, 103);

[0037] The detection conditions for the gas chromatography-ion mobility spectrometer were as follows: column: MXT-wax, 15m, ID: 0.53mm, df: 1.0um; column temperature: 60℃; carrier gas / drift gas: N2; IMS temperature: 45℃; injection volume in the automatic headspace sampling unit: 1000uL; incubation time: 15min; incubation temperature: 60℃; injection needle temperature: 85℃; incubation speed: 500rpm; gas chromatography conditions: E1 drift gas flow rate: 75mL / min; E2 carrier gas flow rate: 0~2min 2mL / min, 2~10min 2~10mL / min, 10~20min 10~100mL / min, 20~30min 100mL / min.

[0038] Fingerprint patterns of different pure milk samples, as follows Figure 1 As shown in Table 3, the compounds detected by fingerprint analysis were analyzed, and the results are as follows:

[0039] Table 3 Summary of compounds detected by gas chromatography-ion mobility spectrometry

[0040]

[0041] according to Figure 1 The peak height data of the fingerprint spectrum shown were used to model OPLS-DA on 10 pure milk samples, and the resulting OPLS-DA model diagram is shown below. Figure 2 As shown, model R 2 X=0.76, Q 2 =0.98, the accuracy of 7-fold cross-validation is 100%, AUC=1.00. (From...) Figure 2 It can be seen that pure milk with higher and lower preference scores are distributed on both sides of the OPLS-DA model diagram. Pure milk with higher preference scores is clustered on the left (green dots), while pure milk with lower preference scores is clustered on the right (blue dots). This indicates that pure milk with different preference levels has differences in the composition of aroma components.

[0042] Sensory preference scores were combined with multivariate statistical results of VOC content. Supervised learning (OPLS-DA) and unsupervised learning (PCA) methods were used to study compounds associated with sensory scores. Through OPLS-DA VIP and PCA loading, seven key aroma compounds were identified: nonanone, octanone, heptanone, pentanone, 2,3-butanedione, acetaldehyde, and 2-pentylfuran. The content of seven key aroma compounds was higher in milk samples 7-10 than in samples 1-6. The average contents of nonanone, caprylicone, heptanone, pentanone, 2,3-butanedione, acetaldehyde, and 2-pentylfuran in milk samples 7-10 were 0.18, 0.04, 0.46, 0.48, 0.02, 0.08, and 0.006 mg / kg, respectively; while the average contents in samples 1-6 were 0.13, 0.02, 0.43, 0.44, 0.01, 0.07, and 0.003 mg / kg, respectively. These seven key aroma compounds were used as characteristic volatile components. The fingerprint spectrum of the characteristic volatile components is shown below. Figure 3 As shown, OPLS-DA modeling is performed based on the peak height of the characteristic volatile components, and the resulting OPLS-DA model diagram is shown below. Figure 4 As shown; Model R 2 X=0.96, Q 2 =0.82, the accuracy of 7-fold cross-validation is 100%, AUC=1.00. (From...) Figure 3 and Figure 4 It can be seen that the content of the seven characteristic volatile components differs between the two types of samples (high preference and low preference), and this difference can be used to distinguish and predict the preference for pure milk through statistical models.

[0043] Example 2

[0044] OPLS-DA model validation with 7 markers

[0045] Nine different brands of UHT pure milk (numbered n1~n9) were purchased, and an OPLS-DA model was used to analyze seven characteristic volatile components. Figure 4 Verification will be performed.

[0046] The method described in Example 1 above was used to perform gas chromatography-ion mobility spectrometry (GC-IMS) on pure milk samples numbered n1 to n9 (no 2-octanol was added during sample pretreatment). The peak heights of the seven characteristic volatile components were then included. Figure 4 The model shown, the regression prediction value Y result graph is as follows: Figure 5 As shown (a positive regression predicted value Y indicates a high degree of preference; a negative regression predicted value Y indicates a low degree of preference), from Figure 5It can be seen that n1, n2, n3, n4, n5, n6, and n8 are pure milk samples with high preference, while n7 and n9 are pure milk samples with low preference.

[0047] Ten volunteers were invited to conduct CATA sensory evaluations of nine different brands of UHT milk (numbered n1 to n9). First, the scores given by each volunteer for the nine milk samples were centered, removing the highest and lowest scores. Then, the mean score for each milk sample was calculated. The score table for different milk preferences is shown in Table 4. Table 4 shows that n1, n2, n3, n4, n5, n6, and n8 had high preference levels (scores greater than 0), while n7 and n9 had low preference levels (scores less than 0). Substituting this into the above... Figure 5 The consistent prediction results further prove that... Figure 4 The accuracy of the model shown.

[0048] Table 4. Preference Scores for Different Brands of Pure Milk

[0049]

[0050] The results show that, by combining sensory evaluation scores, the OPLS-DA model established using seven characteristic volatile components can predict consumers' preference for whole milk. These characteristic volatile components (markers) are closely related to the lipid content and changes in whole milk, indicating that lipids in whole milk have a significant impact on sensory perception and need to be studied and controlled during product development and production.

Claims

1. A rapid digital evaluation method for the sensory flavor of pure milk based on GC-IMS, characterized in that, The pure milk sample to be tested was detected by headspace injection using gas chromatography-ion mobility spectrometry to obtain a fingerprint spectrum. The fingerprint spectrum data or the characteristic volatile component data in the fingerprint spectrum were input into the OPLS-DA model, and the preference for pure milk was evaluated based on the clustering results. The method for constructing the OPLS-DA model is as follows: (1) Select different pure milk samples and use the CATA sensory evaluation method to evaluate the pure milk samples; (2) Different pure milk samples were detected by headspace injection using gas chromatography-ion mobility spectrometry to obtain fingerprint spectra. Partial least squares-discriminant analysis was used to perform regression modeling based on fingerprint spectra data or characteristic volatile component data in fingerprint spectra to construct the OPLS-DA model. Before gas chromatography-ion mobility spectrometry (GC-IMS) analysis, the pretreatment method for pure milk samples is as follows: heating at 60℃ for 15 min followed by headspace injection. The detection conditions of the gas chromatography-ion mobility spectrometer are as follows: column: MXT-wax, 15m, ID: 0.53mm, df: 1.0um; column temperature: 60℃; carrier gas / drift gas: N2; IMS temperature: 45℃; in the automatic headspace sampling unit, injection volume: 1000uL; incubation time: 15min; incubation temperature: 60℃; injection needle temperature: 85℃; incubation speed: 500rpm; drift gas flow rate: 75mL / min; carrier gas flow rate: 0~2min 2mL / min, 2~10min 2~10mL / min, 10~20min 10~100mL / min, 20~30min 100mL / min; The characteristic volatile components are nonanone, octanone, heptanone, pentanone, 2,3-butanedione, acetaldehyde, and 2-pentylfuran; The pure milk with higher and lower preference ratings is distributed on both sides of the OPLS-DA model diagram.

2. The rapid digital evaluation method for the sensory flavor of pure milk based on GC-IMS according to claim 1, characterized in that, The levels of nonanone, octanone, heptanone, pentanone, 2,3-butanedione, acetaldehyde, and 2-pentylfuran in pure milk samples with high sensory evaluation scores were significantly different from those in pure milk samples with low sensory evaluation scores.

Images