Fluorescent Chemical Sensor Array for Detecting Infant Formula and Method for Detecting Infant Formula Using the Same

Abstract

The present invention provides a fluorescent chemical sensor array for detecting infant formula milk powder and a method for detecting infant formula milk powder using the same. The fluorescent chemical sensor array includes the following sensor units: a first sensor unit, which is a compound having the structure shown in Formula I (i.e., PPE2); a second sensor unit, which is a compound having the structure shown in Formula II (i.e., PPE-SO3); and a third sensor unit, which is a compound having the structure shown in Formula III (i.e., PPE-N1). In Formula I, Formula II, and Formula III, n is independently 9-60. The present invention can use the fluorescent chemical sensor array and the LDA pattern recognition algorithm to detect and distinguish different types, different origins, and different brands of infant formula milk powder, and can optimize the sensor array to achieve the purpose of differentiating detection with the least number of sensing units.

Description

Technical Field

[0001] The present invention belongs to the field of chemical sensors, and particularly relates to a fluorescent chemical sensor array for detecting infant formula milk powder, a method for detecting infant formula milk powder by using the same, and a method for evaluating and optimizing the sensor array for detecting infant formula milk powder. Background Art

[0002] The identification of infant formula milk powder is crucial for the health of newborns and the market economy. When the mother is unable to breastfeed for various reasons, the growth and development of newborns are closely related to the quality of infant formula milk powder. Currently, there are many commercial infant formula milk powders on the market with different brands, types, origins, and price ranges. For higher profits, the price differences of infant formula milk powders from different sources pose a potential risk of counterfeit and inferior products. Nowadays, people's emphasis on the healthy growth of infants and the requirement for high-quality infant milk powder have led to an increasing social demand for simple, portable, and sensitive milk powder detection devices.

[0003] The existing milk powder detection methods are to detect multiple items in milk powder, such as fat, protein, calcium, phosphorus, iron, etc., one by one, so as to evaluate the quality. Although a high detection accuracy can be achieved, such detection methods rely on professional technicians and large, expensive equipment, and the sample pretreatment is very cumbersome, making their use mainly limited to laboratories and not suitable for on-site detection.

[0004] Traditional specific single sensors follow the "key-lock" model of "one sensor for one analyte", and use highly selective specific molecules to detect analytes. However, this method requires designing corresponding specific sensors for each analyte. The process of sensor design and optimization is relatively complex, time-consuming, and laborious.

[0005] A chemical sensor array is constructed by using a series of non-selective or semi-selective sensor units into an array. Each sensor unit individually generates different reaction signals with the analyte. The sum of these reaction signals constitutes the response pattern of the sensor array to the detected substance. By analyzing the response pattern of each analyte to be detected and comparing it with the known patterns in the database, the identification and detection of substances are realized, thus avoiding the dependence of a single sensor on specific receptors and greatly expanding the range of detectable substances. A fluorescent chemical sensor array uses a series of fluorescent molecules with cross-response to construct a sensor array to detect analytes, and realizes the identification of components to be detected by analyzing the changes in fluorescence signals. It has the characteristics of high sensitivity, no need for a reference system, rich output signals, and rapid response, which well meets the current social needs. Summary of the Invention

[0006] The object of the present invention is to provide a fluorescent chemical sensor array for detecting infant formula milk powder, a method for detecting infant formula milk powder using the same, and a method for evaluating and optimizing the sensor array for detecting infant formula milk powder.

[0007] The inventors of the present invention found in their research that substances and fluorescent molecules contained in infant formula milk powder interact through electrostatic and hydrophilic-hydrophobic interactions, thereby causing changes in the fluorescence signals of the fluorescent molecules. By using machine learning algorithms to identify and analyze the changes in the signals, it is possible to achieve differential detection between different milk powders. Since different fluorescent molecules have different contributions to the differential detection of milk powder, the contribution degrees of different fluorescent molecules are evaluated using the within-group variance / between-group variance of the algorithm as the evaluation criterion, the molecules with the greatest response contribution are screened out, and the differential detection of different types, different origins, and different brands is achieved using the smallest number of molecules. Thus, the present invention is proposed.

[0008] To achieve the above object, a first aspect of the present invention is to provide a fluorescent chemical sensor array for detecting infant formula milk powder, the fluorescent chemical sensor array comprising the following sensor units:

[0009] A first sensor unit, the first sensor unit being a compound having the structure shown in Formula I (i.e., PPE2),

[0010]

[0011] A second sensor unit, the second sensor unit being a compound having the structure shown in Formula II (i.e., PPE-SO3),

[0012]

[0013] A third sensor unit, the third sensor unit being a compound having the structure shown in Formula III (i.e., PPE-N1),

[0014]

[0015] Wherein, in Formula I, Formula II, and Formula III, n is independently 9-60.

[0016] According to the present invention, preferably, the first sensor unit is a compound having the structure shown in Formula I, the second sensor unit is a compound having the structure shown in Formula II, and the third sensor unit is a compound having the structure shown in Formula III. The first sensor unit, the second sensor unit, and the third sensor unit are all used in the form of a compound solution, and the concentration of the solution is calibrated with the absorbance peak value A = 0.2 of the compound.

[0017] According to the present invention, preferably, the first sensor unit, the second sensor unit and the third sensor unit are each independently arranged. For example, they are arranged in different wells of a 96-well plate.

[0018] A second aspect of the present invention is to provide a method for detecting infant formula, the method comprising the following steps:

[0019] (1) Mix each sensor unit in the above-mentioned sensor array with the test solution respectively and perform fluorescence scanning to obtain the fluorescence data of the test solution, and then calculate the ratio of the fluorescence data of the sensor array in the test solution group to the fluorescence data of the sensor array in the blank control group added with an equal amount of deionized water to obtain the fluorescence fold change data;

[0020] (2) Process the fluorescence fold change data with the LDA algorithm to obtain the LDA map of the fluorescence fold change data;

[0021] (3) Compare the position of the unknown sample data point in the LDA map with the data points in the LDA map of the known type of infant formula to determine the type, origin and brand of the infant formula in the test solution.

[0022] In the above method step (3), the LDA map of the known type of infant formula is obtained by measuring the known type of infant formula according to the methods of steps (1) and (2);

[0023] The test solution is a solution obtained by dissolving at least one commercial infant formula product with deionized water;

[0024] The LDA (Linear Discriminant Analysis) algorithm involved in the present invention is well-known to those skilled in the art and can be implemented in MATLAB language. In the present invention, the LDA algorithm is a supervised dimensionality reduction algorithm. At the beginning, labels are given to the data, and the projection direction is selected for data dimensionality reduction with the aim of small within-class variance and large between-class variance after projection.

[0025] Specifically, calculate the fold change I / I0 of the fluorescence value of the experimental group compared to the blank control group from the scanning data obtained by the microplate reader. Wherein, I represents the fluorescence intensity value of the experimental group, and I0 represents the fluorescence intensity value of the blank control group;

[0026] Then process the obtained fold change with the LDA algorithm written in MATLAB, and two-dimensional and three-dimensional graphs in the new coordinate system can be obtained. Among them, the coordinate factors one, two, and three of the three-dimensional graph respectively correspond to the first factor, the second factor, and the third factor of the LDA, that is, the coordinates of the new coordinate system obtained by orthogonal transformation for each purpose, such as Figure 2 , Figure 4 , Figure 6 , Figures 8 - 10, the percentage in parentheses represents the amount of information of the original data contained in the new coordinates. For example, 79.74% means that the new coordinates obtained by orthogonal transformation contain 79.74% of the information of the original data.

[0027] In the present invention, LDA processing is adopted, and the purpose is to obtain a visual classification effect after dimensionality reduction. Therefore, the LDA spectrum includes both LDA 2D spectrum and LDA 3D spectrum. The purpose of adopting HCA processing is to incorporate the features of all dimensions of the data into the analysis, perform independent feature analysis in the absence of data labels, and perform clustering.

[0028] In the present invention, the judgment method of "evaluating the detection effect of the present sensor array on infant formula milk powder according to the degree of dispersion of data points in the LDA spectrum" is well-known to those skilled in the art, and the degree of dispersion is usually based on the criterion that different points can be distinguished by the naked eye.

[0029] The method of the present invention is applicable to various target infant formula milk powders, especially the simultaneous detection of multiple target infant formula milk powders. The target infant formula milk powder is preferably ordinary infant formula milk powder, rather than premature infant formula milk powder, lactose-free infant formula milk powder, hydrolyzed protein formula milk powder, or other non-ordinary infant formula milk powders. The type of the target infant formula milk powder is preferably at least one of milk powder, sheep milk powder, goat milk powder, and soybean milk powder, and the origin is preferably at least one country among New Zealand, the Netherlands, Switzerland, Germany, Ireland, Denmark, Spain, and Australia.

[0030] The third aspect of the present invention is to provide a method for evaluating and optimizing a sensor array for detecting infant formula milk powder, and the method includes the following steps:

[0031] S1. Determine at least one target infant formula milk powder;

[0032] S2. Establish a sensor array, the sensor array includes a plurality of sensor units, each sensor unit corresponds to a fluorescent compound, and the fluorescent compound has a fluorescent response to the target infant formula milk powder;

[0033] S3. Prepare solutions A1-A of the target infant formula milk powder m ;

[0034] S4. Mix the sensor array with the target infant formula milk powder solutions A1-A m respectively, and perform fluorescence scanning to obtain experimental group fluorescence data D A1 -D Am , and then calculate the ratio of the fluorescence data of each experimental group to the fluorescence data of the sensor array blank control group to obtain fluorescence fold change data I A1 -I Am ;

[0035] S5. Process the fluorescence multiple change data I of the target infant formula milk powder using the LDA algorithm A1 -I Am to obtain the fluorescence multiple change data I of the target infant formula milk powder A1 -I Am of the LDA spectrum;

[0036] S6. Evaluate the detection effect of the sensor array on the infant formula milk powder according to the dispersion degree of the data points in the LDA spectrum.

[0037] According to a preferred embodiment of the present invention, the method further includes: calculating the contribution degrees of different sensor units, including:

[0038] 1) Calculate the within-group variance S w , and the calculation formula is

[0039] 2) Calculate the between-group variance S b , and the calculation formula is

[0040] 3) Calculate the contribution degree S w / S b ;

[0041] where n represents the samples of the infant formula milk powder; m represents the number of groups of experimental repetitions; represents the average value matrix of the characteristic values when the sensor unit distinguishes the i-th infant formula milk powder; represents the average value matrix of the characteristic values when the sensor unit distinguishes all infant formula milk powders; M i,j represents the j-th repeated experiment matrix when the sensor unit distinguishes the i-th infant formula milk powder.

[0042] Furthermore, the method further includes sorting the sensor units according to the contribution degrees, comparing the detection effects of the sensor arrays formed by different combinations of sensor units on the infant formula milk powder, and obtaining an optimized sensor array. The different combinations of sensor units can be determined based on the contribution degrees of the sensor units.

[0043] According to the present invention, preferably, the concentration of the solution of the target infant formula milk powder can be 1.5 mg / mL.

[0044] The present invention can realize the detection and distinction of different types, different origins, and different brands of infant formula milk powder by using a fluorescence-based chemical sensor array and an LDA pattern recognition algorithm, and can optimize the sensor array to achieve the purpose of completing the discrimination detection with the least number of sensing units.

[0045] Other features and advantages of the present invention will be described in detail in the following specific implementation section. Brief Description of the Drawings

[0046] Figure 1 (a)-1(b) shows the fluorescence responses of a sensor array including four sensor units to 8 infant formula milks from 4 different types. Among them, Figure 1 (a) is a picture of the original data of the fluorescence emission intensity. Figure 1 (b) is a picture of the logarithm of the fluorescence fold change.

[0047] Figure 2 A two-dimensional LDA analysis diagram showing the fluorescence responses of a sensor array including four sensor units to 8 infant formula milks from 4 different types.

[0048] Figure 3 (a)-3(b) shows the fluorescence responses of a sensor array including four sensor units to 18 infant formula milks from 9 different origins. Among them, Figure 3 (a) is a picture of the original data of the fluorescence emission intensity. Figure 3 (b) is a picture of the logarithm of the fluorescence fold change.

[0049] Figure 4 A three-dimensional LDA analysis diagram showing the fluorescence responses of a sensor array including four sensor units to 18 infant formula milks from 9 different origins.

[0050] Figure 5 (a)-5(b) shows the fluorescence responses of a sensor array including four sensor units to 6 infant formula milks from 2 different brands. Among them, Figure 5 (a) is a picture of the original data of the fluorescence emission intensity. Figure 5 (b) is a picture of the logarithm of the fluorescence fold change.

[0051] Figure 6 A one-dimensional LDA analysis diagram showing the fluorescence responses of a sensor array including four sensor units to 6 infant formula milks from 2 different brands.

[0052] Figure 7 Shows the calculation results of the contribution degrees of the four sensor units when the sensor array differentiates and detects the formulas of infant milk powders from different origins.

[0053] Figure 8 A two-dimensional LDA analysis diagram showing the fluorescence responses of the optimized sensor array of the present invention including three sensor units to 8 infant formula milks from 4 different types.

[0054] Figure 9The three - dimensional LDA analysis diagram shows the fluorescence response of the optimized sensor array of the present invention containing three sensor units to 18 infant formula powders from 9 different origins.

[0055] Figure 10 The one - dimensional LDA analysis diagram shows the fluorescence response of the optimized sensor array of the present invention containing three sensor units to 6 infant formula powders from 3 different brands. Detailed implementation manners

[0056] The present invention will be described below through specific embodiments, but the present invention is not limited thereto.

[0057] Unless otherwise specified, the experimental methods used in the following embodiments are all conventional methods; unless otherwise specified, the reagents, biological materials, etc. used in the following embodiments can all be obtained from commercial channels.

[0058] The first aspect of the present invention is to provide a fluorescence - type chemical sensor array for detecting infant formula powders. The fluorescence - type chemical sensor array includes the following sensor units:

[0059] The first sensor unit, and the first sensor unit is a compound having the structure shown in Formula I (i.e., PPE2),

[0060]

[0061] The second sensor unit, and the second sensor unit is a compound having the structure shown in Formula II (i.e., PPE - SO3),

[0062]

[0063] The third sensor unit, and the third sensor unit is a compound having the structure shown in Formula III (i.e., PPE - N1),

[0064]

[0065] Wherein, in Formula I, Formula II, and Formula III, n is independently 9 - 60.

[0066] The first sensor unit is a compound having the structure shown in Formula I, the second sensor unit is a compound having the structure shown in Formula II, and the third sensor unit is a compound having the structure shown in Formula III. The first sensor unit, the second sensor unit, and the third sensor unit are all used in the form of a compound solution, and the concentration of the solution is calibrated with the absorbance peak value A = 0.2 of the compound.

[0067] According to the present invention, preferably, the first sensor unit, the second sensor unit, and the third sensor unit are independently arranged.

[0068] The second aspect of the present invention is to provide a method for detecting infant formula milk powder, the method comprising the following steps:

[0069] (1) Mix the above-mentioned sensor array with the test solution and perform fluorescence scanning to obtain the fluorescence data of the test solution, and then calculate the ratio of the fluorescence data of the sensor array in the test solution group to the fluorescence data of the sensor array in the blank control group added with an equal amount of deionized water to obtain the fluorescence fold change data;

[0070] (2) Process the fluorescence fold change data with the LDA algorithm to obtain the LDA map of the fluorescence fold change data;

[0071] (3) Compare the positions of the data points of the unknown sample in the LDA map with the data points in the LDA map of the known type of infant formula milk powder to determine the type, origin and brand of the infant formula milk powder in the test solution.

[0072] In the above method step (3), the LDA map of the known type of infant formula milk powder is obtained by measuring the known type of infant formula milk powder according to the methods of steps (1) and (2);

[0073] The test solution is a solution obtained by dissolving at least one commercial infant formula milk powder product with deionized water;

[0074] Specifically, calculate the fold change I / I0 of the fluorescence value of the experimental group compared to the blank control group from the scanning data obtained by the microplate reader. Wherein, I represents the fluorescence intensity value of the experimental group, and I0 represents the fluorescence intensity value of the blank control group;

[0075] Then process the obtained fold change with the LDA algorithm written in MATLAB to obtain two-dimensional and three-dimensional graphs in a new coordinate system. Among them, the coordinate factors one, two, and three of the three-dimensional graph respectively correspond to the first factor, the second factor, and the third factor of the LDA, that is, the coordinates of the new coordinate system obtained by orthogonal transformation for each purpose, such as Figure 2 , Figure 4 , Figure 6 , Figures 8 - 10 , and the percentage in the parentheses represents the information amount of the original data contained in the new coordinate. For example, 79.74% means that the new coordinate obtained by orthogonal transformation contains 79.74% of the information of the original data.

[0076] The method of the present invention is applicable to various target infant formula milks, especially for the simultaneous detection of multiple target infant formula milks. The target infant formula milk is preferably ordinary infant formula milk, rather than premature infant formula milk, lactose-free infant formula milk, hydrolyzed protein formula milk, or other non-ordinary infant formula milks. The types of the target infant formula milk are preferably at least one of milk powder, sheep milk powder, goat milk powder, and soybean milk powder, and the place of origin is preferably at least one country among New Zealand, the Netherlands, Switzerland, Germany, Ireland, Denmark, Spain, and Australia.

[0077] The third aspect of the present invention is to provide a method for evaluating and optimizing a sensor array for detecting infant formula milk, and the method includes the following steps:

[0078] S1. Determine at least one target infant formula milk;

[0079] S2. Establish a sensor array, where the sensor array includes multiple sensor units, and each sensor unit corresponds to a fluorescent compound, and the fluorescent compound has a fluorescent response to the target infant formula milk;

[0080] S3. Prepare solutions A1-A of the target infant formula milk m ;

[0081] S4. Mix the sensor array with the target infant formula milk solutions A1-A respectively m and perform fluorescence scanning to obtain experimental group fluorescence data D A1 -D Am , and then calculate the ratio of the fluorescence data of each experimental group to the fluorescence data of the sensor array blank control group to obtain fluorescence fold change data I A1 -I Am ;

[0082] S5. Process the target infant formula milk fluorescence fold change data I A1 -I Am using the LDA algorithm to obtain the LDA map of the target infant formula milk fluorescence fold change data I A1 -I Am ;

[0083] S6. Evaluate the detection effect of the sensor array on infant formula milk according to the degree of dispersion of the data points in the LDA map.

[0084] According to a preferred embodiment of the present invention, the method further includes: calculating the contribution degrees of different sensor units, including:

[0085] 4) Calculate the within-group variance S w , and the calculation formula is

[0086] 5) Calculate the between-group variance S b , and the calculation formula is

[0087] 6) Calculate the contribution degree S w / S b ;

[0088] where n represents the samples of infant formula; m represents the number of groups of experimental repetitions; represents the average value matrix of the characteristic values when the sensor unit distinguishes the i-th infant formula; represents the average value matrix of the characteristic values when the sensor unit distinguishes all infant formulas; M i,j represents the j-th repeated experiment matrix when the sensor unit distinguishes the i-th infant formula.

[0089] Furthermore, the method further includes sorting the sensor units according to the contribution degree, and comparing the detection effects of the sensor arrays formed by different combinations of sensor units on infant formula to obtain an optimized sensor array. The different combinations of sensor units can be determined based on the contribution degree of the sensor units.

[0090] According to the present invention, preferably, the concentration of the solution of the target infant formula can be 1.5 mg / mL.

[0091] The present invention can use a fluorescent chemical sensor array and an LDA pattern recognition algorithm to realize the detection and differentiation of infant formulas of different types, different origins, and different brands, and can optimize the sensor array to achieve the purpose of distinguishing detection with the least number of sensing units.

[0092] Example 1

[0093] 1. Selection, preparation, and concentration setting of sensor units

[0094] The chemical fluorescent molecules respond to the protein on the surface of milk powder based on its different hydrophilicity-hydrophobicity, polarity, viscosity, and electrostatic interaction with the protein in the milk powder. Based on this principle, 4 chemical fluorescent molecules with different charge properties and different main and side chain structures are selected to build a sensor array.

[0095] The sensor array includes 4 fluorescent molecules: PPE2 (n = 9 - 60), PPE-SO3 (n = 9 - 60), PPE-N1 (n = 9 - 60), ANS (ammonium anthraquinone-1-sulfonate). The specific structures are as follows.

[0096] The preparation method of the fluorescent molecule PPE2 refers to: Fluorescence Array-Based Sensing of MetalIons Using Conjugated Polyelectrolytes. Wu, Y.; Tan, Y.; Wu, J.; Chen, S.; Chen, Y.Z.; Zhou, X.; Jiang, J.; Tan, C. ACS Appl. Mater. Interfaces 2015, 7, 6882 - 6888.

[0097] The preparation method of the fluorescent molecule PPE-SO3 refers to: Photophysics, aggregation and amplified quenching of a water-soluble poly(phenylene ethynylene). Tan, C.; Pinto, M.R.; Schanze, K.S. Chem. Comm. 2002, 446 - 447.

[0098] The preparation method of the fluorescent molecule PPE-N1 refers to: Fluorescence array-based sensing of nitroaromatics using conjugated polyelectrolytes. Wu, J.; Tan, C.; Chen, Z.; Chen, Y.Z.; Tan, Y.; Jiang, Y. Analyst 2016, 141, 3242 - 3245.

[0099] The fluorescent molecule ANS was purchased from Shanghai Aladdin Biochemical Technology Co., Ltd.

[0100] The concentration of the fluorescent molecule solution is calibrated with the peak absorbance A = 0.2 of the fluorescent molecule.

[0101]

[0102] 2. Selection and preparation concentration setting of infant formula

[0103] The infant formula was randomly purchased from the market as shown in Table 1. The concentration of the prepared solution of the infant formula is 1.5 mg / mL.

[0104] Table 1 Production, types and brands of commercial infant formula products *

[0105]

[0106] * All milk powders are selected as stage 2 products.

[0107] The brand name is represented by numbers, and milk powder samples with the same number represent different series from the same brand.

[0108] 3. Record the signal of the milk powder solution on a microplate reader using a 96-well plate.

[0109] (1) Add 100 μL of the detection solution of the sensor unit to each well. Add 100 μL of the milk powder solution to be detected to the experimental group, and add 100 μL of deionized water to the blank control group.

[0110] (2) Oscillate for 5 minutes;

[0111] (3) Set the scanning parameters (excitation wavelength 390 nm, scanning range 410 - 650 nm);

[0112] (4) Fluorescence scanning and data acquisition.

[0113] 4. Process the data

[0114] Calculate the fold change I / I0 of the fluorescence value of the experimental group compared to the blank control group from the scanning data obtained by the microplate reader. Here, I represents the fluorescence intensity value of the experimental group, and I0 represents the fluorescence intensity value of the blank control group.

[0115] Figure 1 (a)-(b) show the fluorescence responses of a sensor array containing four sensor units to 8 infant formula milk powders from 4 different types. Among them, Figure 1 (a) is a picture of the original data of the fluorescence emission intensity. Figure 1 (b) is a picture of the logarithm of the fluorescence fold change.

[0116] Then, the obtained fold change is processed using the LDA algorithm written in MATLAB to obtain a two-dimensional graph (such as Figure 2 ) and a three-dimensional graph (such as Figure 4 ). Among them, the first, second, and third coordinate factors of the three-dimensional graph correspond to the first factor, second factor, and third factor of LDA respectively. That is, the coordinates of the new coordinate system obtained through orthogonal transformation for each respective purpose. Such as Figure 2 , Figure 4 , Figure 6 , Figures 8 - 10 , and the percentage in the parentheses represents the amount of information of the original data contained in the new coordinates. For example, 79.74% means that the new coordinates obtained through orthogonal transformation contain 79.74% of the information of the original data.

[0117] (1) Differential detection of different types of infant formula milk powders by a fluorescence-based chemical sensor array

[0118] Figure 1(a)-1(b) shows the fluorescence responses of this fluorescent sensor array to different types of infant formula milk powder. It can be seen that the fluorescence responses of different sensor units to different types of infant formula milk powder are different. After processing with LDA, the results obtained are Figure 2 as shown. It can be found that this method can well distinguish different types of infant formula milk powder.

[0119] (2) Differential detection of infant formula milk powder from different origins by the fluorescent chemical sensor array

[0120] Figure 3 (a)-3(b) shows the fluorescence responses of this fluorescent sensor array to infant formula milk powder from different origins. It can be seen that the fluorescence responses of different sensor units to infant formula milk powder from different origins are different. After processing with LDA, the results obtained are Figure 4 as shown. It can be found that this method can well distinguish infant formula milk powder from different origins.

[0121] (3) Differential detection of infant formula milk powder from different brands by the fluorescent chemical sensor array

[0122] Figure 5 (a)-5(b) shows the fluorescence responses of this fluorescent sensor array to infant formula milk powder from different brands. It can be seen that the fluorescence responses of different sensor units to infant formula milk powder from different brands are different. After processing with LDA, the results obtained are Figure 6 as shown. It can be found that this method can well distinguish infant formula milk powder from different brands.

[0123] 7. Optimization of the sensor array

[0124] Figure 7 The calculation results of the contribution degrees of the four sensing units of this sensor array for differential detection of infant formula milk powder from different origins are shown. The smaller the value, the greater the contribution degree. Therefore, the contribution degree of PPE2 is the largest and that of ANS is the smallest. When the sensing units are added one by one from the largest to the smallest contribution degree to construct the sensor array, it is found that when performing data analysis with LDA for differential detection of different types, different origins, and different brands of infant formula milk powder. The results are as Figures 8 - 10 shown. Only a sensor array composed of three fluorescent molecules, PPE2 + PPE-N1 + PPE-SO3, is sufficient to distinguish different types, different origins, and different brands of infant formula milk powder. After optimizing the sensor array, not only the detection accuracy is not reduced, but also the number of sensing units is reduced, making the detection faster and more efficient.

[0125] The embodiments of the present invention have been described above. The above description is exemplary and not exhaustive, and is also not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for detecting infant formula, comprising the following steps: (1) Mix each sensor unit in the fluorescent chemical sensor array with the test solution respectively and perform fluorescence scanning to obtain the fluorescence data of the test solution. Then calculate the ratio of the fluorescence data of the sensor array in the test solution group to the fluorescence data of the blank control group sensor array added with an equal amount of deionized water to obtain the fluorescence multiple change data; The fluorescent chemical sensor array is composed of the following sensor units: The first sensor unit, and the first sensor unit is a compound having the structure shown in Formula I, that is, PPE2, Formula I The second sensor unit, and the second sensor unit is a compound having the structure shown in Formula II, that is, PPE-SO3, Formula II The third sensor unit, and the third sensor unit is a compound having the structure shown in Formula III, that is, PPE-N1, Formula III Wherein, in Formula I, Formula II, and Formula III, n is independently 9-60; (2) Process the fluorescence multiple change data with the LDA algorithm to obtain the LDA map of the fluorescence multiple change data; (3) Compare the position of the unknown sample data point in the LDA map with the data points in the LDA map of the known type of infant formula to determine the type, origin, and brand of the infant formula in the test solution.

2. The method according to claim 1, wherein: In step (1), each sensor unit is used in the form of a compound solution, and the concentration of the solution is calibrated with the absorbance peak value A = 0.2 of the compound; The test solution is a solution obtained by dissolving at least one commercial infant formula product with deionized water.

3. The method according to claim 1, wherein: In step (3), the LDA map of the known type of infant formula is obtained by measuring the known type of infant formula according to the methods of steps (1) and (2).

4. The method according to claim 1, characterized in that: The operation of the method is: calculate the multiple change I / I0 of the fluorescence value of the experimental group compared with the blank control group from the scanning data obtained by the microplate reader, where I represents the fluorescence intensity value of the experimental group and I0 represents the fluorescence intensity value of the blank control group; Then process the obtained multiple change with the LDA algorithm written in MATLAB to obtain two-dimensional and three-dimensional maps in a new coordinate system. The coordinate factors one, two, and three of the three-dimensional map respectively correspond to the first factor, the second factor, and the third factor of LDA, that is, the coordinates of the new coordinate system obtained by orthogonal transformation for each purpose.

5. The method according to claim 1, wherein: The infant formula is ordinary infant formula, The types of the infant formula are at least one of milk powder, sheep milk powder, goat milk powder, and soy milk powder, The origin is at least one country among New Zealand, the Netherlands, Switzerland, Germany, Ireland, Denmark, Spain, and Australia.

6. A method for evaluating and optimizing a sensor array for detecting infant formula, comprising the following steps: S1. Determine at least one target infant formula; S2. Establish a sensor array, and the sensor array includes multiple sensor units, and each sensor unit corresponds to a fluorescent compound, and the fluorescent compound has a fluorescence response to the target infant formula; S3. Prepare Solution A1-A of the target infant formula m ; S4. Respectively mix the said sensor array with the said target infant formula solution A1 - A m and perform fluorescence scanning to obtain the fluorescence data D A1 - D Am of the experimental group. Then calculate the ratio of the fluorescence data of each experimental group to the fluorescence data of the blank control group of the sensor array to obtain the fluorescence fold change data I A1 - I Am ; S5. Process the fluorescence multiple change data I of the target infant formula milk powder with the LDA algorithm A1 -I Am , and obtain the LDA spectrogram of the fluorescence multiple change data I of the target infant formula milk powder A1 -I Am ; S6. Evaluate the detection effect of the sensor array on infant formula according to the dispersion degree of data points in the LDA spectrum; After optimization, a sensor array composed of only three fluorescent molecules, PPE2 + PPE-N1 + PPE-SO3, is sufficient to distinguish different types, different origins, and different brands of infant formula; The PPE2 is the first sensor unit, and the first sensor unit is a compound having the structure shown in Formula I, Formula I The PPE-SO3 is the second sensor unit, and the second sensor unit is a compound having the structure shown in Formula II, Formula II The PPE-N1 is the third sensor unit, and the third sensor unit is a compound having the structure shown in Formula III, Formula III Wherein, in Formula I, Formula II, and Formula III, n is independently 9-60.

7. The method according to claim 6, wherein: The method further includes: calculating the contribution degrees of different sensor units, including: 1) Calculate the within-group variance , and the calculation formula is ; 2) Calculate the between-group variance , and the calculation formula is ; 3) Calculate the contribution degree ; where n represents the sample of infant formula; m represents the number of groups of experimental repetitions; represents the average value matrix of the characteristic values when the sensor unit distinguishes the i-th infant formula; represents the average value matrix of the characteristic values when the sensor unit distinguishes all infant formulas; represents the j-th repeated experiment matrix when the sensor unit distinguishes the i-th infant formula.

8. The method according to claim 7, wherein: The method further includes sorting the sensor units according to the contribution degrees and comparing the detection effects of sensor arrays formed by different combinations of sensor units on infant formula to obtain an optimized sensor array; The different combinations of sensor units are determined based on the contribution degrees of the sensor units.

Images