Method, device and equipment for predicting bitterness of bovine casein zymolyte and storage medium

Abstract

The invention relates to the technical field of dairy product production, and provides a method, device and equipment for predicting bitterness of bovine casein zymolyte and a storage medium, and the method comprises the following steps: obtaining the component content of each amino acid peptide fragment generating bitterness in bovine casein zymolyte to be detected; inputting the component content of each amino acid peptide fragment into a bitterness prediction model to obtain a bitterness value of the to-be-detected bovine casein enzymatic hydrolysate output by the bitterness prediction model, the bitterness prediction model being based on the sample component content of the amino acid peptide fragment generating bitterness in the sample bovine casein enzymatic hydrolysate, and the bitterness value of the to-be-detected bovine casein enzymatic hydrolysate being the bitterness value of the to-be-detected bovine casein enzymatic hydrolysate. And the bitterness label value corresponding to the sample bovine casein zymolyte is trained. According to the method, the bitterness value of the bovine casein zymolyte can be rapidly and efficiently measured, the bitterness value is predicted through the bitterness prediction model, the problem that the consistency of an artificial tasting mode is poor is avoided, and optimization of a subsequent enzymolysis process can be better guided.

Description

Technical Field

[0001] The invention relates to the technical field of dairy product production, and in particular to a method, device, equipment and storage medium for predicting the bitterness of bovine casein enzymatic hydrolysate. Background Art

[0002] Cow's milk is a food with high nutritional value, which contains a variety of proteins, such as α-lactalbumin, β-lactoglobulin and lactoferrin. However, the proteins in cow's milk, especially cow's milk casein, will produce bitter taste after enzymatic hydrolysis. This is mainly because the released bitter peptides contain hydrophobic amino acids. These hydrophobic amino acids are exposed in the polypeptide chain and can stimulate the taste buds and produce bitterness. Bitterness is a taste with a very low threshold (i.e., easy to perceive), which will lead to a decrease in sensory quality. When the content of bitter substances is low, it can be detected, affecting the acceptance of food and further affecting the application of cow's milk casein enzymatic hydrolysate. Therefore, the determination of the bitterness value of cow's milk casein enzymatic hydrolysate plays a vital role in the subsequent application of cow's milk casein enzymatic hydrolysate.

[0003] At present, the bitterness value of bovine casein enzymatic hydrolysate is mostly determined by sensory measurement, that is, professional staff tastes the bovine casein enzymatic hydrolysate and scores its bitterness value (for example, 1-10 points). This method is subject to the influence of staff experience, physical condition and subjectivity, and the bitterness values ​​measured by different staff may be different, and the final measured bitterness value consistency is poor. In addition, each measurement requires the participation of staff, which has high labor costs. It may take multiple tastings to get the results during the measurement, and the measurement efficiency is low. Summary of the invention

[0004] The invention provides a method, device, equipment and storage medium for predicting the bitterness of bovine casein enzymatic hydrolysate, which are used to solve the problems of poor consistency and low measurement efficiency in the prior art for measuring the bitterness value of bovine casein enzymatic hydrolysate.

[0005] The invention provides a method for predicting the bitterness of bovine casein enzymatic hydrolysate, comprising the following steps.

[0006] Obtain the content of each bitter amino acid peptide in the tested bovine casein enzymatic hydrolysate.

[0007] The component content of each amino acid peptide segment is input into a bitterness prediction model to obtain the bitterness value of the bovine casein protein hydrolysate to be tested output by the bitterness prediction model.

[0008] The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

[0009] According to a method for predicting the bitterness of bovine casein enzymatic hydrolysate provided by the present invention, the amino acid peptide segments that produce bitterness include: the following seven types of amino acid peptide segments.

[0010] The C-terminus is a basic amino acid and the N-terminus is a hydrophobic amino acid.

[0011] The N-terminus is a basic amino acid and the C-terminus is a hydrophobic amino acid.

[0012] Both the C-terminus and the N-terminus are hydrophobic amino acids.

[0013] Both the C-terminus and the N-terminus are phenylalanine.

[0014] There are at least two hydrophobic amino acids at the C-terminus.

[0015] There are two phenylalanines at the C-terminus.

[0016] Arginine and proline are adjacent in a polypeptide.

[0017] Wherein, the basic amino acids include at least one of His, Arg and Lys, and the hydrophobic amino acids include at least one of Phe, Ala, Leu, Met, Ile, Trp, Pro and Val.

[0018] According to a method for predicting the bitterness of bovine milk casein enzymatic hydrolysate provided by the present invention, the content of each amino acid peptide segment producing bitterness in the bovine milk casein enzymatic hydrolysate to be tested is obtained, comprising the following steps.

[0019] A data table of all amino acid peptides in the bovine casein enzymatic hydrolysate to be tested is obtained, wherein the data table includes sequence expressions of all amino acid peptides and their corresponding component contents.

[0020] Based on the sequence expression of the amino acid peptide segment and the seven types of amino acid peptide segments, the component contents corresponding to the seven types of amino acid peptide segments are determined from the data table, and the component content of each type of amino acid peptide segment is the sum of the component contents of all amino acid peptide segments of the corresponding type.

[0021] According to a method for predicting the bitterness of bovine casein enzymatic hydrolysate provided by the present invention, the bitterness prediction model includes: a linear regression model or a random forest model.

[0022] According to a method for predicting the bitterness of bovine casein enzymatic hydrolysate provided by the present invention, the bitterness prediction model is a random forest model, and the training method of the bitterness prediction model is as follows.

[0023] The sample component contents of each bitter amino acid peptide segment in a plurality of sample bovine casein enzymatic hydrolysates are obtained respectively to form a training data set.

[0024] The method of sampling with replacement is adopted, and the sample component content of each amino acid peptide segment of some samples of bovine casein enzymatic hydrolysate in the training data set is randomly extracted each time to train the decision tree. For each decision tree, the sample component content of some amino acid peptide segments is randomly selected as the feature set. During the training process, any feature is selected from the feature set for splitting each time. When the loss function corresponding to the decision tree converges, the training of the decision tree is completed. After all the decision trees are trained, the training of the bitterness prediction model is completed.

[0025] According to a method for predicting the bitterness of bovine casein enzymatic hydrolysate provided by the present invention, before training the decision tree, the method further comprises: preprocessing the training data set to remove duplicate data and data with missing features.

[0026] According to a method for predicting the bitterness of bovine casein enzymatic hydrolysate provided by the present invention, for each sample bovine casein enzymatic hydrolysate, the corresponding bitterness label value is the average value of manual tasting results of the sample bovine casein enzymatic hydrolysate by multiple different staff members.

[0027] The present invention also provides a device for predicting the bitterness of bovine casein enzymatic hydrolysate, comprising the following modules.

[0028] The test data acquisition module is used to obtain the content of each amino acid peptide segment that produces bitterness in the bovine milk casein enzymatic hydrolysate to be tested.

[0029] The model execution module is used to input the component content of each amino acid peptide segment into the bitterness prediction model to obtain the bitterness value of the bovine milk casein enzymatic hydrolysate to be tested output by the bitterness prediction model.

[0030] The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

[0031] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the program, the method for predicting the bitterness of bovine casein enzymatic hydrolysate as described in any one of the above is implemented.

[0032] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for predicting the bitterness of bovine casein enzymatic hydrolysate as described in any one of the above is implemented.

[0033] The bitterness prediction method, device, equipment and storage medium of bovine milk casein protein hydrolysate provided by the present invention obtain the component content of each amino acid peptide segment that produces bitterness in the bovine milk casein protein hydrolysate to be tested; input the component content of each amino acid peptide segment into a bitterness prediction model to obtain the bitterness value of the bovine milk casein protein hydrolysate to be tested output by the bitterness prediction model, and the bitterness prediction model is based on the sample component content of the amino acid peptide segment that produces bitterness in the sample bovine milk casein protein hydrolysate, and the bitterness label value corresponding to the sample bovine milk casein protein hydrolysate is trained. This method can quickly and efficiently determine the bitterness value of the bovine milk casein protein hydrolysate, and predicts the bitterness value through the bitterness prediction model, avoiding the problem of poor consistency of the manual tasting method, and can better guide the optimization of the subsequent enzymatic hydrolysis process. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 This is one of the schematic flow diagrams of the bitterness prediction method for bovine casein enzymatic hydrolysate provided by the present invention.

[0036] Figure 2 This is the second schematic flow diagram of the bitterness prediction method of bovine casein enzymatic hydrolysate provided by the present invention.

[0037] Figure 3 The present invention is a schematic diagram of the structure of a bitterness prediction device for bovine casein enzymatic hydrolysate provided by the present invention.

[0038] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] The method for predicting bitterness of bovine casein protein hydrolysate according to the embodiment of the present invention is as follows: Figure 1 As shown, the process includes the following steps S110 and S120.

[0041] Step S110: Obtain the content of each bitter amino acid peptide in the tested bovine milk casein enzymatic hydrolysate. Specifically, for the tested bovine milk casein enzymatic hydrolysate, the composition and content of each bitter amino acid peptide in the tested bovine milk casein enzymatic hydrolysate can be obtained by peptidomics analysis methods such as liquid chromatography (HPLC), Q-Exactive HF-X mass spectrometer, MALDI TOF / TOF mass spectrometer, etc.

[0042] Step S120: Input the component content of each amino acid peptide segment into a bitterness prediction model to obtain the bitterness value of the bovine milk casein protein hydrolysate to be tested output by the bitterness prediction model. The bitterness prediction model is trained based on the sample component content of the amino acid peptide segment that produces bitterness in the sample bovine milk casein protein hydrolysate and the bitterness label value corresponding to the sample bovine milk casein protein hydrolysate. The bitterness label value corresponding to the sample bovine milk casein protein hydrolysate can be manually tasted and calibrated by professional staff, and the bitterness degree is scored from 0 to 10. The higher the bitterness degree, the higher the score.

[0043] In the method for predicting the bitterness of bovine casein protein hydrolysate of the present embodiment, the component content of each amino acid peptide segment that produces bitterness in the bovine casein protein hydrolysate to be tested is obtained; the component content of each amino acid peptide segment is input into a bitterness prediction model to obtain the bitterness value of the bovine casein protein hydrolysate to be tested output by the bitterness prediction model, and the bitterness prediction model is based on the sample component content of the amino acid peptide segment that produces bitterness in the sample bovine casein protein hydrolysate, and the bitterness label value corresponding to the sample bovine casein protein hydrolysate is trained. This method can quickly and efficiently determine the bitterness value of bovine casein protein hydrolysate, and predicting the bitterness value through the bitterness prediction model avoids the problem of poor consistency of the manual tasting method, and can better guide the optimization of the subsequent enzymatic hydrolysis process.

[0044] Since a large number of amino acid peptides are produced in the enzymatic hydrolysate after the enzymatic hydrolysis of bovine casein, there are amino acid peptides that do not produce bitterness, and there are also amino acid peptides that produce bitterness. The number of amino acid peptides that produce bitterness is also very large, and the types and quantities of amino acid peptides that produce bitterness in different bovine casein enzymatic hydrolysates are also different. The specific amino acid peptides in the bovine casein enzymatic hydrolysate have a certain randomness. Therefore, in order to facilitate model training and reduce the complexity of model training, in some embodiments, the amino acid peptides are divided into the following seven categories. Specifically, the amino acid peptides that produce bitterness include: the following seven types of amino acid peptides.

[0045] The C-terminus (i.e. carbon terminus) is a basic amino acid (H, R, K), and the N-terminus (i.e. nitrogen terminus) is a hydrophobic amino acid (F, A, L, M, I, W, P, V).

[0046] The N-terminus is a basic amino acid (H, R, K), and the C-terminus is a hydrophobic amino acid (F, A, L, M, I, W, P, V).

[0047] Both the C-terminus and the N-terminus are hydrophobic amino acids (F, A, L, M, I, W, P, V).

[0048] Both the C-terminus and the N-terminus are phenylalanine (F).

[0049] There are at least two hydrophobic amino acids at the C-terminus.

[0050] There are two phenylalanines (F) at the C-terminus.

[0051] Arginine (R) and proline (P) are adjacent to each other in a polypeptide.

[0052] Wherein, the basic amino acids include: at least one of histidine (His, abbreviated as H), arginine (Arg, abbreviated as R) and lysine (Lys, abbreviated as K), and the hydrophobic amino acids include: at least one of phenylalanine (Phe, abbreviated as F), alanine (Ala, abbreviated as A), leucine (Leu, abbreviated as L), methionine (Met, abbreviated as M), isoleucine (Ile, abbreviated as I), tryptophan (Trp, abbreviated as W), proline (Pro, abbreviated as P) and valine (Val, abbreviated as V). The component content of each amino acid peptide segment that produces bitterness is the component content of each major category of amino acid peptide segments, which is convenient for organizing the training data set for the bitterness prediction model.

[0053] In some embodiments, Figure 2 As shown, step S110 specifically includes the following steps S210 to S220.

[0054] Step S210: Obtain a data table of all amino acid peptides in the bovine milk casein enzymatic hydrolysate to be tested, the data table including sequence expressions of all amino acid peptides and their corresponding component contents, for example, a data table of all amino acid peptides in the bovine milk casein enzymatic hydrolysate to be tested can be obtained by peptidomics analysis such as MALDI TOF / TOF mass spectrometer. The sequence expression of the amino acid peptide is represented by a string formed by the abbreviation letters of each amino acid, for example: LKK.

[0055] Step S220: Based on the sequence expression of the amino acid peptide segment and the rules of the seven types of amino acid peptide segments, determine the component contents corresponding to the seven types of amino acid peptide segments from the data table. Specifically, the amino acid sequence expression in the data table is screened according to the rule of left nitrogen and right carbon. As long as it meets one of the types, it is screened out and classified into that type, thereby obtaining amino acid peptide segments that meet the seven types and their corresponding component contents. For example: LKK, the nitrogen end on the left is the hydrophobic amino acid L (leucine), and the carbon end on the right is the basic amino acid K (lysine). Therefore, the amino acid peptide segment LKK belongs to one of the seven types of amino acid peptide segments, and the component content of LKK is extracted. The component content of each type of amino acid peptide segment is the sum of the component contents of all amino acid peptide segments of the corresponding type.

[0056] Through the steps of this embodiment, the content of the components of the seven types of amino acid peptides that produce bitterness can be quickly and accurately screened from all the amino acid peptides in the bovine casein enzymatic hydrolysate to be tested.

[0057] It should be noted that it is possible that an amino acid peptide segment meets multiple types. Amino acid peptide segments that meet multiple types are only counted once and can be randomly assigned to one of the types they meet. Since the random forest model has good generalization, even if amino acid peptide segments that meet multiple types are randomly assigned to one type, a relatively accurate bitterness prediction result can be obtained.

[0058] In some embodiments, the bitterness prediction model includes: a linear regression model or a random forest model. Specifically, for bovine milk casein enzymatic hydrolysate, the more bitter amino acid peptides there are, the more obvious the bitterness is, that is, the greater the bitterness value is, and the overall linear relationship is present, so it can be modeled and predicted based on a linear regression model.

[0059] Furthermore, since the bovine milk casein protein hydrolysate contains a large amount of amino acid peptides that produce bitterness, the bitterness degree is also different, and different bovine milk casein protein hydrolysates contain different amino acid peptides that produce bitterness, which has a certain randomness. Therefore, in this embodiment, the random forest model is preferably used as the basis, and the bitterness prediction model is constructed based on the component content of the above seven major categories of amino acid peptides that produce bitterness. The introduction of randomness makes the random forest model not easy to overfit, has strong noise resistance, and has a very high prediction accuracy.

[0060] In some embodiments, the bitterness prediction model is a random forest model, and the bitterness prediction model is trained as follows: the sample component content of each amino acid peptide segment that produces bitterness in multiple sample bovine casein enzymatic hydrolysates is obtained respectively to form a training data set. Specifically, each amino acid peptide segment that produces bitterness can be obtained from a public bitter peptide component database (e.g., BIOPEP-UWM, BitterX, iBitter-Fuse, BitterDB, etc.), bitter peptide segments shown in various relevant documents, laboratory test data (e.g., peptideomics analysis results), etc., and then each amino acid peptide segment that produces bitterness is classified according to the above seven categories to obtain the sample component content corresponding to each category of amino acid peptide segment (for sample data, it can also be processed according to the above steps S210 and S220).

[0061] The method of sampling with replacement is adopted, and the sample component content of each amino acid peptide segment of some samples of bovine casein enzymatic hydrolysate in the training data set is randomly extracted each time to train the decision tree. For each decision tree, the sample component content of some amino acid peptide segments is randomly selected as the feature set. During the training process, any feature is selected from the feature set for splitting each time. When the loss function corresponding to the decision tree converges, the training of the decision tree is completed. After all the decision trees are trained, the training of the bitterness prediction model is completed.

[0062] Specifically, the loss function corresponding to the decision tree is the mean squared error (MSE). During the training of the decision tree, the bitterness prediction value and bitterness label value of the sample bovine casein enzymatic hydrolysate are substituted into the MSE. When the MSE converges, the decision tree training is completed. It should be noted that the process of decision tree training is also the process of decision tree creation. When the loss function converges, the decision tree creation is completed. Each random sampling can obtain a decision tree, and finally a trained random forest model is obtained.

[0063] In some embodiments, before training the decision tree, the process also includes: preprocessing the training data set to remove duplicate data and data with missing features. For example, for a sample of bovine casein enzymatic hydrolysate, if the sample component content of a certain amino acid peptide segment is missing, the sample is removed from the sample data set, thereby making the training sample data more accurate and the prediction results of the bitterness prediction model obtained through training more accurate.

[0064] In some embodiments, for each sample bovine casein enzymatic hydrolysate, the corresponding bitterness label value is the average value of the manual tasting results of the sample bovine casein enzymatic hydrolysate by multiple different staff members. For any sample bovine casein enzymatic hydrolysate, multiple professional staff members perform manual tasting on the sample bovine casein enzymatic hydrolysate, take the average value of the manual tasting results, and use the average value as the bitterness label value of the corresponding sample bovine casein enzymatic hydrolysate, further avoiding the problem of inaccurate bitterness label value caused by a single manual tasting result as a bitterness label value due to artificial subjective factors, making the bitterness label value more accurate, and the model predicts the bitterness value of the bovine casein enzymatic hydrolysate to be tested more accurately.

[0065] The bitterness prediction device for bovine casein enzymatic hydrolysate provided by the present invention is described below. The bitterness prediction device for bovine casein enzymatic hydrolysate described below and the bitterness prediction method for bovine casein enzymatic hydrolysate described above can be referenced to each other.

[0066] The bitterness prediction device for bovine casein enzymatic hydrolysate according to the embodiment of the present invention is as follows: Figure 3 As shown, it includes the following modules.

[0067] The test data acquisition module 310 is used to obtain the content of each bitter amino acid peptide segment in the tested bovine casein enzymatic hydrolysate.

[0068] The model execution module 320 is used to input the component content of each amino acid peptide segment into the bitterness prediction model to obtain the bitterness value of the bovine casein protein hydrolysate to be tested output by the bitterness prediction model.

[0069] The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

[0070] In some embodiments, the amino acid peptides that produce bitter taste include: the following seven types of amino acid peptides.

[0071] The C-terminus is a basic amino acid and the N-terminus is a hydrophobic amino acid.

[0072] The N-terminus is a basic amino acid and the C-terminus is a hydrophobic amino acid.

[0073] Both the C-terminus and the N-terminus are hydrophobic amino acids.

[0074] Both the C-terminus and the N-terminus are phenylalanine.

[0075] There are at least two hydrophobic amino acids at the C-terminus.

[0076] There are two phenylalanines at the C-terminus.

[0077] Arginine and proline are adjacent in a polypeptide.

[0078] Wherein, the basic amino acids include at least one of His, Arg and Lys, and the hydrophobic amino acids include at least one of Phe, Ala, Leu, Met, Ile, Trp, Pro and Val.

[0079] In some embodiments, the test data acquisition module 310 is specifically used to obtain a data table of all amino acid peptides in the bovine casein enzymatic hydrolysate to be tested, the data table including sequence expressions of all amino acid peptides and their corresponding component contents; based on the sequence expressions of the amino acid peptides and seven types of amino acid peptides, the component contents corresponding to the seven types of amino acid peptides are determined from the data table.

[0080] In some embodiments, the bitterness prediction model comprises: a linear regression model or a random forest model.

[0081] In some embodiments, the bitterness prediction model is a random forest model, and the training method of the bitterness prediction model is as follows.

[0082] The sample component contents of each bitter amino acid peptide segment in a plurality of sample bovine casein enzymatic hydrolysates are obtained respectively to form a training data set.

[0083] The method of sampling with replacement is adopted, and the sample component content of each amino acid peptide segment of some samples of bovine casein enzymatic hydrolysate in the training data set is randomly extracted each time to train the decision tree. For each decision tree, the sample component content of some amino acid peptide segments is randomly selected as the feature set. During the training process, any feature is selected from the feature set for splitting each time. When the loss function corresponding to the decision tree converges, the training of the decision tree is completed. After all the decision trees are trained, the training of the bitterness prediction model is completed.

[0084] In some embodiments, before training the decision tree, the process further includes: preprocessing the training data set to remove duplicate data and data with missing features.

[0085] In some embodiments, for each sample of bovine casein enzymatic hydrolysate, the corresponding bitterness label value is the average value of manual tasting results of the sample of bovine casein enzymatic hydrolysate by multiple different staff members.

[0086] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communications interface 420 and the memory 430 communicate with each other via the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the bitterness prediction method of bovine casein enzymatic hydrolysate, and the method includes the following steps.

[0087] Obtain the content of each bitter amino acid peptide in the tested bovine casein enzymatic hydrolysate.

[0088] The component content of each amino acid peptide segment is input into a bitterness prediction model to obtain the bitterness value of the bovine casein protein hydrolysate to be tested output by the bitterness prediction model.

[0089] The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

[0090] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0091] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the bitterness prediction method of bovine casein enzymatic hydrolysate provided by the above-mentioned methods, which includes the following steps.

[0092] Obtain the content of each bitter amino acid peptide in the tested bovine casein enzymatic hydrolysate.

[0093] The component content of each amino acid peptide segment is input into a bitterness prediction model to obtain the bitterness value of the bovine casein protein hydrolysate to be tested output by the bitterness prediction model.

[0094] The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

[0095] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the method for predicting the bitterness of bovine casein enzymatic hydrolysate provided by the above methods, and the method includes the following steps.

[0096] Obtain the content of each bitter amino acid peptide in the tested bovine casein enzymatic hydrolysate.

[0097] The component content of each amino acid peptide segment is input into a bitterness prediction model to obtain the bitterness value of the bovine casein protein hydrolysate to be tested output by the bitterness prediction model.

[0098] The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

[0099] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0100] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the bitterness of bovine casein enzymatic hydrolysate, characterized in that: include: Obtaining the content of each bitter amino acid peptide in the tested bovine casein enzymatic hydrolysate; The component content of each amino acid peptide segment is input into a bitterness prediction model to obtain the bitterness value of the bovine milk casein enzymatic hydrolysate to be tested output by the bitterness prediction model, The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

2. The method for predicting bitterness of bovine casein protein hydrolysate according to claim 1, characterized in that: The amino acid peptide segments producing bitter taste include the following seven types of amino acid peptide segments: The C-terminus is a basic amino acid, and the N-terminus is a hydrophobic amino acid; The N-terminus is a basic amino acid, and the C-terminus is a hydrophobic amino acid; Both the C-terminus and the N-terminus are hydrophobic amino acids; Both the C-terminus and the N-terminus are phenylalanine; There are at least two hydrophobic amino acids at the C-terminus; There are two phenylalanines at the C-terminus; Arginine and proline are adjacent in a polypeptide; Wherein, the basic amino acids include at least one of His, Arg and Lys, and the hydrophobic amino acids include at least one of Phe, Ala, Leu, Met, Ile, Trp, Pro and Val.

3. The method for predicting bitterness of bovine casein protein hydrolysate according to claim 2, characterized in that: Obtain the content of each bitter amino acid peptide in the tested bovine milk casein enzymatic hydrolysate, including: Obtaining a data table of all amino acid peptides in the bovine casein enzymatic hydrolysate to be tested, wherein the data table includes sequence expressions of all amino acid peptides and their corresponding component contents; Based on the sequence expression of the amino acid peptide segment and the seven types of amino acid peptide segments, the component contents corresponding to the seven types of amino acid peptide segments are determined from the data table, and the component content of each type of amino acid peptide segment is the sum of the component contents of all amino acid peptide segments of the corresponding type.

4. The method for predicting bitterness of bovine casein enzymatic hydrolysate according to any one of claims 1 to 3, characterized in that: The bitterness prediction model includes: a linear regression model or a random forest model.

5. The method for predicting bitterness of bovine casein protein hydrolysate according to claim 4, characterized in that: The bitterness prediction model is a random forest model, and the training method of the bitterness prediction model is as follows: Respectively obtaining the sample component content of each bitter amino acid peptide segment in a plurality of sample bovine milk casein enzymatic hydrolysates to form a training data set; The method of sampling with replacement is adopted, and the sample component content of each amino acid peptide segment of some samples of bovine casein enzymatic hydrolysate in the training data set is randomly extracted each time to train the decision tree. For each decision tree, the sample component content of some amino acid peptide segments is randomly selected as the feature set. During the training process, any feature is selected from the feature set for splitting each time. When the loss function corresponding to the decision tree converges, the training of the decision tree is completed. After the training of all decision trees is completed, the training of the bitterness prediction model is completed.

6. The method for predicting bitterness of bovine casein protein hydrolysate according to claim 5, characterized in that: Before training the decision tree, the training data set is also preprocessed to remove duplicate data and data with missing features.

7. The method for predicting bitterness of bovine casein enzymatic hydrolysate according to any one of claims 1 to 3, characterized in that: For each sample of bovine milk casein enzymatic hydrolysate, the corresponding bitterness label value is the average value of the manual tasting results of a plurality of different staff members on the sample of bovine milk casein enzymatic hydrolysate.

8. A bitterness prediction device for bovine casein enzymatic hydrolysate, characterized in that: include: The test data acquisition module is used to obtain the content of each amino acid peptide segment producing bitterness in the bovine milk casein enzymatic hydrolysate to be tested; A model execution module is used to input the component content of each amino acid peptide segment into a bitterness prediction model to obtain the bitterness value of the bovine milk casein protein hydrolysate to be tested output by the bitterness prediction model, The bitterness prediction model is obtained by training based on the sample component content of bitter amino acid peptides in the sample bovine milk casein enzymatic hydrolysate and the bitterness label value corresponding to the sample bovine milk casein enzymatic hydrolysate.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method for predicting the bitterness of bovine casein enzymatic hydrolysate according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the bitterness of bovine casein enzymatic hydrolysate according to any one of claims 1 to 7 is implemented.

Images