Screening method of high-digestibility plant source protein and related device

Abstract

The invention discloses a high-digestibility plant source protein screening method and a related device, and relates to the technical field of high-digestibility plant source protein screening. Digestibility influence data obtained by measuring plant source protein to be judged are obtained; the digestibility influence data comprises the protein content, the secondary structure ratio, the intermolecular force, the solubility, the Zeta potential and the particle size of the plant source protein, determining the digestibility of the plant source protein by taking the digestibility influence data as input and utilizing the trained digestibility prediction model, judging whether the digestibility is greater than a preset value or not, and if yes, determining the digestibility of the plant source protein. According to the method, the digestion characteristics of the plant-derived proteins can be determined in vitro, the high-digestibility plant-derived proteins are screened, and the problems that in-vivo experiments are long in experimental period, high in cost, large in technical difficulty, poor in reproducibility and large in ethical dispute are solved.

Description

Technical Field

[0001] The present application relates to the technical field of screening high-digestibility plant-based proteins, and in particular to a screening method and related devices for high-digestibility plant-based proteins. Background Art

[0002] Compared with animal-based proteins, plant-based proteins have the advantages of promoting human health, being environmentally friendly, sustainable resources, and improving animal welfare. Therefore, the demand for plant-based proteins has reached an unprecedented high in recent years, and plant-based proteins have huge market potential. Compared with animal-based proteins, plant-based proteins are rich in dietary fiber that is beneficial to intestinal health and do not contain cholesterol. Increasing their daily intake can reduce the risk of chronic diseases. However, the molecular structure of plant-based proteins is complex and contains anti-nutrients, which reduces their solubility, digestibility, and absorption availability in the human body. Therefore, research on the digestibility characteristics of various types of plant-based proteins is urgent.

[0003] There are certain limitations in conducting in vivo experiments with animals or humans to study the digestion characteristics of plant-based proteins, which are manifested in long experimental cycles, high costs, high technical difficulties, poor reproducibility, and great ethical controversy. Therefore, a technology that can determine the digestion characteristics of plant-based proteins in vitro is needed to screen for high-digestibility plant-based proteins. Although there is currently no in vitro model that can replicate the entire complex process of plant-based protein digestion in vivo, it has been proven to be an effective alternative to in vivo experiments and an operational screening tool to solve problems such as food digestibility, bioaccessibility, bioavailability, release of bioactive compounds, and changes in food structure. Summary of the invention

[0004] The purpose of the present application is to provide a method and related device for screening plant-based proteins with high digestibility, which can determine the digestion characteristics of plant-based proteins in vitro and screen plant-based proteins with high digestibility.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a method for screening a high digestibility plant-based protein, the method comprising:

[0007] Obtaining digestibility influence data obtained by measuring the plant-based protein to be judged; the digestibility influence data includes protein content, secondary structure ratio, intermolecular force, solubility, Zeta potential and particle size of the plant-based protein; the intermolecular force includes ionic bond content, hydrogen bond content, hydrophobic interaction content and disulfide bond content;

[0008] Using the digestibility impact data as input, the digestibility of the plant-based protein is determined using a trained digestibility prediction model;

[0009] Determine whether the digestibility is greater than a preset value; if so, regard the plant-based protein as a high-digestibility plant-based protein.

[0010] Optionally, before using the digestibility influencing data as input and using a trained digestibility prediction model to determine the digestibility of the plant-based protein, the method for screening plant-based protein with high digestibility further includes:

[0011] Acquire a data set; the data set includes sample digestibility influencing data and sample digestibility of each of the plurality of plant-based protein samples; the sample digestibility influencing data includes protein content, secondary structure proportion, intermolecular force, solubility, Zeta potential and particle size of the plant-based protein sample;

[0012] Constructing an initial digestibility prediction model; the initial digestibility prediction model adopts a neural network model;

[0013] The initial digestibility prediction model is trained using the data set to obtain a trained digestibility prediction model.

[0014] Optionally, the method for obtaining the sample digestibility includes:

[0015] Digest the plant-based protein sample in vitro using a gastrointestinal simulated digestion model to obtain a digested sample;

[0016] Determining the digested sample to obtain the protein content of the digested sample;

[0017] The sample digestibility is determined based on the protein content of the digested sample and the protein content of the plant-based protein sample.

[0018] Optionally, before using the data set to train the initial digestibility prediction model, the method for screening high digestibility plant-based protein further comprises:

[0019] The data set is cleaned and normalized to obtain a normalized data set, and the normalized data set is used as a new data set; the cleaning includes filling in missing values ​​in the data set.

[0020] Optionally, before using the data set to train the initial digestibility prediction model, the method for screening high digestibility plant-based protein further comprises:

[0021] Taking all the sample digestibility influencing data in the data set as input, feature selection is performed using a correlation analysis method, a principal component analysis method or a recursive feature elimination method to obtain sample feature data; the sample feature data is several types of data in the sample digestibility influencing data;

[0022] Combining the sample characteristic data and sample digestibility of each of the plurality of plant-based protein samples into a new data set;

[0023] At this time, the digestibility of the plant-based protein is determined using the digestibility impact data as input and the trained digestibility prediction model, specifically including:

[0024] The digestibility influencing data are screened based on the sample characteristic data to obtain characteristic data; the characteristic data includes data of the same type as the data included in the sample characteristic data;

[0025] The characteristic data is used as input and the digestibility of the plant-based protein is determined using a trained digestibility prediction model.

[0026] Optionally, when using the data set to train the initial digestibility prediction model, the data set is divided into a training set, a validation set and a test set, the training set is used to train the model, the validation set is used to adjust the model hyperparameters, and the test set is used to evaluate the model performance; the loss function is mean square error or cross entropy loss; and the optimization algorithm is Adam algorithm or SGD algorithm.

[0027] Optionally, after obtaining the trained digestibility prediction model, the method for screening high digestibility plant-based protein further comprises:

[0028] Determine whether the time for obtaining the trained digestibility prediction model reaches a preset time;

[0029] If not, collecting new plant-based protein samples will affect the data and sample digestibility;

[0030] If so, the data set is expanded using the sample digestibility influence data and sample digestibility of the new plant-based protein sample to obtain an expanded data set, and the initial digestibility prediction model is trained again using the expanded data set to obtain a new trained digestibility prediction model.

[0031] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for screening high-digestibility plant-based protein described in any one of the above.

[0032] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for screening high-digestibility plant-based proteins as described above.

[0033] In a fourth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method for screening high-digestibility plant-based protein described in any one of the above.

[0034] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0035] The present application provides a method and related device for screening plant-based proteins with high digestibility, and obtains digestibility influence data obtained by measuring the plant-based proteins to be judged, the digestibility influence data including the protein content, secondary structure proportion, intermolecular force, solubility, Zeta potential and particle size of the plant-based proteins. The digestibility influence data is used as input, and a trained digestibility prediction model is used to determine the digestibility of the plant-based proteins, and it is determined whether the digestibility is greater than a preset value. If so, the plant-based proteins are regarded as plant-based proteins with high digestibility. The present application directly uses the digestibility influence data of the plant-based proteins to be judged as input, and uses a trained digestibility prediction model to determine the digestibility of the plant-based proteins, thereby determining the digestibility characteristics of the plant-based proteins in vitro, and further regarding the plant-based proteins with a digestibility greater than a preset value as plant-based proteins with high digestibility, thereby screening the plant-based proteins with high digestibility in vitro, and solving the problems of long experimental cycles, high costs, high technical difficulties, poor reproducibility and large ethical disputes existing in in vivo experiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 This is an application environment diagram of a method for screening plant-based proteins with high digestibility provided in Example 1 of the present application.

[0038] Figure 2 A schematic flow chart of a method for screening high-digestibility plant-based proteins provided in Example 1 of the present application.

[0039] Figure 3 This is a schematic diagram of the secondary structure ratio of the plant-based protein sample provided in Example 1 of the present application.

[0040] Figure 4 This is a schematic diagram of the solubility of the plant-based protein sample provided in Example 1 of the present application.

[0041] Figure 5 Schematic diagram of the Zeta potential and particle size of the plant-based protein sample provided in Example 1 of the present application.

[0042] Figure 6 A schematic diagram of the structure of a computer device provided in Example 2 of the present application. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0044] Example 1

[0045] The screening method for high digestibility plant-based protein provided in the embodiments of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal communicates with the server through the network. The data storage system can store data that the server needs to process. The data storage system can be set up separately, integrated on the server, or placed on the cloud or other servers. The terminal can send the plant-based protein to be judged to the server. After the server receives the plant-based protein to be judged, for the plant-based protein to be judged, the server obtains the digestibility influence data obtained by measuring the plant-based protein to be judged; using the digestibility influence data as input, the digestibility of the plant-based protein is determined using the trained digestibility prediction model; judging whether the digestibility is greater than the preset value; if so, the plant-based protein is regarded as a high digestibility plant-based protein. The server can feedback the judgment result of whether the plant-based protein to be judged is a high digestibility plant-based protein to the terminal.

[0046] In addition, in some embodiments, the screening method for high-digestibility plant-based proteins can also be implemented solely by a server or a terminal. For example, the terminal can directly screen the plant-based proteins to be judged, or the server can obtain the plant-based proteins to be judged from the data storage system and screen the plant-based proteins to be judged.

[0047] The terminal may be, but is not limited to, various desktop computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The server may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0048] like Figure 2 As shown, a method for screening high digestibility plant-based proteins is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server in the example is used to illustrate the following steps:

[0049] Step S1, obtaining digestibility influence data obtained by measuring the plant-based protein to be judged; the digestibility influence data includes the protein content, secondary structure proportion, intermolecular force, solubility, Zeta potential and particle size of the plant-based protein; the intermolecular force includes ionic bond content, hydrogen bond content, hydrophobic interaction content and disulfide bond content.

[0050] Step S2, taking the digestibility impact data as input, and using a trained digestibility prediction model to determine the digestibility of the plant-based protein.

[0051] Step S3, determining whether the digestibility is greater than a preset value; if so, treating the plant-based protein as a plant-based protein with a high digestibility.

[0052] By implementing the above-mentioned steps S1 to S3, this embodiment can directly use the digestibility influence data of the plant-based protein to be judged as input, and use the trained digestibility prediction model to determine the digestibility of the plant-based protein, so as to determine the digestibility characteristics of the plant-based protein in vitro, and further regard the plant-based protein with a digestibility greater than a preset value as a high-digestibility plant-based protein, so as to screen the high-digestibility plant-based protein in vitro, thereby solving the problems of long experimental cycle, high cost, high technical difficulty, poor reproducibility and large ethical controversy existing in in vivo experiments.

[0053] Globally, the plant-based proteins consumed in people's daily diets mainly come from cereals, legumes, seeds and nuts, and pseudo-cereals (also known as pseudo-cereals) and algae account for a relatively small proportion of the total intake. Due to differences in protein content, structural composition (secondary structure proportion and intermolecular forces), physicochemical properties (solubility, Zeta potential and particle size) and other aspects of various types of plant-based proteins, there are differences in their digestion degree in the human body. Therefore, this embodiment uses an alkali-dissolution and acid precipitation method to extract protein from fifteen common plant raw materials such as soybeans, peas and broad beans, and determines the protein content, structural composition and physicochemical properties of the above fifteen plant raw materials, and determines the digestibility of different plant-based proteins in vitro through a gastrointestinal simulation digestion model to train the machine learning model, and then the trained machine learning model can be used to screen high-digestibility plant-based proteins and develop highly digestible and absorbable plant protein bases.

[0054] Before using the digestibility influencing data as input and using the trained digestibility prediction model to determine the digestibility of the plant-based protein, the method for screening the high digestibility plant-based protein of this embodiment further includes:

[0055] (1) Obtaining a data set, wherein the data set includes sample digestibility influence data and sample digestibility of each of the multiple plant-based protein samples, wherein the sample digestibility influence data includes protein content, secondary structure ratio, intermolecular force, solubility, Zeta potential and particle size of the plant-based protein sample.

[0056] In this embodiment, fifteen common plant raw materials, including soybeans, peas, broad beans, chickpeas, lupines, white kidney beans, rice, brown rice, barley, wheat, oats, corn, potatoes, walnuts, and black sesame, are selected as plant raw material samples. The various plant raw material samples are extracted and purified through operations such as crushing, defatting, alkali dissolution and acid precipitation to generate a variety of plant-based protein samples with high purity. The protein content of the plant-based protein samples with high purity is at least 80%, and the protein content can be subsequently determined by the Kjeldahl method.

[0057] Each plant raw material sample was ground and pulverized by a pulverizer and then passed through a 70-mesh sieve to obtain plant powder. The plant powder was mixed with n-hexane to obtain defatted plant powder. Specifically, the plant powder and n-hexane were mixed at a ratio of 1:3 (w / v, mass concentration), and the mixed plant powder and n-hexane were magnetically stirred for 2 hours at room temperature, and the n-hexane was discarded. The mixing-magnetic stirring-discarding of n-hexane was repeated twice to fully remove the fat in the plant powder, and the solvent was evaporated in a fume hood overnight to obtain defatted plant powder. According to 1:20 (w / v), defatted plant powder is added to a beaker containing deionized water, mixed evenly, and a dispersion is obtained. The pH value of the dispersion is adjusted to 10.0 with 1M NaOH, and magnetic stirring is performed at room temperature for 1 hour. After centrifugation at a speed of 3000r / min for 15 minutes, the insoluble matter is removed to obtain a supernatant, and the pH value of the supernatant is adjusted to 4.0 with 1M HCl. After centrifugation at a speed of 3000r / min for 15 minutes, the supernatant is removed to obtain a precipitate. The precipitate is washed 3 times with deionized water, and dried after being frozen in a -80°C refrigerator for 1 day to obtain a plant-based protein powder, which is placed in a dryer for storage. By completing the above steps of extracting and purifying plant-based protein, the obtained plant-based protein powder is the plant-based protein sample corresponding to the plant raw material sample, so there are a total of fifteen plant-based protein samples in this embodiment.

[0058] For each plant-based protein sample, the plant-based protein sample was measured to obtain the sample digestibility impact data corresponding to the plant-based protein sample, as follows:

[0059] 1) Protein content

[0060] The protein content of the plant protein sample was determined by the Kjeldahl method, specifically the semi-automatic Kjeldahl method, which was a method slightly modified from the first method in GB 5009.5-2016 "National Food Safety Standard Determination of Protein in Food". The specific operation was as follows: 0.230 g (accurate to 0.001 g) of the plant protein sample was weighed into a digestion tube, and 3.5 g K2SO4, 0.4 g CuSO4·5H2O and 12.5 mL 1.84 g / L H2SO4 were added to the digestion tube for digestion. The digestion product was placed in a semi-automatic Kjeldahl nitrogen analyzer to realize the process of automatic addition of sodium hydroxide solution (400 g / L) and automatic distillation, and the distilled ammonia was purged with 30 mL 40 g / LH3BO3 is used for absorption, and the absorbent is placed in a conical flask. Two drops of 0.1% bromocresol green and 0.1% methyl red are added to the conical flask, and it is titrated with 0.1022 mol / L hydrochloric acid to calculate the nitrogen content in the plant-based protein sample and obtain the protein content of the plant-based protein sample.

[0061] 2) Proportion of secondary structure

[0062] The plant-derived protein sample is prepared into a liquid sample, and the liquid sample is measured by a circular dichroism spectrometer to obtain the secondary structure ratio of the plant-derived protein sample. The secondary structure ratio includes the ratio of α-helix, β-fold, β-turn and random coil (also called random coil). The secondary structure ratio of various plant-derived protein samples is as follows: Figure 3 shown.

[0063] The secondary structure ratio was determined as follows:

[0064] The plant-derived protein sample was dissolved in 0.01M pH 7 phosphate buffer to prepare a 1 mg / mL protein solution. After centrifugation at 3000 r / min for 15 min, the supernatant was loaded onto a circular dichroism spectrometer to determine the secondary structure ratio of the plant-derived protein sample.

[0065] 3) Intermolecular forces

[0066] The intermolecular forces of the plant-derived protein samples were measured to obtain the ionic bond content, hydrogen bond content, hydrophobic interaction content and disulfide bond content of the plant-derived protein samples. The ionic bond content, hydrogen bond content, hydrophobic interaction content and disulfide bond content are all intermolecular forces. The intermolecular forces of various plant-derived protein samples are shown in Table 1 below.

[0067] Table 1 Intermolecular forces

[0068]

[0069] The intermolecular forces were determined as follows:

[0070] Protein dissolving solutions with different selectivities were prepared, including SA solution, SB solution, SC solution, SD solution and SE solution. SA solution was 0.05mol / L NaCl solution, SB solution was 0.6mol / L NaCl solution, SC solution was 0.6mol / L NaCl+1.5mol / L urea solution, SD solution was 0.6mol / LNaCl+8mol / L urea solution, and SE solution was 0.6mol / LNaCl+8.0mol / L urea+1.5mol / Lβ-mercaptoethanol solution. All of the above five solutions were prepared in 0.01MpH7 phosphate buffer.

[0071] For each of the five solutions mentioned above, 0.5 g of plant-based protein sample was homogenously mixed with 5 mL of solution, allowed to stand for 2 h at 4 °C, and then centrifuged at 3000 r / min for 20 min. The protein content in the supernatant was determined by the Bradford method (Coomassie Brilliant Blue staining method) to obtain the protein content corresponding to each solution.

[0072] The ionic bond content, hydrogen bond content, hydrophobic interaction content and disulfide bond content are expressed by the differences in protein contents between SB solution and SA solution, SC solution and SB solution, SD solution and SC solution, and SE solution and SD solution, respectively, that is, the ionic bond content is equal to the difference between the protein content corresponding to SB solution and the protein content corresponding to SA solution, the hydrogen bond content is equal to the difference between the protein content corresponding to SC solution and the protein content corresponding to SB solution, the hydrophobic interaction content is equal to the difference between the protein content corresponding to SD solution and the protein content corresponding to SC solution, and the disulfide bond content is equal to the difference between the protein content corresponding to SE solution and the protein content corresponding to SD solution.

[0073] 4) Solubility

[0074] The solubility of the plant-based protein samples was measured to obtain the solubility of the plant-based protein samples. Figure 4 shown.

[0075] The solubility determination is as follows:

[0076] The solubility of the plant-based protein sample was characterized by measuring the content of soluble protein in the plant-based protein sample. Specifically, 0.30 g of the plant-based protein sample was mixed with 2.0 mL of potassium hydroxide solution (0.042 mol / L), and then oscillated for extraction for 1 hour. The supernatant was retained after centrifugation at 10,000 r / min for 10 minutes. The protein content in the supernatant was determined by the Bradford method, and then the solubility of the plant-based protein sample (the mass ratio of unit mass of protein that can be dissolved in the solution) was calculated based on the total amount of the plant-based protein sample added.

[0077] 5) Zeta potential and particle size

[0078] The plant-derived protein sample is prepared into a liquid sample, and the liquid sample is measured by a nanometer particle size Zeta potential instrument to obtain the Zeta potential and particle size of the plant-derived protein sample, and the particle size can be a nanometer particle size. The Zeta potential and particle size of various plant-derived protein samples are as follows: Figure 5 shown.

[0079] Zeta potential and particle size determination are as follows:

[0080] The plant-based protein sample was dissolved in 0.01M pH 7 phosphate buffer to prepare a 1 mg / mL protein solution. After centrifugation at 3000 r / min for 15 min, the supernatant was taken and placed on a nanoparticle size Zeta potential instrument. The supernatant was measured using the nanoparticle size Zeta potential instrument to obtain the Zeta potential and particle size of the plant-based protein sample.

[0081] The method for obtaining the sample digestibility includes: using a gastrointestinal simulated digestion model to digest the plant-based protein sample in vitro to obtain a digested sample. The digested sample is measured to obtain the protein content of the digested sample. Based on the protein content of the digested sample and the protein content of the plant-based protein sample, the sample digestibility is determined, and the difference between the protein content of the plant-based protein sample and the protein content of the digested sample is specifically calculated, and the ratio of the difference to the protein content of the plant-based protein sample is calculated to obtain the sample digestibility, and the in vitro digestibility determination is completed. The digestibility of various plant-based protein samples is shown in Table 2 below.

[0082] Table 2 Digestibility

[0083] Types of plant raw materials Digestibility pea 88.51％ Walnut 72.54％ oat 69.20％ broad bean 69.01％ Lupine 68.77％ Soybean 66.48％ Black sesame 65.99％ Rice 55.54％ white kidney beans 54.29％ wheat 52.35％ Chickpeas 49.61％ potato 49.55％ corn 37.49％ barley 35.06％ brown rice 34.54％

[0084] Specifically, this embodiment can use INFOGEST (standard static in vitro digestion model) 2.0, a gastrointestinal simulated digestion model, to digest the plant-based protein sample in vitro to generate a digested plant-based protein sample, that is, to obtain a digested sample. The digested sample is divided into a solid component and a liquid component. The protein content of the liquid component in the digested sample is determined by the Bradford method to obtain the protein content of the digested sample. Pepsin and pancreatic enzyme preparations are used to digest the plant-based protein sample in vitro.

[0085] The in vitro digestion of plant protein samples using INFOGEST 2.0 is as follows:

[0086] Simulated gastric stage: Weigh 1 g of plant-based protein sample, add 4 mL of distilled water, mix thoroughly, and then mix evenly with 5 mL of simulated gastric digestive fluid (including 4 mL of SGF (Simulated Gastric Fluid, simulated gastric fluid) stock solution, 5 μL of 0.3 M CaCl2 (H2O)2, 6.67 mg of pepsin (3000 U / mg), and a certain volume of deionized water), adjust the solution pH to 3.0 with 1 M hydrochloric acid, shake in a 37 ° C water bath for 2 hours, and then adjust the pH to 7.0 with 1 M sodium hydroxide solution to inactivate the enzyme.

[0087] Simulated intestinal stage: Add 10 mL of simulated intestinal digestive fluid (including 9 mL of SIF (Simulated Intestinal Fluid) stock solution, 40 μL of 0.3 M CaCl2(H2O)2, 20 mg of pancreatic enzyme, 80 mg of bile salt, and a certain volume of deionized water) to the product of the gastric stage, mix well, and place in a 37°C water bath for 2 hours, and then boil in water for 10 minutes to inactivate the enzyme.

[0088] The final product of the gastric and intestinal stages was mixed with an equal volume of 10% (w / v) trichloroacetic acid solution, allowed to stand for 30 min, and then centrifuged at 12000×g for 15 min at 4°C to collect the supernatant. The protein content of the supernatant was determined by the Bradford method to obtain the protein content of the digested sample.

[0089] In addition to the above-mentioned determination method, this embodiment can also collect data on the protein content, physical properties (structural composition and physicochemical properties) and digestibility of different plant protein source samples, and the data can come from scientific research, literature, databases, etc.

[0090] (2) Construct an initial digestibility prediction model, which uses a neural network model.

[0091] This embodiment selects a suitable neural network model architecture as the initial digestibility prediction model, such as a fully connected neural network (FCNN), a convolutional neural network (CNN) or a recurrent neural network (RNN).

[0092] (3) The initial digestibility prediction model is trained using the data set to obtain a trained digestibility prediction model.

[0093] Receive sample digestibility prediction data of each of the multiple plant-based protein samples, and link it with the sample digestibility of the corresponding plant-based protein sample, and establish a prediction relationship between the sample digestibility prediction data of each of the multiple plant-based protein samples and the sample digestibility of the corresponding plant-based protein sample to generate a trained digestibility prediction model.

[0094] Before using the data set to train the initial digestibility prediction model, the screening method for high digestibility plant-based protein in this embodiment also includes: data preprocessing, specifically including: cleaning and normalizing the data set to obtain a normalized data set, and using the normalized data set as a new data set to train the initial digestibility prediction model. Among them, cleaning includes filling missing values ​​in the data set, and the missing values ​​can be filled by interpolation or average value, and the interpolation can be determined by any existing interpolation method. The normalization process can use any existing normalization processing method, such as minimum-maximum normalization, etc. Data normalization can make model training more stable and efficient.

[0095] Before using the data set to train the initial digestibility prediction model, the screening method of high digestibility plant-based protein in this embodiment also includes: feature selection, specifically including: using all sample digestibility influencing data in the data set as input, using correlation analysis method, principal component analysis method (Principal Components Analysis, PCA) or recursive feature elimination method (Recursive Feature Elimination, RFE) to perform feature selection to obtain sample feature data, the sample feature data is several types of data in the sample digestibility influencing data, and the sample feature data and sample digestibility of each plant-based protein sample in the multiple plant-based protein samples are combined into a new data set, and the initial digestibility prediction model is trained using the new data set. Feature selection is to select the most important features, which can reduce model complexity and training time.

[0096] At this time, the digestibility influencing data is taken as input, and the digestibility of the plant-based protein is determined using a trained digestibility prediction model, specifically including: screening the digestibility influencing data based on the sample characteristic data to obtain characteristic data, wherein the data included in the characteristic data is of the same type as the data included in the sample characteristic data; taking the characteristic data as input, and determining the digestibility of the plant-based protein using a trained digestibility prediction model.

[0097] Before performing feature selection, this embodiment may also extract key features as needed, specifically extracting features that may affect the digestibility of plant-based protein from the original sample digestibility impact data.

[0098] When using the data set to train the initial digestibility prediction model, this embodiment divides the data set into a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used to adjust the model hyperparameters, and the test set is used to evaluate the model performance.

[0099] When using the training set to train the model, the loss function is the mean square error (MSE) or the cross entropy loss, and the optimization algorithm is the Adam algorithm or the SGD (Stochastic Gradient Descent) algorithm. During the training process, the sample digestion rate influencing data (or sample feature data) is used as input, and the sample digestion rate is used as a label, and the optimization algorithm is used to minimize the loss function.

[0100] When using the validation set to adjust the model hyperparameters, the training set is used to train models with different hyperparameters to obtain trained models with different hyperparameters. The validation set is used to compare the trained models with different hyperparameters and select the trained model with the smallest error as the validated model. The error can be calculated using R 2 (Coefficient of Determination, coefficient of determination), RMSE (Root Mean Square Error, root mean square error), MAE, etc.

[0101] When evaluating model performance using the test set, the evaluation index can be R 2 , RMSE, MAE, etc., and a relationship diagram (scatter plot) between the predicted value and the true value of the digestibility rate can also be drawn to visualize the model performance. According to the evaluation results of the evaluation indicators, if the accuracy requirements are not met, the model architecture, hyperparameters and training strategies can be further adjusted to further improve the model performance until the accuracy requirements are met. In this embodiment, cross-validation can be used to obtain a more robust model performance evaluation.

[0102] In this embodiment, early stopping, learning rate scheduling and regularization techniques can be used to prevent overfitting. Early stopping refers to training on the training set, and calculating the error of the trained model on the validation set after each training cycle. For example, every 15 epochs is a training cycle. When the error of the trained model on the validation set is larger than the error of the trained model on the validation set in the previous training cycle, the training is stopped, and the parameters in the previous training cycle are used as the final parameters of the model. Learning rate scheduling refers to gradually reducing the learning rate according to the number of epochs. Regularization technology refers to adding an L1 regularization term or an L2 regularization term to the loss function.

[0103] This embodiment further deploys the trained digestibility prediction model to a production environment, which can be a cloud server or a local server, so that new plant-based proteins can be predicted and analyzed at any time. Specifically, the digestibility impact data of the new plant-based protein is input into the trained digestibility prediction model to predict its digestibility, and based on the prediction results, high digestibility plant-based proteins are screened out. The method for determining the digestibility impact data of the new plant-based protein is the same as the method for determining the sample digestibility impact data of the plant-based protein sample.

[0104] This embodiment can also collect new data, that is, continuously collect new plant-based protein digestibility impact data and digestibility data, and regularly update the data set and retrain the model to maintain the accuracy and predictive ability of the model.

[0105] After obtaining the trained digestibility prediction model, the screening method for high digestibility plant-based protein in this embodiment also includes: determining whether the time for obtaining the trained digestibility prediction model reaches a preset time; if not, collecting sample digestibility influence data and sample digestibility of new plant-based protein samples; if so, using the sample digestibility influence data and sample digestibility of the new plant-based protein samples to expand the data set to obtain an expanded data set, and using the expanded data set to train the initial digestibility prediction model again to obtain a new trained digestibility prediction model.

[0106] This embodiment can also monitor and maintain the model, that is, monitor the performance of the trained digestibility prediction model in actual application, and perform maintenance and optimization as needed to ensure that the model can always provide reliable results. Through this neural network-based strategy, high digestibility plant-based proteins can be effectively screened out, providing important reference data for the food and feed industry.

[0107] This embodiment provides a screening method for screening high digestibility plant-based proteins to develop highly digestible and absorbable plant-based proteins. Highly digestible plant-based proteins can be screened in the following manner: using alkali dissolution and acid precipitation to extract protein from fifteen common plant raw materials, including soybeans, peas, broad beans, chickpeas, lupines, white kidney beans, rice, brown rice, barley, wheat, oats, corn, potatoes, walnuts, and black sesame, to determine the protein content, structural composition (secondary structure proportion and intermolecular force) and physicochemical properties (solubility, zeta potential and particle size) of the plant-based proteins, and using INFOGEST 2.0 This gastrointestinal simulation digestion model determines the in vitro digestibility of different plant-based proteins to train the machine learning model and obtain a trained digestibility prediction model. The trained digestibility prediction model is then used to determine the digestibility of new plant-based proteins, screen out high-digestibility plant-based proteins, and develop highly digestible and absorbable plant protein bases. It is an in vitro method for analyzing the digestibility of nutrients in plant-based proteins, and uses machine learning models to screen high-digestibility plant-based proteins to achieve the goals of being fast and accurate, saving time and effort, and reducing ethical disputes.

[0108] The screening method of this embodiment has the following advantages:

[0109] (1) This embodiment develops a method for screening high-digestibility plant-based proteins. The screening method combines chemical analysis with machine learning, which can significantly enhance the efficiency and accuracy of screening high-digestibility plant-based proteins, thereby achieving full coverage of rapid analysis and screening of the digestibility of plant-based proteins.

[0110] (2) This embodiment provides a method for screening highly digestible plant-based proteins, which is an efficient in vitro screening method. In vitro has been proven to be an effective alternative to in vivo experiments and an operational screening tool for solving problems such as food digestibility, bioaccessibility, bioavailability, release of bioactive compounds and changes in food structure. Therefore, it has significant advantages such as short test cycle, low operating cost and good reproducibility, and effectively avoids major ethical controversies.

[0111] (3) This embodiment uses a machine learning method to establish a predictive relationship between the protein content, structural composition and physicochemical properties of each plant-based protein sample in a plurality of extracted and purified plant-based protein samples and the in vitro digestibility of the corresponding plant-based protein sample, train the corresponding model and make further predictions, thereby effectively reducing the demand for real samples, improving analysis efficiency and increasing the accuracy of analysis results.

[0112] The embodiment of the present application also provides an application scenario, which applies the above-mentioned method for screening plant-based proteins with high digestibility. Specifically, the method for screening plant-based proteins with high digestibility provided in this embodiment can be applied in the plant-based protein screening scenario. The plant-based protein screening scenario includes an acquisition link, a screening link, and a display link. The acquisition link is used to acquire the plant-based protein to be judged, the screening link is used to use the trained digestibility prediction model to judge whether the plant-based protein is a plant-based protein with high digestibility, and the display link is used to display the judgment result of whether the plant-based protein is a plant-based protein with high digestibility. The method for screening plant-based proteins with high digestibility provided in this embodiment belongs to the screening link.

[0113] Example 2

[0114] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store plant-based proteins to be judged and related data thereof. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a screening method for high-digestibility plant-based proteins is implemented.

[0115] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0116] In an exemplary embodiment, a computer device is also provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for screening high-digestibility plant-based protein described in Example 1.

[0117] Example 3

[0118] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for screening high-digestibility plant-based protein described in Example 1 is implemented.

[0119] Example 4

[0120] An embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for screening high-digestibility plant-based proteins described in Example 1.

[0121] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0122] References to memory used in the various embodiments provided in the present application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0123] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0124] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for screening plant-based proteins with high digestibility, characterized in that: The screening method for the high digestibility plant-based protein comprises: Obtaining digestibility influence data obtained by measuring the plant-based protein to be judged; the digestibility influence data includes protein content, secondary structure ratio, intermolecular force, solubility, Zeta potential and particle size of the plant-based protein; the intermolecular force includes ionic bond content, hydrogen bond content, hydrophobic interaction content and disulfide bond content; Using the digestibility impact data as input, the digestibility of the plant-based protein is determined using a trained digestibility prediction model; Determine whether the digestibility is greater than a preset value; if so, regard the plant-based protein as a high-digestibility plant-based protein.

2. The method for screening high digestibility plant-based protein according to claim 1, characterized in that: Before determining the digestibility of the plant-based protein using the trained digestibility prediction model using the digestibility impact data as input, the method for screening plant-based protein with high digestibility further includes: Acquire a data set; the data set includes sample digestibility influencing data and sample digestibility of each of the plurality of plant-based protein samples; the sample digestibility influencing data includes protein content, secondary structure proportion, intermolecular force, solubility, Zeta potential and particle size of the plant-based protein sample; Constructing an initial digestibility prediction model; the initial digestibility prediction model adopts a neural network model; The initial digestibility prediction model is trained using the data set to obtain a trained digestibility prediction model.

3. The method for screening high digestibility plant-based protein according to claim 2, characterized in that: The method for obtaining the sample digestibility comprises: Digest the plant-based protein sample in vitro using a gastrointestinal simulated digestion model to obtain a digested sample; Determining the digested sample to obtain the protein content of the digested sample; The sample digestibility is determined based on the protein content of the digested sample and the protein content of the plant-based protein sample.

4. The method for screening high digestibility plant-based protein according to claim 2, characterized in that: Before using the data set to train the initial digestibility prediction model, the method for screening high digestibility plant-based protein further includes: The data set is cleaned and normalized to obtain a normalized data set, and the normalized data set is used as a new data set; the cleaning includes filling in missing values ​​in the data set.

5. The method for screening high digestibility plant-based protein according to claim 2, characterized in that: Before using the data set to train the initial digestibility prediction model, the method for screening high digestibility plant-based protein further includes: Taking all the sample digestibility influencing data in the data set as input, feature selection is performed using a correlation analysis method, a principal component analysis method or a recursive feature elimination method to obtain sample feature data; the sample feature data is several types of data in the sample digestibility influencing data; Combining the sample characteristic data and sample digestibility of each of the plurality of plant-based protein samples into a new data set; At this time, the digestibility of the plant-based protein is determined using the digestibility impact data as input and the trained digestibility prediction model, specifically including: The digestibility influencing data are screened based on the sample characteristic data to obtain characteristic data; the characteristic data includes data of the same type as the data included in the sample characteristic data; The characteristic data is used as input and the digestibility of the plant-based protein is determined using a trained digestibility prediction model.

6. The method for screening high digestibility plant-based protein according to claim 2, characterized in that: When using the data set to train the initial digestibility prediction model, the data set is divided into a training set, a validation set and a test set, the training set is used to train the model, the validation set is used to adjust the model hyperparameters, and the test set is used to evaluate the model performance; the loss function is mean square error or cross entropy loss; the optimization algorithm is Adam algorithm or SGD algorithm.

7. The method for screening high digestibility plant-based protein according to claim 2, characterized in that: After obtaining the trained digestibility prediction model, the method for screening high digestibility plant-based protein further includes: Determine whether the time for obtaining the trained digestibility prediction model reaches a preset time; If not, collecting new plant-based protein samples will affect the data and sample digestibility; If so, the data set is expanded using the sample digestibility influence data and sample digestibility of the new plant-based protein sample to obtain an expanded data set, and the initial digestibility prediction model is trained again using the expanded data set to obtain a new trained digestibility prediction model.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for screening high-digestibility plant-based protein according to any one of claims 1 to 7.

9. A 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 screening high-digestibility plant-based protein according to any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for screening high-digestibility plant-based protein according to any one of claims 1 to 7 is implemented.

Images