Method for auxiliary screening of grease antioxidant peptides based on artificial intelligence

Abstract

The invention discloses a method for auxiliary screening of grease antioxidant peptides based on artificial intelligence, and belongs to the technical field of bioinformatics. According to the invention, deep learning and virtual enzymolysis are combined, so that a novel method for preparing the antioxidant peptide by rapidly and efficiently screening protein enzymolysis is researched and developed. According to the method, an antioxidant peptide is obtained as a target, an antioxidant peptide data set and a test set are constructed, a Bi-LSTM high-precision deep learning strategy is provided based on the data set, and the Bi-LSTM high-precision deep learning strategy is trained to recognize the antioxidant peptide. And then, predicting the antioxidant activity of the polypeptide obtained through virtual enzymolysis by utilizing Bi-LSTM (Bidirectional Long Short Term Memory). Taking black bean protein as an example, the strategy of the invention discovers that the antioxidant peptide can be better obtained by enzymolysis of the black bean protein through trypsin. After ABTS and DPPH activity verification, the antioxidant peptide is used for lipophilicity determination, the peptide with lipophilicity can play a better grease antioxidant effect, and a research basis is provided for safety control of a grease hot working process.

Description

Technical Field

[0001] The present invention relates to a method for screening fat antioxidant peptides based on artificial intelligence assistance, and belongs to the technical field of bioinformatics. Background Art

[0002] During the production and storage process, vegetable oils and fats are easily oxidized and rancidified under the influence of external environments such as light, heat, moisture, oxygen, metal ions, and enzymes, thereby producing a variety of small molecules such as aldehydes, ketones, and acids. While reducing the quality of the oils and fats, they can also cause certain toxic hazards when ingested by the human body.

[0003] In order to inhibit the oxidation of plant oil products, existing research mainly uses antioxidants, including synthetic antioxidants and natural antioxidants. In order to avoid the potential hazards brought by synthetic antioxidants such as BHA, BHT, TBHQ, etc., their use is gradually being replaced by green and safe natural antioxidants.

[0004] Natural antioxidant peptides are a class of small molecule peptide chains with antioxidant activity, which can inhibit lipid oxidation by quenching free radicals, inhibiting oxidation reactions, capturing oxidation products, etc. However, there are tens of thousands of antioxidant peptide structures, and their biological activity effects vary with their structures. The current screening experimental methods for antioxidant peptides are time-consuming and laborious.

[0005] At present, artificial intelligence is being widely used to screen bioactive peptides. It is fast, efficient and highly accurate. Through virtual enzymatic hydrolysis and deep learning, biologically active peptides can be quickly and accurately obtained from proteins. Virtual enzymatic hydrolysis is an efficient protein hydrolysis preparation sequence prediction technology based on protein amino acid sequence and protease cleavage site. It is considered to be an effective alternative to traditional enzymatic hydrolysis, but it can only provide peptide sequences and cannot predict activity. As an important branch of artificial intelligence, deep learning can extract patterns, laws and knowledge from large amounts of data by automatically learning algorithms and models from data, and use this knowledge for tasks such as prediction, classification and identification. In the field of bioactive peptide screening, deep learning technology can further explore the potential relationship between peptide sequences and their biological activities, thereby achieving efficient screening and prediction of bioactive peptides.

[0006] Although artificial intelligence technology has made significant progress in the screening of bioactive peptides, it is still insufficient in the screening and prediction of antioxidant peptides. In particular, in recent years, antioxidant peptides from various sources have emerged in an endless stream, but few have used artificial intelligence methods for systematic screening and prediction. In addition, the application effects of the same antioxidant peptide in different environments also vary significantly, which makes the screening and application of antioxidant peptides more complex and challenging.

[0007] In summary, there is an urgent need for a method based on artificial intelligence to assist in the screening of antioxidant peptides in oils and fats. By combining the advantages of artificial intelligence technology, a large amount of information on antioxidant peptides can be quickly obtained, and antioxidant peptides that can play a role in oils and fats can be efficiently screened, which will play an important role in promoting the widespread application and in-depth research of antioxidant peptides in oils and fats. Summary of the invention

[0008] To solve the above problems, the present invention provides a method for artificial intelligence-assisted screening of oil antioxidant peptides. This method obtains antioxidant peptides based on deep learning and virtual enzymatic hydrolysis, and can ensure a high accuracy rate. Combined with oil antioxidant experiments, new oil antioxidant peptides are discovered, providing a research basis for oil antioxidant.

[0009] In a first aspect, the present invention provides a method for screening lipid antioxidant peptides based on artificial intelligence, the method comprising:

[0010] Step S1, collecting antioxidant peptides from various databases, deleting duplicate data and removing samples with inconsistent experimental results, obtaining a positive data set containing antioxidant peptides, randomly extracting an equal amount of polypeptide sequences matching the length of the positive data set as a negative control, and dividing the positive and negative sample data sets into a training set and a test set;

[0011] Step S2, inputting the data set in step S1 into a bidirectional long short-term memory network, using the training set to learn and summarize the characteristics of antioxidant peptides, and using the validation set data for verification, to obtain an antioxidant peptide prediction model;

[0012] Step S3, obtaining the amino acid sequence of the raw protein from the NCBI website, using different enzymes in the Novo Pro labs virtual enzymatic hydrolysis program to virtually hydrolyze the raw protein, and obtaining peptides after the raw protein is virtually hydrolyzed by different enzymes;

[0013] Step S4, using the antioxidant peptide prediction model obtained in step S2 to predict the antioxidant activity of the polypeptide obtained in step S3, and calculating the percentage of polypeptides with antioxidant effects, and using the enzyme with the highest percentage for actual enzymatic hydrolysis of the raw material protein to obtain antioxidant peptides;

[0014] Step S5, randomly selecting the antioxidant peptides obtained in step S4 to measure the ABTS and DPPH free radical scavenging activities to verify the antioxidant activity;

[0015] Step S6, importing the antioxidant peptide sequence obtained in step S5 into ACD / Labs software to calculate the logP of the compound;

[0016] Step S7, using the enzyme that can obtain more antioxidant peptides screened in step S3 to perform actual enzymatic hydrolysis on the protein, and perform antioxidant peptide identification, separation and purification;

[0017] The antioxidant peptides prepared in step S8 and step S7 are added to vegetable oil, and heating and frying simulation is performed to determine the inhibition rate of the antioxidant peptides on the formation of primary and secondary oxidation products after thermal processing of the vegetable oil and the inhibitory effect on the formation of harmful volatile aldehydes.

[0018] In one embodiment of the present invention, in step S1, each peptide contains 2-20 amino acids, and the positive and negative sample data sets are divided into a training set and a test set at a random ratio of 8:2.

[0019] In one embodiment of the present invention, in step S2, the features of the polypeptides in the training set are learned and summarized, and the polypeptides in the verification set are verified, and "0" and "1" are output, "0" indicates that the polypeptide has no antioxidant activity, and "1" indicates that the polypeptide has antioxidant activity.

[0020] In one embodiment of the present invention, in step S3, the length of the polypeptide sequence after virtual enzymatic hydrolysis of the raw material protein by different enzymes is less than 20.

[0021] In one embodiment of the present invention, in step S4, whether the polypeptide has antioxidant activity is determined based on the output result of the antioxidant peptide prediction model, "0" indicates that the polypeptide has no antioxidant activity, and "1" indicates that the polypeptide has antioxidant activity.

[0022] In one embodiment of the present invention, in step S4, a logP greater than 0 indicates fat solubility, and a higher value indicates greater fat solubility.

[0023] In one embodiment of the present invention, in step S8, the types of vegetable oils include but are not limited to one or more of peanut oil, sunflower oil, palm oil, soybean oil, olive oil, rapeseed oil, cottonseed oil, perilla seed oil, rice bran oil, linseed oil, safflower seed oil, tea seed oil, palm fruit oil, coconut oil, cocoa oil, almond oil, tung seed oil, corn germ oil, wheat germ oil, sesame seed oil, pumpkin seed oil, walnut oil, grape seed oil, sesame seed oil, pumpkin seed oil, cocoa butter and algae oil.

[0024] In one embodiment of the present invention, in step S2, the bidirectional long short-term memory network merges the outputs of two directional RNNs, one for forward processing of the sequence and the other for backward processing of the sequence.

[0025] In a second aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the method for artificial intelligence-assisted screening of lipid antioxidant peptides are implemented.

[0026] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for artificial intelligence-assisted screening of lipid antioxidant peptides.

[0027] The present invention has the following beneficial effects:

[0028] The present invention combines deep learning with virtual enzymatic hydrolysis to develop a new method for rapid and efficient screening of antioxidant peptides prepared by protein enzymatic hydrolysis. The present invention aims to obtain antioxidant peptides. By constructing an antioxidant peptide dataset and a test set, the present invention proposes a Bi-LSTM high-precision deep learning strategy based on the dataset and trains it to identify antioxidant peptides. Then, Bi-LSTM is used to predict the antioxidant activity of peptides obtained by virtual enzymatic hydrolysis.

[0029] Taking black bean protein as an example, the strategy of the present invention found that black bean protein can better obtain antioxidant peptides after trypsin hydrolysis. After the antioxidant peptides were verified by ABTS and DPPH activity, they were used for lipophilicity determination. Lipophilic peptides can exert better antioxidant effects on oils and fats, providing a research basis for the safety control of oil thermal processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings required to be used in the embodiments or the prior art description are briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without creative work.

[0031] Figure 1 This is a flow chart of the antioxidant peptide Bi-LSTM deep learning model provided by the present invention.

[0032] Figure 2 This is a graph of the 5-fold cross-training and validation results of the antioxidant peptide Bi-LSTM deep learning model provided by the present invention.

[0033] Figure 3 This is a diagram verifying the ABTS and DPPH free radical scavenging activity of the antioxidant peptide of the present invention.

[0034] Figure 4 This is a diagram showing the antioxidant activity of the primary and secondary products of the antioxidant peptide oil of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments 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.

[0036] Unless otherwise specified, the experimental materials, reagents, and instruments used in the examples of the present invention are all commercially available. Unless otherwise specified, the technical means used in the examples are all conventional means well known to those skilled in the art.

[0037] Example 1

[0038] The present invention provides a method for screening lipid antioxidant peptides based on artificial intelligence, the method comprising:

[0039] Step S1, collecting antioxidant peptides from various databases, deleting duplicate data and removing samples with inconsistent experimental results, obtaining a positive data set containing antioxidant peptides, randomly extracting an equal amount of polypeptide sequences matching the length of the positive data set as a negative control, and dividing the positive and negative sample data sets into a training set and a test set;

[0040] Step S2, inputting the data set in step S1 into a bidirectional long short-term memory network, using the training set to learn and summarize the characteristics of antioxidant peptides, and using the validation set data for verification, to obtain an antioxidant peptide prediction model;

[0041] Step S3, obtaining the amino acid sequence of the raw protein from the NCBI website, using different enzymes in the Novo Pro labs virtual enzymatic hydrolysis program to virtually hydrolyze the raw protein, and obtaining peptides after the raw protein is virtually hydrolyzed by different enzymes;

[0042] Step S4, using the antioxidant peptide prediction model obtained in step S2 to predict the antioxidant activity of the polypeptide obtained in step S3, and calculating the percentage of polypeptides with antioxidant effects, and using the enzyme with the highest percentage for actual enzymatic hydrolysis of the raw material protein to obtain antioxidant peptides;

[0043] Step S5, randomly selecting the antioxidant peptides obtained in step S4 to measure the ABTS and DPPH free radical scavenging activities to verify the antioxidant activity;

[0044] Step S6, importing the antioxidant peptide sequence obtained in step S5 into ACD / Labs software to calculate the logP of the compound;

[0045] Step S7, using the enzyme that can obtain more antioxidant peptides screened in step S3 to perform actual enzymatic hydrolysis on the protein, and perform antioxidant peptide identification, separation and purification;

[0046] The antioxidant peptides prepared in step S8 and step S7 are added to vegetable oil, and heating and frying simulation is performed to determine the inhibition rate of the antioxidant peptides on the formation of primary and secondary oxidation products after thermal processing of the vegetable oil and the inhibitory effect on the formation of harmful volatile aldehydes.

[0047] Optionally, in step S1, each peptide contains 2-20 amino acids, and the positive and negative sample data sets are divided into a training set and a test set at a random ratio of 8:2.

[0048] Optionally, in step S2, by learning and summarizing the features of the polypeptides in the training set and verifying the polypeptides in the verification set, "0" and "1" are output, where "0" indicates that the polypeptide has no antioxidant activity and "1" indicates that the polypeptide has antioxidant activity.

[0049] Optionally, in step S3, the length of the polypeptide sequence after virtual enzymatic hydrolysis of the raw material protein by different enzymes is less than 20.

[0050] Optionally, in step S4, whether the polypeptide has antioxidant activity is determined according to the output result of the antioxidant peptide prediction model, "0" indicates that the polypeptide has no antioxidant activity, and "1" indicates that the polypeptide has antioxidant activity.

[0051] Optionally, in step S4, a logP value greater than 0 indicates fat solubility, and a higher value indicates greater fat solubility.

[0052] Optionally, in step S8, the types of vegetable oils include but are not limited to one or more of peanut oil, sunflower oil, palm oil, soybean oil, olive oil, rapeseed oil, cottonseed oil, perilla seed oil, rice bran oil, linseed oil, safflower seed oil, tea seed oil, palm fruit oil, coconut oil, cocoa oil, almond oil, tung seed oil, corn germ oil, wheat germ oil, sesame seed oil, pumpkin seed oil, walnut oil, grape seed oil, sesame seed oil, pumpkin seed oil, cocoa butter and algae oil.

[0053] Optionally, in step S2, the bidirectional long short-term memory network merges the outputs of two directional RNNs, one for forward processing of the sequence and the other for backward processing of the sequence.

[0054] Example 2

[0055] like Figure 1-4 As shown, the present invention provides a method for screening lipid antioxidant peptides, the method comprising:

[0056] Step 1: In DFBP (http: / / www.cqudfbp.net / ), BIOPEP-UWM

[0057] The “Antioxidative peptide” was input into multiple databases, including the Biochemia database (https: / / biochemia.uwm.edu.pl / biopep / start_biopep.php), the Antimicrobial Peptide Database (https: / / aps.unmc.edu / ), PlantPepDB (http: / / 14.139.61.8 / PlantPepDB / index.php), and FermFooDb (https: / / webs.iiitd.edu.in / raghava / fermfoodb / ), to collect peptides with antioxidant activity, delete duplicate data, and remove samples with inconsistent experimental results. A positive data set containing 1,511 antioxidant peptides was obtained, each containing 2-20 amino acids. Equal amounts of peptide sequences matching the length of the positive data set were randomly extracted as negative controls. The positive and negative sample data sets were divided into training and test sets at a random ratio of 8:2. The training and test sets were divided by 5-fold cross validation. The composition of the model data set is shown in Table 1.

[0058] Table 1 Composition of model dataset

[0059]

[0060] Step 2: The constructed bidirectional long short-term memory network (Bi-LSTM) is a recurrent neural network (RNN) that combines the outputs of two directional RNNs, one for forward processing and the other for backward processing, which can effectively capture the long-term dependencies in sequence data. Figure 1 As shown. The amino acid sequence of the peptide is input into the Bi-LSTM neural network. By learning and summarizing the features of the peptides in the training set and verifying the peptides in the validation set, "0" and "1" are output. "0" means that the peptide has no antioxidant activity, and "1" means that the peptide has antioxidant activity, and the antioxidant peptide prediction model is obtained. The accuracy of the Bi-LSTM deep learning model on the 5-fold cross-validation set reached 0.8969. The accuracy (0.9043), precision (0.9767), sensitivity (0.8284), specificity (0.9802), MCC (0.8181) and AUC (0.9043) of the model on the independent test data set all show that it has excellent predictive ability. Among them, the accuracy evaluation of the Bi-LSTM deep learning model is shown in Table 2.

[0061] Table 2 Bi-LSTM model accuracy

[0062]

[0063] Step 3: The amino acid sequence of black bean protein was searched and obtained from the website of the National Center for Biotechnology Information (NCBI) (https: / / www.ncbi.nlm.nih.gov). The raw protein was virtually hydrolyzed using different enzymes in the virtual enzymatic hydrolysis program of Novo Pro labs / (https: / / www.novoprolabs.com / ), including alkaline protease, trypsin, Arg-C protease, chymotrypsin, Asp-N endopeptidase, and pepsin (PH>1.3), to obtain peptides after virtual enzymatic hydrolysis of the raw protein by different enzymes.

[0064] Step 4: Use the antioxidant peptide prediction model obtained in step 2 to predict the antioxidant activity of the polypeptide obtained in step 3, and calculate the percentage of polypeptides with antioxidant effects. The enzyme used when the percentage is the highest is used for the actual enzymatic hydrolysis of the raw protein to obtain antioxidant peptides. The results are shown in Table 3:

[0065] Table 3 Number of antioxidant peptides obtained by virtual enzyme digestion of black bean protein

[0066]

[0067] Step 5: In order to verify the accuracy of the deep learning model, a portion of the predicted peptides with antioxidant activity were selected for ABTS and DPPH free radical scavenging experiments to verify the antioxidant activity. The results are as follows: Figure 4 shown.

[0068] Step 6. Import the antioxidant peptide sequence obtained in step 5 into ACD / Labs software to calculate the logP of the compound. Generally, a logP greater than 0 indicates fat solubility. The higher the value, the greater the fat solubility. Table 4 shows the calculated lipophilicity values ​​of some antioxidant peptides.

[0069] Table 4 Calculation of lipophilicity of antioxidant peptides

[0070]

[0071] Step 7: Use the trypsin screened in step 3 to hydrolyze the black bean protein, identify the antioxidant peptides, and separate and purify them. The preparation of antioxidant peptides is carried out according to the method for preparing an amphiphilic antioxidant peptide and its application disclosed in the patent document with the publication number CN117567551A.

[0072] Step 8. Randomly select 4 antioxidant peptides in step 7 to conduct experiments on the inhibitory effect on oil oxidation. Before use, add the pre-prepared natural antioxidant peptides to the vegetable oil at an amount of 0.01% of the total oil. Stir evenly. After fully dissolved, heat the vegetable oil to about 180°C and continue heating for 2-12 hours to determine the inhibition rate of the antioxidant peptides on the production of primary and secondary oxidation products of the vegetable oil and the inhibitory effect on harmful volatile aldehydes.

[0073] Wherein, the step 7 specifically includes:

[0074] Step 7-1, determination of primary oxidation product inhibition rate;

[0075] Dissolve the polypeptide in oil, use BHT synthetic antioxidant as positive control, and heat to 180℃ for 6h. After heating, take out 10μL and place it in a 96-well plate. Add 200μL chloroform-methanol solution (7:3), 20μL 3% ammonium thiocyanate solution and 20μL 2mM ferrous chloride solution in sequence. Let stand at room temperature for 3min, and measure the absorbance value at 490nm on an ELISA reader, which is recorded as the peroxide value (A). The inhibition rate (%) of the sample on the primary oxidation products of oils and fats is calculated using the following formula:

[0076]

[0077] Where: I is the inhibition rate; A0 is the absorbance value of the oil; A is the absorbance value of the oil after adding the amphiphilic polypeptide;

[0078] Step 7-2, determination of inhibition rate of secondary oxidation products;

[0079] Weigh 37.5 mg of thiobarbituric acid and 1.5 g of trichloroacetic acid, dissolve them in deionized water, add 100 μL of hydrochloric acid, and dilute to 10 mL with deionized water. Take 50 μL of the heated oil and add 100 μL of thiobarbituric acid reagent, mix well and place in a boiling water bath for 15 minutes, then put in ice water for 10 minutes to cool to room temperature, centrifuge at 2000 r / min for 10 minutes, collect the supernatant, and measure the absorbance at 532 nm on an enzyme reader. The inhibition rate (%) of the sample on the secondary oxidation products of oil is calculated using the following formula:

[0080]

[0081] Where: I is the inhibition rate; A0 is the absorbance value of the oil; A is the absorbance value of the oil after adding the amphiphilic antioxidant peptide;

[0082] Step 7-3, determination of the content of harmful volatile aldehydes in frying oil:

[0083] Blank control: Take 100g blank antibiotic-free rapeseed oil sample, place it in a 500ml beaker, place it on a heatable magnetic stirrer, heat it to 180±5℃ within 10min, and heat it in 4 groups of 2h, 4h, 6h, and 8h. Take oil samples at the end of each frying to determine the content of oxidized α, β-unsaturated aldehydes in the oil samples;

[0084] Frying test with peptide: before heating, 100 ppm of the prepared antioxidant peptide was added to 100 g of sample oil to prepare the oil sample for frying experiment, and BHT synthetic antioxidant was used as positive control. Frying experiment was carried out according to the frying experiment method described in the blank control;

[0085] Accurately weigh 2g of the oxidized oil sample into a 20mL headspace bottle, add a stirrer to mix evenly; keep the sample in a constant temperature device at 50℃ for 15min, then insert the solid phase microextraction head into the sample bottle and extract it steadily for 60min. After the extraction is completed, insert it into the GC injection port and desorb for 10min;

[0086] Gas phase conditions: carrier gas is He (>99.999%), flow rate is 0.8mL / min. Column oven temperature program: initial temperature is 40℃, maintained for 6min, then increased to 80℃ at 4℃ / min; then increased to 86℃ at 2℃ / min, maintained for 3min; then increased to 120℃ at 4℃ / min, maintained for 2min; finally increased to 250℃ at 12℃ / min, maintained for 2min. Inlet: 230℃, split ratio is 5:1. Chromatographic column is HP-5MS conventional capillary column (30m×0.25mm×0.25μm);

[0087] Mass spectrometry conditions: transfer line 280°C, quadrupole 150°C, ion source 230°C; electron impact (EI) ion source; electron quantification 70 eV; solvent delay 3 min; full scan mode; mass scan range m / z 35-500.

[0088] Among them, the total amount of harmful volatile aldehydes after frying in oil is shown in Table 5, and the inhibition of the generation of harmful volatile aldehydes in oil by antioxidant peptides is shown in Table 6.

[0089] Table 5 Total amount of harmful volatile aldehydes after frying

[0090]

[0091] Table 6 Inhibition of harmful volatile aldehydes produced by antioxidant peptides in oils

[0092]

[0093]

[0094] From the above, it can be seen that the four oil antioxidant peptides not only have a better inhibitory effect on the primary and secondary oxidation products of vegetable oils than the synthetic antioxidant BHT, but also reduce the total oxidized α, β-unsaturated aldehydes to varying degrees at the same frying time. Among them, YAP showed a relatively good antioxidant effect at most time points, and its total oxidized α, β-unsaturated aldehyde content was lower than that of the control group CON and BHT in most cases. The antioxidant effects of IFS and BHT are relatively similar. At different frying times, their total oxidized α, β-unsaturated aldehyde content is reduced to a certain extent relative to CON, but not as good as YAP. The antioxidant effects of IFQ and YPL are not very stable, but overall they show a certain antioxidant effect.

[0095] In summary, trypsin hydrolysis of black bean protein is the best strategy for producing and screening oil antioxidant peptides, and the verified polypeptide sequence is quite consistent with the theoretical antioxidant peptide effect. The oil antioxidant property of the polypeptide obtained by enzymatic hydrolysis has also been well demonstrated. Therefore, the method for obtaining oil antioxidant peptides based on deep learning and virtual enzymatic hydrolysis provided by the present invention, in addition to providing guidance for the selection of enzymatic proteases, can also obtain polypeptide sequences with antioxidant activity, and can screen out oil antioxidant peptides with excellent effects.

[0096] In addition, the present invention also provides a computer device, which may include a processor, a memory, a network interface and a database connected via a system bus. 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 network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the processor executes the steps of a method for screening lipid antioxidant peptides based on artificial intelligence as described in any of the above embodiments.

[0097] The working process, working details and technical effects of the computer equipment provided in this embodiment can be found in the above embodiment of a method for artificial intelligence-assisted screening of lipid antioxidant peptides, which will not be described in detail here.

[0098] In addition, the present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of a method for screening lipid antioxidant peptides based on artificial intelligence as described in any of the above embodiments are implemented. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive and / or a memory stick, etc., and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0099] The working process, working details and technical effects of the computer-readable storage medium provided in this embodiment can be found in the above embodiment of a method for artificial intelligence-assisted screening of lipid antioxidant peptides, which will not be described in detail here.

[0100] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0101] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent replacements for some of the technical features therein. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for screening lipid antioxidant peptides based on artificial intelligence, characterized in that: The method comprises: Step S1, collecting antioxidant peptides from various databases, deleting duplicate data and removing samples with inconsistent experimental results, obtaining a positive data set containing antioxidant peptides, randomly extracting an equal amount of polypeptide sequences matching the length of the positive data set as a negative control, and dividing the positive and negative sample data sets into a training set and a test set; Step S2, inputting the data set in step S1 into a bidirectional long short-term memory network, using the training set to learn and summarize the characteristics of antioxidant peptides, and using the validation set data for verification, to obtain an antioxidant peptide prediction model; Step S3, obtaining the amino acid sequence of the raw protein from the NCBI website, using different enzymes in the Novo Pro labs virtual enzymatic hydrolysis program to virtually hydrolyze the raw protein, and obtaining peptides after the raw protein is virtually hydrolyzed by different enzymes; Step S4, using the antioxidant peptide prediction model obtained in step S2 to predict the antioxidant activity of the polypeptide obtained in step S3, and calculating the percentage of polypeptides with antioxidant effects, and using the enzyme with the highest percentage for actual enzymatic hydrolysis of the raw material protein to obtain antioxidant peptides; Step S5, randomly selecting the antioxidant peptides obtained in step S4 to measure the ABTS and DPPH free radical scavenging activities to verify the antioxidant activity; Step S6, importing the antioxidant peptide sequence obtained in step S5 into ACD / Labs software to calculate the logP of the compound; Step S7, using the enzyme that can obtain more antioxidant peptides screened in step S3 to perform actual enzymatic hydrolysis on the protein, and perform antioxidant peptide identification, separation and purification; The antioxidant peptides prepared in step S8 and step S7 are added to vegetable oil, and heating and frying simulation is performed to determine the inhibition rate of the antioxidant peptides on the formation of primary and secondary oxidation products after thermal processing of the vegetable oil and the inhibitory effect on the formation of harmful volatile aldehydes.

2. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 1, characterized in that: In step S1, each peptide contains 2-20 amino acids, and the positive and negative sample data sets are randomly divided into a training set and a test set at a ratio of 8:

2.

3. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 2, characterized in that: In step S2, by learning and summarizing the features of the polypeptides in the training set and verifying the polypeptides in the verification set, "0" and "1" are output, where "0" indicates that the polypeptide has no antioxidant activity, and "1" indicates that the polypeptide has antioxidant activity.

4. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 3, characterized in that: In step S3, the length of the polypeptide sequence after virtual enzymatic hydrolysis of the raw protein by different enzymes is less than 20.

5. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 4, characterized in that: In step S4, whether the polypeptide has antioxidant activity is determined based on the output result of the antioxidant peptide prediction model, "0" indicates that the polypeptide has no antioxidant activity, and "1" indicates that the polypeptide has antioxidant activity.

6. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 5, characterized in that: In step S4, if logP is greater than 0, it means fat solubility, and the higher the value, the greater the fat solubility.

7. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 6, characterized in that: In step S8, the types of vegetable oils include but are not limited to one or more of peanut oil, sunflower oil, palm oil, soybean oil, olive oil, rapeseed oil, cottonseed oil, perilla seed oil, rice bran oil, linseed oil, safflower seed oil, tea seed oil, palm fruit oil, coconut oil, cocoa bean oil, almond oil, tung seed oil, corn germ oil, wheat germ oil, sesame seed oil, pumpkin seed oil, walnut oil, grape seed oil, sesame seed oil, pumpkin seed oil, cocoa butter and algae oil.

8. The method for screening lipid antioxidant peptides based on artificial intelligence according to claim 6, characterized in that: In step S2, the bidirectional long short-term memory network merges the outputs of two directional RNNs, one for forward processing of the sequence and the other for backward processing of the sequence.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for screening lipid antioxidant peptides based on artificial intelligence are implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for artificial intelligence-assisted screening of lipid antioxidant peptides are implemented as described in any one of claims 1 to 7.

Images