Method and equipment for screening xanthine oxidase inhibitory peptide

Abstract

The invention provides a method and equipment for screening xanthine oxidase inhibitory peptides. The method comprises the following steps: acquiring a first peptide fragment with xanthine oxidase inhibitory activity as a positive sample, and acquiring a second peptide fragment without activity as a negative sample; respectively carrying out activity labeling; performing quantization coding on the two peptide fragments by adopting amino acid physicochemical property descriptors to generate an initial feature matrix, and performing pooling operation on the matrix to obtain a fixed-length peptide fragment feature vector; training an ensemble learning classification model to obtain a xanthine oxidase inhibitory peptide prediction model by taking the fixed-length peptide fragment feature vector as input and the activity label as output; quantizing and coding the candidate peptide fragment set and then inputting the candidate peptide fragment set into the prediction model to obtain an inhibitory activity prediction probability; and screening out a xanthine oxidase inhibitory peptide candidate sequence with high confidence from the candidate peptide fragment set. Based on the method, the invention further provides screening equipment of the xanthine oxidase inhibitory peptide. According to the method, a large number of candidate peptide fragments can be quickly predicted, and a high-confidence candidate sequence is obtained.

Description

Technical Field

[0001] This invention belongs to the field of inhibitory peptide screening technology, and specifically relates to a method and apparatus for screening xanthine oxidase inhibitory peptides. Background Technology

[0002] Hyperuricemia is a metabolic disease characterized by elevated serum uric acid levels due to abnormal purine metabolism, and is a significant contributing factor to gout and various metabolic syndromes. Xanthine oxidase (XOD), a key rate-limiting enzyme in uric acid production, is a crucial target for current uric acid-lowering drug development. While clinical drugs such as allopurinol and febuxostat can effectively inhibit xanthine oxidase activity, safety concerns remain, including skin rash, hepatotoxicity, and cardiovascular risks. Therefore, identifying safer natural xanthine oxidase inhibitory peptides from dietary proteins has become a current research hotspot.

[0003] Existing active peptide screening technologies are mostly based on virtual enzymatic digestion, molecular docking, and activity prediction methods related to binding energy. For example, some technologies use virtual enzymatic digestion of protein substrates, combined with binding energy calculated by molecular docking, and use a small amount of IC50 data to construct a linear model to predict potential active peptides, where IC50 is the half-inhibitory concentration. However, such methods usually have the following shortcomings: (1) the correlation between binding energy and real biological activity is limited and easily affected by various factors; (2) training labels are mostly derived from estimated values ​​rather than experimental measurements, affecting the generalization ability of the model; (3) linear models are difficult to characterize complex structure-activity relationships, thus limiting the prediction accuracy.

[0004] Sheep milk proteins have a distinct amino acid composition and peptide release characteristics compared to cow milk, and are considered a potentially important source of functional peptides. However, current methods for precisely screening xanthine oxidase-inhibiting peptides from sheep milk are still lacking, and related technologies cannot yet effectively support efficient and accurate activity prediction. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a method and apparatus for screening xanthine oxidase-inhibiting peptides. This method enables rapid screening of a large number of peptides, significantly reducing experimental costs.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: This invention proposes a method for screening xanthine oxidase inhibitory peptides, comprising the following steps: The first peptide with xanthine oxidase inhibitory activity was obtained as a positive sample, and the second peptide without xanthine oxidase inhibitory activity was obtained as a negative sample; both the first and second peptides were labeled with activity. Using amino acid physicochemical property descriptors, both the first and second peptide segments are quantized and encoded to generate an initial feature matrix characterizing the physicochemical properties of the peptide segments. The initial feature matrix is ​​then pooled to obtain a fixed-length peptide segment feature vector. Using fixed-length peptide feature vectors as input features and activity labels as output labels, an ensemble learning classification model is trained to obtain a trained xanthine oxidase inhibitory peptide prediction model. The candidate peptide set to be screened is quantized and encoded and then input into the xanthine oxidase inhibitory peptide prediction model to obtain the predicted probability of inhibitory activity of each candidate peptide; based on the preset activity probability threshold, high-confidence xanthine oxidase inhibitory peptide candidate sequences are screened from the candidate peptide set.

[0007] The present invention also proposes a screening device for xanthine oxidase inhibitory peptides, comprising at least one processor and a memory, wherein the memory stores a computer program, and the computer program, when executed by the at least one processor, implements the screening method for xanthine oxidase inhibitory peptides.

[0008] The effects described in the invention are merely those of the embodiments, and not all the effects of the invention. One of the above technical solutions has the following advantages or beneficial effects: This invention proposes a method and device for screening xanthine oxidase inhibitory peptides. The method includes the following steps: obtaining a first peptide fragment with xanthine oxidase inhibitory activity as a positive sample, and obtaining a second peptide fragment without xanthine oxidase inhibitory activity as a negative sample; both the first and second peptide fragments are labeled with activity; using amino acid physicochemical property descriptors, both the first and second peptide fragments are quantized and encoded to generate an initial feature matrix characterizing the physicochemical properties of the peptides; the initial feature matrix is ​​pooled to obtain a fixed-length peptide feature vector; using the fixed-length peptide feature vector as input features and the activity label as output label, an ensemble learning classification model is trained to obtain a trained xanthine oxidase inhibitory peptide prediction model; the candidate peptide set to be screened is quantized and encoded and input into the xanthine oxidase inhibitory peptide prediction model to obtain the predicted probability of inhibitory activity of each candidate peptide fragment; based on a preset activity probability threshold, high-confidence xanthine oxidase inhibitory peptide candidate sequences are screened from the candidate peptide set. Based on this method for screening xanthine oxidase inhibitory peptides, a screening device for xanthine oxidase inhibitory peptides is also proposed. This invention can quickly predict a large number of candidate peptides, rapidly obtain high-confidence candidate sequences, and significantly reduce the workload and cost of experimental screening. Attached Figure Description

[0009] Figure 1 This is a flowchart of a method for screening xanthine oxidase inhibitory peptides proposed in Example 1 of the present invention; Figure 2This is a schematic diagram illustrating the performance evaluation results of the CatBoost model based on VHSE features proposed in Embodiment 1 of the present invention. Figure 3 The ROC curves and AUC comparison charts of different models proposed in Example 1 of this invention in the prediction of xanthine oxidase XOD inhibitory peptides are shown. Figure 4 This is a schematic diagram showing the distribution of peptide prediction probabilities output by the xanthine oxidase inhibitory peptide prediction model proposed in Example 1 of this invention. Figure 5 This is a schematic diagram of the molecular docking structure of TNAYF and xanthine oxidase XOD proposed in Example 1 of the present invention; Figure 6 This is a schematic diagram of the experimental results of the inhibition of xanthine oxidase XOD activity by TNAYF proposed in Example 1 of the present invention; Figure 7 This is a schematic diagram of a screening device for xanthine oxidase inhibitory peptides proposed in Embodiment 2 of the present invention. Detailed Implementation

[0010] To clearly illustrate the technical features of this solution, the invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the invention. To simplify the disclosure of the invention, components and arrangements of specific examples are described below. Furthermore, reference numerals and / or letters may be repeated in different examples. This repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. Descriptions of well-known components, processing techniques, and processes are omitted in this invention to avoid unnecessarily limiting the invention.

[0011] Example 1 Example 1 of this invention proposes a screening method for xanthine oxidase inhibitory peptides. This method is based on real experimental labels, combined with VHSE feature encoding and CatBoost classification model, which improves the accuracy and generalization ability of peptide activity prediction and is applicable to the rapid screening of peptides derived from sheep milk.

[0012] Figure 1 This is a flowchart of a method for screening xanthine oxidase inhibitory peptides proposed in Example 1 of the present invention; In step S1, a first peptide with xanthine oxidase inhibitory activity is obtained as a positive sample, and a second peptide without xanthine oxidase inhibitory activity is obtained as a negative sample; and both the first and second peptides are labeled with activity. Peptides with xanthine oxidase (XOD) inhibitory activity as determined in vitro from publicly available literature were collected as positive samples, while peptides with no or low activity as determined in vitro from publicly available literature were collected as negative samples.

[0013] In this application, positive and negative samples are measured at half-inhibition concentrations. The inhibition rate or other experimental indicators are used to determine this. For example, the half-inhibitory concentration of the obtained peptide. When the value is ≤ the first threshold, the peptide is labeled as a positive sample; when the half-inhibitory concentration of the obtained peptide is... If the value is greater than or equal to the second threshold, the peptide is labeled as a negative sample; wherein the second threshold is greater than the first threshold. The determination of the first and second thresholds in this application can be reasonably selected based on the actual situation.

[0014] Excluding half-inhibitory concentration Peptides in the "gray area" with values ​​between the first and second thresholds, as well as entries containing non-standard amino acids or with questionable activity data, are identified. The collected sequences are then deduplicated.

[0015] In this application, after deduplication and cleaning, 202 positive samples and 202 negative samples were obtained, with peptide lengths ranging from 3 to 10 amino acids. The scope of protection of this invention is not limited to the values ​​listed in Example 1, and those skilled in the art can make reasonable selections based on the actual situation.

[0016] In step S2, amino acid physicochemical property descriptors are used to quantize and encode both the first and second peptide segments to generate an initial feature matrix characterizing the physicochemical properties of the peptide segments. The initial feature matrix is ​​then pooled to obtain a fixed-length peptide feature vector.

[0017] In the implementation process of this application, based on the amino acid sequence of any peptide, the physicochemical property descriptor vector corresponding to each amino acid residue in the sequence is extracted sequentially; the amino acid physicochemical property descriptor is a VHSE descriptor; the VHSE descriptor characterizes the hydrophobicity, steric hindrance, and electronic effects of the amino acid through eight principal components. All extracted descriptor vectors are arranged and combined according to the order of the amino acid sequence to generate the initial feature matrix of any peptide. Table 1 below shows the eight-dimensional VHSE descriptor values ​​for different amino acids.

[0018] Table 1: Eight-dimensional VHSE descriptor values ​​for different amino acids

[0019] This application aims to unify peptide feature vectors of different lengths by performing pooling operations on the initial feature matrix of the feature dimension of peptides to obtain fixed-length peptide feature vectors. The specific process is as follows: The initial feature matrix is ​​input into a preset pooling function; the feature values ​​of each physicochemical property dimension are aggregated and calculated along the sequence length dimension of the initial feature matrix; a fixed-length peptide feature vector with dimension 1×N is output; where N is the vector dimension of the amino acid physicochemical property descriptor, and the pooling function includes mean pooling function, max pooling function and summation pooling function.

[0020] Sequence padding can also be used to fill in short peptide sequences to a set length and flatten the feature matrix to obtain peptide feature vectors of 8 to 80 dimensions.

[0021] This application also employs a dynamic length-adaptive pooling strategy, which intelligently selects the optimal pooling method based on peptide length: for short peptides (2-5 amino acids), maximum pooling is used to retain the prominent contribution of key active residues; for medium-length peptides (6-10 amino acids), weighted average pooling is used, where the weights are calculated through a lightweight attention mechanism to give higher weight to features at important positions; for longer peptides (11-15 amino acids), hierarchical convolutional pooling is used, first extracting local sequence patterns using one-dimensional convolution, and then performing global pooling on the convolutional features. This setup can more accurately capture the activity-determining features of peptides of different lengths.

[0022] The specific coding process implemented in this application is as follows: For any peptide in the dataset, such as the sequence "TNAYF" According to Table 1, the 8-dimensional VHSE vector corresponding to each amino acid (T, N, A, Y, F) is obtained sequentially.

[0023] Stack these vectors in sequence to form a 5-row (peptide length) 8-column (VHSE dimension) matrix, denoted as the initial feature matrix.

[0024] Because different peptides have different lengths, the number of rows in the generated matrix can vary. To obtain input features with fixed dimensions, mean pooling is used. That is, for each column of the above matrix (each VHSE dimension), the arithmetic mean of the values ​​of all amino acids in that dimension is calculated. Finally, a 1×8 dimensional global feature vector is obtained, which represents the average physicochemical properties of the entire peptide across the 8 VHSE dimensions.

[0025] To further enhance feature representation capabilities, additional features such as the amino acid composition (20 dimensions), molecular weight, and average hydrophobicity of the peptide can be concatenated onto the 8-dimensional VHSE mean vector to form a higher-dimensional fused feature vector. All features are Z-score normalized before being input into the model.

[0026] The scope of protection of this invention is not limited to the examples listed in Embodiment 1, and those skilled in the art can make coding according to the actual situation.

[0027] In step S3, a training ensemble learning classification model is obtained by using fixed-length peptide feature vectors as input features and activity labels as output labels to train an ensemble learning classification model, thus obtaining a trained xanthine oxidase inhibitory peptide prediction model.

[0028] In this application, the ensemble learning classification model is a gradient boosting decision tree model; the gradient boosting decision tree model is a CatBoost model, which uses an ordered boosting strategy to process features and makes predictions through a symmetric tree structure.

[0029] The method also includes: optimizing the CatBoost model parameters through cross-validation to obtain a trained xanthine oxidase inhibitory peptide prediction model; specifically: The training dataset is randomly divided into K mutually exclusive subsets of equal size; Perform K rounds of iterations. In each round of iteration: designate one subset as the validation set, and combine the remaining subsets as the training set for this round; train a model based on the training set for this round using the hyperparameters of the current group of models; evaluate the model using the validation set and record the preset performance evaluation metrics. Calculate the average value of the performance evaluation index in K iterations as the generalization performance estimate of the hyperparameters of the current group; Based on the generalization performance estimation, a better set of hyperparameters is searched using the selected optimization strategy; the performance evaluation index is the area under the receiver operating characteristic curve; the optimization strategy is grid search, random search, or Bayesian optimization. The iterative process is repeated until the hyperparameter set meets the stopping condition, and the obtained optimal hyperparameter set is applied to all the training datasets for final training to obtain the xanthine oxidase inhibitory peptide prediction model.

[0030] In this step, positive and negative samples are randomly shuffled and divided into a training set (for model training and parameter tuning) and an independent test set (for final performance evaluation) in an 8:2 ratio, while maintaining a consistent ratio of positive to negative samples during the division process.

[0031] The CatBoost algorithm was chosen as the classifier. CatBoost is a high-performance gradient boosting decision tree library, characterized by its ability to natively and efficiently handle class features (suitable for certain encoded peptide features), employing an ordered boosting strategy to reduce overfitting, and exhibiting good robustness to imbalanced data.

[0032] Using fixed-length peptide feature vectors as input features and corresponding activity labels as outputs, the CatBoost model is initially trained on the training set.

[0033] To achieve the best generalization performance, a combination of 5-fold cross-validation and grid search is used for hyperparameter optimization.

[0034] Divide the training set into 5 equal parts (K=5). Perform 5 iterations, selecting one part as the validation set in each round and using the remaining 4 parts as the training subset. Train the model on the training subset using the current parameter combination and calculate the AUC value on the validation set. After 5 rounds, calculate the average of the 5 AUC values ​​as the performance evaluation of the parameter combination.

[0035] The search range is set for key hyperparameters (such as learning rate, tree depth, L2 regularization coefficient, etc.). By using a grid search strategy, different parameter combinations are traversed, and the combination with the highest average AUC is selected as the optimal hyperparameter.

[0036] This invention evaluates performance using metrics such as Area Under the Receiver Operating Characteristic (AUC), Accuracy, Recall, and Precision. Simultaneously, XGBoost, LightGBM, and GBDT models are constructed as controls, trained using the same features and data partitioning methods, such as... Figure 2 A schematic diagram illustrating the performance evaluation results of the CatBoost model based on VHSE features is provided. The CatBoost model performs well across multiple metrics.

[0037] Figure 3 ROC curves and AUC comparisons of different models in the prediction of xanthine oxidase (XOD) inhibitory peptides are presented. Under identical training data, VHSE feature encoding, and testing conditions, the CatBoost model preferred in this invention achieved an AUC value of 0.902, which is significantly higher than the comparative models LightGBM (0.879), XGBoost (0.864), and GBDT (0.861).

[0038] In step S4, the set of candidate peptides to be screened is quantized and encoded and then input into the xanthine oxidase inhibitory peptide prediction model to obtain the predicted probability of the inhibitory activity of each candidate peptide. The methods for obtaining the set of candidate peptides to be screened in this application include: Virtual enzymatic digestion of the amino acid sequence of a target protein, wherein the target protein includes sheep milk protein; Proteins are subjected to in vitro enzymatic hydrolysis, and the hydrolysis products are then separated or sequenced. Direct extraction and identification of endogenous peptides from biological samples; A set of peptides of a specific length range obtained by screening from a known peptide database.

[0039] This application describes the virtual hydrolysis of whey protein, αs1-casein, β-casein, etc. in sheep milk using single-enzyme, double-enzyme, or triple-enzyme combinations to obtain a database of candidate peptide fragments for xanthine oxidase.

[0040] After the obtained candidate peptide set is encoded with VHSE features using the method disclosed in step S2, a prediction probability threshold of 0.8 is set. When the prediction probability is greater than or equal to the threshold, the peptide is identified as a high-confidence potential xanthine oxidase (XOD) inhibitory peptide. Figure 4 This is a schematic diagram showing the distribution of predicted peptide probabilities output by the xanthine oxidase inhibitory peptide prediction model proposed in Example 1 of the present invention. The scope of protection of the present invention is not limited to the prediction probability thresholds listed in Example 1, and those skilled in the art can make reasonable selections based on actual circumstances.

[0041] In step S5, based on a preset activity probability threshold, high-confidence xanthine oxidase inhibitory peptide candidate sequences are screened from the candidate peptide set.

[0042] The peptide TNAYF with the highest prediction score was selected for molecular docking analysis.

[0043] Using AutoDockVina software, TNAYF was docked with the xanthine oxidase crystal structure (PDB: 3NVW). Figure 5 This is a schematic diagram of the molecular docking structure of TNAYF binding to xanthine oxidase XOD as proposed in Example 1 of this invention; TNAYF can stably bind to the active site pocket of XOD, and the binding energy is... 9.5 kcal / mol.

[0044] TNAYF was synthesized using the solid-phase peptide synthesis method (Fmoc method) and dissolved in buffer to prepare the required concentration. Xanthine oxidase (XOD) working solution was prepared according to the kit instructions, and blank control, activity control, inhibitor control, and sample group were set up in 96-well plates. After adding xanthine oxidase (XOD) and peptide solution, the plates were incubated at 37°C for 10 min, followed by the addition of the colorimetric system and incubation for another 30 min. The absorbance was measured at 450 nm. Figure 6 This is a schematic diagram of the experimental results of the inhibition of xanthine oxidase (XOD) activity by TNAYF as proposed in Example 1 of the present invention. TNAYF has a significant inhibitory effect on XOD activity.

[0045] This invention not only protects the core screening method and system, but also extends protection to the specific products obtained by the method (TNAYF and its analogues) and their key uses (functional foods and drugs), constructing a complete patent chain of "method-product-use", which greatly enhances the commercial value and market barriers of the technology.

[0046] The scope of protection of this invention is not limited to the examples and values ​​listed in Example 1, and those skilled in the art can make reasonable selections based on the actual situation.

[0047] The method for screening xanthine oxidase inhibitory peptides proposed in Example 1 of this invention can quickly predict a large number of candidate peptides, rapidly obtain high-confidence candidate sequences, and significantly reduce the workload and cost of experimental screening.

[0048] Example 2 Example 2 of this invention proposes a screening device for xanthine oxidase inhibitory peptides. Figure 7 This is a schematic diagram of a screening device for xanthine oxidase inhibitory peptides according to Embodiment 2 of the present invention. The device includes: Memory, used to store computer programs; When a processor executes the computer program, the method steps are as follows: In step S1, a first peptide with xanthine oxidase inhibitory activity is obtained as a positive sample, and a second peptide without xanthine oxidase inhibitory activity is obtained as a negative sample; and both the first and second peptides are labeled with activity.

[0049] In step S2, amino acid physicochemical property descriptors are used to quantize and encode both the first and second peptide segments to generate an initial feature matrix characterizing the physicochemical properties of the peptide segments. The initial feature matrix is ​​then pooled to obtain a fixed-length peptide feature vector.

[0050] In step S3, a training ensemble learning classification model is obtained by using fixed-length peptide feature vectors as input features and activity labels as output labels to train an ensemble learning classification model, thus obtaining a trained xanthine oxidase inhibitory peptide prediction model.

[0051] In step S4, the set of candidate peptides to be screened is quantified and encoded and then input into the xanthine oxidase inhibitory peptide prediction model to obtain the predicted probability of the inhibitory activity of each candidate peptide.

[0052] In step S5, based on a preset activity probability threshold, high-confidence xanthine oxidase inhibitory peptide candidate sequences are screened from the candidate peptide set.

[0053] It should be noted that the present invention also provides an electronic device, including: a communication interface capable of interacting with other devices such as network devices; and a processor connected to the communication interface to enable information interaction with other devices, used to execute a screening method for xanthine oxidase inhibitory peptides provided by one or more of the above technical solutions when running a computer program, wherein the computer program is stored in a memory. In practical applications, the various components of the electronic device are coupled together through a bus system. It is understood that the bus system is used to realize the connection and communication between these components. In addition to a data bus, the bus system also includes a power bus, a control bus, and a status signal bus. The memory in the embodiments of this application is used to store various types of data to support the operation of the electronic device. Examples of this data include any computer program used to operate on the electronic device. It is understood that the memory can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory, flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache.By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM). The memories described in the embodiments of this application are intended to include, but are not limited to, these and any other suitable types of memory. The methods disclosed in the embodiments of this application can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by integrated logic circuits in the processor hardware or by instructions in software form. The processor can be a general-purpose processor, a DSP (Digital Signal Processing, i.e., a chip capable of implementing digital signal processing technology), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules can be located in a storage medium, which is located in memory. The processor reads the program from the memory and, in conjunction with its hardware, completes the steps of the aforementioned method. When the processor executes the program, it implements the corresponding processes in the various methods of the embodiments of this application; for simplicity, these will not be elaborated further here.

[0054] The description of the relevant parts of the screening device for xanthine oxidase inhibitory peptides provided in Embodiment 2 of this application can be found in the detailed description of the corresponding parts of the screening method for xanthine oxidase inhibitory peptides provided in Embodiment 1 of this application, and will not be repeated here.

[0055] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that the elements inherent in a process, method, article, or apparatus that includes a list of elements are included. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Additionally, portions of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of corresponding technical solutions in the prior art have not been described in detail to avoid excessive elaboration.

[0056] While specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art can make other modifications or variations based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for screening xanthine oxidase inhibitory peptides, characterized in that, Includes the following steps: The first peptide with xanthine oxidase inhibitory activity was obtained as a positive sample, and the second peptide without xanthine oxidase inhibitory activity was obtained as a negative sample; both the first and second peptides were labeled with activity. Using amino acid physicochemical property descriptors, both the first and second peptide segments are quantized and encoded to generate an initial feature matrix characterizing the physicochemical properties of the peptide segments. The initial feature matrix is ​​then pooled to obtain a fixed-length peptide segment feature vector. Using fixed-length peptide feature vectors as input features and activity labels as output labels, an ensemble learning classification model is trained to obtain a trained xanthine oxidase inhibitory peptide prediction model. The candidate peptide set to be screened is quantized and encoded and then input into the xanthine oxidase inhibitory peptide prediction model to obtain the predicted probability of inhibitory activity of each candidate peptide; based on the preset activity probability threshold, high-confidence xanthine oxidase inhibitory peptide candidate sequences are screened from the candidate peptide set.

2. The method according to claim 1, characterized in that, The method for determining whether the obtained peptide is active is as follows: When the half-inhibitory concentration of the obtained peptide is... If the peptide is ≤ the first threshold, then the peptide is marked as a positive sample; When the half-inhibitory concentration of the obtained peptide is... If the peptide is ≥ the second threshold, it will be labeled as a negative sample; Wherein, the second threshold is greater than the first threshold.

3. The method according to claim 1, characterized in that, Using amino acid physicochemical property descriptors, both the first and second peptide segments are quantized and encoded to generate an initial feature matrix characterizing the physicochemical properties of the peptide segments, specifically: Based on the amino acid sequence of any peptide segment, extract the amino acid physicochemical property descriptor vector corresponding to each amino acid residue in the sequence in sequence; All extracted descriptor vectors are arranged and combined according to the order of the amino acid sequence to generate an initial feature matrix for any peptide segment.

4. The method according to claim 3, characterized in that, The amino acid physicochemical property descriptor is the VHSE descriptor; The VHSE descriptor characterizes the hydrophobicity, steric hindrance, and electronic effects of amino acids through eight principal components.

5. The method according to claim 1, characterized in that, The initial feature matrix is ​​pooled to obtain a fixed-length peptide feature vector, specifically: The initial feature matrix is ​​input into a preset pooling function; Along the sequence length dimension of the initial feature matrix, the feature values ​​of each physicochemical property dimension are aggregated and calculated; Output a fixed-length peptide feature vector with dimension 1×N; where N is the vector dimension of the amino acid physicochemical property descriptor, and the pooling function includes mean pooling function, max pooling function and summation pooling function.

6. The method according to claim 1, characterized in that, The ensemble learning classification model is a gradient boosting decision tree model. The gradient boosting decision tree model is a CatBoost model, which uses an ordered boosting strategy to process features and makes predictions using a symmetric tree structure.

7. The method according to claim 6, characterized in that, The method further includes: optimizing the CatBoost model parameters through cross-validation to obtain a trained xanthine oxidase inhibitory peptide prediction model; specifically: The training dataset is randomly divided into K mutually exclusive subsets of equal size; Perform K rounds of iterations. In each round of iteration: designate one subset as the validation set, and combine the remaining subsets as the training set for this round; train a model based on the training set for this round using the hyperparameters of the current group of models; evaluate the model using the validation set and record the preset performance evaluation metrics. Calculate the average value of the performance evaluation index in K iterations as the generalization performance estimate of the hyperparameters of the current group; Based on the generalization performance estimate, a better set of hyperparameters is searched using the selected optimization strategy; The iterative process is repeated until the hyperparameter set meets the stopping condition, and the obtained optimal hyperparameter set is applied to all the training datasets for final training to obtain the xanthine oxidase inhibitory peptide prediction model.

8. The method according to claim 7, characterized in that, The performance evaluation metric is the area under the receiver operating characteristic curve; the optimization strategy is grid search, random search, or Bayesian optimization.

9. The method according to claim 1, characterized in that, The methods for obtaining the set of candidate peptides to be screened include: Virtual enzymatic digestion of the amino acid sequence of a target protein, wherein the target protein includes sheep milk protein; Proteins are subjected to in vitro enzymatic hydrolysis, and the hydrolysis products are then separated or sequenced. Direct extraction and identification of endogenous peptides from biological samples; A set of peptides of a specific length range obtained by screening from a known peptide database.

10. A screening device for xanthine oxidase inhibitory peptides, comprising at least one processor and a memory, said memory storing a computer program, characterized in that, When the computer program is executed by the at least one processor, it implements a method for screening xanthine oxidase inhibitory peptides as described in any one of claims 1 to 9.

Images