Intelligent control method and system for directional enzymolysis of protein peptide with specific molecular weight distribution
By constructing a process prediction model based on temporal convolutional networks and long short-term memory networks, real-time and precise control of the enzymatic hydrolysis process was achieved, solving the problem that it is difficult to accurately prepare protein peptides of specific molecular weights in the traditional enzymatic hydrolysis process, and improving production efficiency and product stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HAINAN YUTIDE BIOPHARMACEUTICAL CO LTD
- Filing Date
- 2026-01-15
- Publication Date
- 2026-05-01
AI Technical Summary
Traditional enzymatic hydrolysis processes cannot be monitored and dynamically adjusted in real time, making it difficult to accurately prepare protein peptides within specific molecular weight ranges. They also suffer from poor batch-to-batch stability, low efficiency, and high cost.
A process prediction molecular weight distribution model is constructed using temporal convolutional networks and long short-term memory networks. Through real-time prediction and feedback adjustment, precise control of the enzymatic hydrolysis process is achieved, including data acquisition, model construction, standardization processing, control command generation, and adjustment feedback.
It enables real-time and precise control of the enzymatic hydrolysis process, improves prediction accuracy, and stably and efficiently prepares target protein peptides with molecular weight distributions that meet specific requirements, reducing reliance on operator experience and material waste.
Smart Images
Figure CN121963835A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of bioengineering technology, and in particular to a method and system for the intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution. Background Technology
[0002] Enzymatic hydrolysis is the mainstream process for preparing protein peptides, including bioactive peptides and nutritional peptides. The molecular weight distribution of the product is a key indicator determining its functional properties and bioavailability. Traditional enzymatic hydrolysis processes typically rely on fixed parameters such as time, temperature, pH, and enzyme dosage, evaluating the product through endpoint sampling and testing. This method has significant drawbacks: firstly, it cannot monitor and dynamically adjust the enzymatic hydrolysis process in real time, resulting in poor batch-to-batch stability; secondly, when the target product is a peptide within a specific molecular weight range (such as antioxidant peptides concentrated in the 1-3 kDa range), traditional methods struggle to accurately target the product, often requiring numerous repeated experiments to determine the optimal conditions, leading to low efficiency and high cost.
[0003] In related technologies, some improvements have been made to the control of enzymatic hydrolysis processes. For example, the end point of enzymatic hydrolysis is controlled by monitoring the degree of hydrolysis, but the degree of hydrolysis is a macroscopic average indicator and cannot accurately reflect the complex molecular weight distribution. pH-stat methods or temperature gradient control are used, but these are still open-loop controls based on a single variable, which cannot cope with the complex process of enzymatic hydrolysis, which is multivariate, nonlinear, and time-varying, nor can they be directly regulated with the final molecular weight distribution as the target. Mathematical models are used to predict the enzymatic hydrolysis process, but these are mostly limited to offline simulation or feedforward control, lacking feedback regulation closed loops based on real-time prediction, and the model's generalization ability is limited, making it difficult to adapt to actual production conditions such as raw material fluctuations. Summary of the Invention
[0004] The purpose of this invention is to provide a method and system for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution, so as to achieve real-time and precise control of the enzymatic hydrolysis process, and to stably and efficiently prepare target protein peptides with molecular weight distribution that meet specific requirements.
[0005] To achieve the above objectives, the present invention provides a method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution, comprising the following steps: S1. Obtain historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data; S2. Based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, construct a process prediction molecular weight distribution model according to temporal convolutional network and long short-term memory network. S3. Obtain the enzymatic hydrolysis process data and specific molecular weight distribution of the target protein peptide, and standardize the enzymatic hydrolysis process data of the target protein peptide to obtain standardized enzymatic hydrolysis process data of the target protein peptide. S4. Input the standardized enzymatic hydrolysis process data of the target protein peptide into the process prediction molecular weight distribution model to obtain the predicted molecular weight distribution of the target protein peptide. S5. Input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model, and obtain the enzymatic hydrolysis process control instructions by comparing it with the specific molecular weight prediction distribution of the target protein peptide. S6. Adjust the enzymatic hydrolysis process data according to the control instructions and feed it back to the process prediction molecular weight distribution model.
[0006] Preferably, the historical enzymatic hydrolysis process data includes the amount of protease added, stirring rate, hydrolysis pH, hydrolysis temperature, and hydrolysis time.
[0007] Preferably, the specific content of constructing the process prediction molecular weight distribution model based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data in S2, according to temporal convolutional networks and long short-term memory networks, includes: Historical enzymatic hydrolysis process data were time-series aligned, missing values were removed, and standardized to obtain a historical standardized enzymatic hydrolysis process parameter sequence. By associating historical standardized enzymatic hydrolysis process parameter sequences with corresponding historical final product molecular weight distribution data, training and test sets are constructed. Based on the training and testing sets, using historical standardized enzymatic hydrolysis process parameter sequences as input and historical final product molecular weight distribution data as output, a temporal convolutional network and a long short-term memory network are trained. By minimizing the loss function between the predicted and actual molecular weight distributions of protein peptides, a process-predicted molecular weight distribution model is obtained.
[0008] Preferably, the expression for the loss function is: ; in, For loss function, This represents the actual molecular weight distribution of protein peptides. To predict the molecular weight distribution of protein peptides, The actual molecular weight distribution of protein peptides With the predicted molecular weight distribution of protein peptides cosine similarity, The actual molecular weight distribution of protein peptides With the predicted molecular weight distribution of protein peptides The mean square error, This represents the peak value of the actual protein peptide molecular weight distribution. To predict the peak molecular weight distribution of protein peptides, Peak value of actual protein peptide molecular weight distribution Compared with the predicted peak molecular weight distribution of protein peptides The average error, , and All are weighting coefficients.
[0009] Preferably, the standardization process in S3 uses the Z-score standardization algorithm, and the calculation formula is as follows: ; in, This refers to the enzymatic hydrolysis process data for the target protein peptides. This represents the average of historical enzymatic hydrolysis process data. The standard deviation of historical enzymatic hydrolysis process data. Standardize the enzymatic hydrolysis process data for the target protein peptide.
[0010] Preferably, in step S5, the specific molecular weight distribution of the target protein peptide is input into the process prediction molecular weight distribution model. By comparing it with the predicted molecular weight distribution of the target protein peptide, the specific content of the enzymatic hydrolysis process control instruction is obtained, including: Input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model; The cosine similarity algorithm is used to calculate the similarity between the specific molecular weight distribution of the target protein peptide and the predicted specific molecular weight distribution of the target protein peptide; Based on the similarity, enzymatic hydrolysis process control instructions are generated; the enzymatic hydrolysis process control instructions include: adjusting the amount of protease added, stirring rate, enzymatic hydrolysis pH, enzymatic hydrolysis temperature, and enzymatic hydrolysis time.
[0011] Preferably, the specific content of the enzymatic hydrolysis process control instructions generated based on similarity includes: When the similarity is greater than or equal to 95%, the current enzymatic hydrolysis process is determined to meet the requirements, and the control instruction is to maintain the existing process parameters. When the similarity is less than 95%, the distribution difference characteristic parameters are calculated, and the process adjustment instructions are generated based on the difference characteristic parameters; the distribution difference characteristic parameters include peak offset, half-peak width deviation and cumulative distribution area difference.
[0012] Preferably, the specific content of the process adjustment instruction generated based on the difference characteristic parameters includes: If the peak offset is greater than 50 Da, an instruction is generated to adjust the amount of protease added, with an adjustment range of ±5%-15%, and to fine-tune the enzymatic hydrolysis temperature by ±1℃-3℃. If the half-peak width deviation is greater than 20%, an instruction will be generated to adjust the stirring speed by ±10r / min-30r / min and the pH value by ±0.1-0.3. If the cumulative distribution area difference is greater than 10%, an instruction to adjust the enzymatic hydrolysis time by ±10min-60min will be generated.
[0013] This invention provides a smart control system for the targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution, used to implement the aforementioned smart control method for the targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution, comprising: The first data acquisition module is used to acquire historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data. The model building module is used to construct a process prediction molecular weight distribution model based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, using temporal convolutional networks and long short-term memory networks. The second data acquisition module is used to acquire the enzymatic hydrolysis process data of the target protein peptide and the specific molecular weight distribution of the target protein peptide, and to standardize the enzymatic hydrolysis process data of the target protein peptide to obtain standardized enzymatic hydrolysis process data of the target protein peptide. The distribution prediction module is used to input the standardized enzymatic hydrolysis process data of the target protein peptide into the process prediction molecular weight distribution model to obtain the predicted distribution of the specific molecular weight of the target protein peptide. The control instruction generation module is used to input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model, and obtain the enzymatic hydrolysis process control instructions by comparing it with the specific molecular weight prediction distribution of the target protein peptide. The adjustment feedback module is used to adjust the enzymatic hydrolysis process data according to control commands and feed it back to the process prediction molecular weight distribution model.
[0014] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor, when calling the computer program in the memory, implements the above-described intelligent control method for the directional enzymatic hydrolysis of protein peptides with specific molecular weight distribution.
[0015] In summary, the present invention provides a method and system for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution. Compared with traditional technologies, the advantages are as follows: The process prediction molecular weight distribution model constructed by long short-term memory network and temporal convolutional network can accurately capture the nonlinear temporal relationship between enzymatic hydrolysis process parameters and molecular weight distribution, realizing closed-loop control of the enzymatic hydrolysis process with predicted molecular weight distribution as the direct control target. This not only achieves real-time and precise control of the enzymatic hydrolysis process, but also improves the prediction accuracy, enabling the stable and efficient preparation of target protein peptides with molecular weight distribution that meets specific requirements.
[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating a method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution. Figure 2This is a schematic diagram of a module of an intelligent control system for the directional enzymatic hydrolysis of protein peptides with specific molecular weight distribution. Detailed Implementation
[0018] The technical method of the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values of the components and steps described in these embodiments do not limit the scope of this application.
[0019] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.
[0020] Techniques, systems, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the instruction manual.
[0021] In all the examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.
[0022] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.
[0023] This invention provides a method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distributions, such as... Figure 1 As shown, it includes: Step S1: Obtain historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data. The historical enzymatic hydrolysis process data includes the amount of protease added, stirring rate, hydrolysis pH, hydrolysis temperature, and hydrolysis time; the final product molecular weight distribution data is obtained through high-performance liquid chromatography or mass spectrometry analysis.
[0024] Step S2: Based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, construct a process prediction molecular weight distribution model using temporal convolutional networks and long short-term memory networks.
[0025] Furthermore, the specific content of step S2 includes: Step S201: Perform time series alignment, missing value processing, and standardization on the historical enzymatic hydrolysis process data to obtain a historical standardized enzymatic hydrolysis process parameter sequence.
[0026] Step S202: Associate the historical standardized enzymatic hydrolysis process parameter sequence with the corresponding historical final product molecular weight distribution data to construct training and testing sets.
[0027] Step S203: Based on the training and testing sets, using historical standardized enzymatic hydrolysis process parameter sequences as input and historical final product molecular weight distribution data as output, train a temporal convolutional network and a long short-term memory network. By minimizing the loss function between the predicted and actual protein peptide molecular weight distributions, a process-predicted molecular weight distribution model is obtained. The above training process terminates when the loss function converges. This loss function must simultaneously measure the similarity of distribution shape and numerical deviation, and its expression is: ; in, For loss function, This represents the actual molecular weight distribution of protein peptides. To predict the molecular weight distribution of protein peptides, The actual molecular weight distribution of protein peptides With the predicted molecular weight distribution of protein peptides Cosine similarity (used to measure the overall shape matching of the distribution). The actual molecular weight distribution of protein peptides With the predicted molecular weight distribution of protein peptides The mean square error (used to measure the overall deviation of the distributed values). This represents the peak value of the actual protein peptide molecular weight distribution. To predict the peak molecular weight distribution of protein peptides, Peak value of actual protein peptide molecular weight distribution Compared with the predicted peak molecular weight distribution of protein peptides The average error (used to focus on the accuracy of the core feature of molecular weight distribution, the "peak") , and All are weighting coefficients, and the weighting coefficients satisfy... .
[0028] Long Short-Term Memory Networks (LSTMs) are networks consisting of embedding layers, single or multiple LSTM units, dropout layers, and fully connected output layers. The output layer uses the Softmax activation function. LSTMs utilize their gating mechanisms (input gate, forget gate, output gate) to effectively capture long-term dependencies in process parameter sequences, such as the persistent impact of early enzyme addition strategies on the distribution of later products.
[0029] Temporal Convolutional Networks (TCNs) are deep networks consisting of causal dilated convolutional layers, weight normalization layers, modified linear unit activation functions, and residual connections. Their final output layer is a fully connected layer using the Softmax activation function to ensure the output represents the probability distribution across molecular weight ranges. TCNs utilize stacked causal dilated convolutional layers, enabling parallel processing of long sequences, exhibiting more stable gradients, and excelling at capturing local patterns and multi-level trends within sequences.
[0030] This invention employs a weighted fusion of three sub-models—long short-term memory network and temporal convolutional network—to construct a prediction model. Combined with a loss function (which simultaneously constrains distribution similarity, overall numerical bias, and peak accuracy), the prediction accuracy is improved compared to a single model. This model can accurately capture the nonlinear temporal relationship between enzymatic hydrolysis process parameters and molecular weight distribution.
[0031] Step S3: Obtain the enzymatic hydrolysis process data and specific molecular weight distribution of the target protein peptide, and standardize the enzymatic hydrolysis process data to obtain standardized enzymatic hydrolysis process data. The standardization process uses the Z-score normalization algorithm, and the calculation formula is as follows: ; in, This refers to the enzymatic hydrolysis process data for the target protein peptides. This represents the average of historical enzymatic hydrolysis process data. The standard deviation of historical enzymatic hydrolysis process data. Standardize the enzymatic hydrolysis process data for the target protein peptide.
[0032] Step S4: Input the standardized enzymatic hydrolysis process data of the target protein peptide into the process prediction molecular weight distribution model to obtain the predicted molecular weight distribution of the target protein peptide. The predicted molecular weight distribution trend of the protein peptide is output as a curve showing the percentage of peptide content changing over time in a specific molecular weight range (e.g., <1kDa, 1kDa-3kDa, 3kDa-5kDa, and >5kDa).
[0033] Step S5: Input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model, and obtain the enzymatic hydrolysis process control instructions by comparing it with the predicted molecular weight distribution of the target protein peptide.
[0034] Step S6: Adjust the enzymatic hydrolysis process data according to the control instructions and feed it back to the process prediction molecular weight distribution model.
[0035] Furthermore, the specific content of step S5 includes: Step S501: Input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model.
[0036] Step S502: Calculate the similarity between the specific molecular weight distribution of the target protein peptide and the predicted specific molecular weight distribution of the target protein peptide using the cosine similarity algorithm.
[0037] Step S503: Generate enzymatic hydrolysis process control instructions based on the similarity. These instructions include adjusting the amount of protease added, stirring rate, hydrolysis pH, hydrolysis temperature, and hydrolysis time. Specifically, when the similarity is greater than or equal to 95%, the current enzymatic hydrolysis process is deemed to meet the requirements, and the control instruction is to maintain the existing process parameters. When the similarity is less than 95%, distribution difference characteristic parameters are calculated, and process adjustment instructions are generated based on these parameters. These distribution difference characteristic parameters include peak offset, half-width deviation, and cumulative distribution area difference.
[0038] It is understandable that the specific content of the process adjustment instructions generated based on the difference characteristic parameters includes: If the peak offset is greater than 50 Da, an instruction is generated to adjust the amount of protease added, with an adjustment range of ±5%-15%, and to fine-tune the enzymatic hydrolysis temperature by ±1℃-3℃.
[0039] If the half-peak width deviation is greater than 20%, an instruction will be generated to adjust the stirring speed by ±10r / min-30r / min and the pH value by ±0.1-0.3.
[0040] If the cumulative distribution area difference is greater than 10%, an instruction to adjust the enzymatic hydrolysis time by ±10min-60min will be generated.
[0041] This invention employs a closed-loop control mechanism with a single closed-loop control cycle of ≤30min. It eliminates the need to wait for long-term feedback from offline detection, allowing for real-time adjustment of process parameters. This prevents the enzymatic hydrolysis process from deviating from the optimal range, reduces raw material waste, and eliminates the need for manual intervention throughout the entire process. It achieves intelligent control of the entire process, from process parameter input and molecular weight distribution prediction to process adjustment, reducing reliance on operator experience and improving product quality stability.
[0042] This invention provides an intelligent control system for the targeted enzymatic hydrolysis of protein peptides with specific molecular weight distributions, such as... Figure 2 As shown, the method for intelligent control of targeted enzymatic hydrolysis of protein peptides with a specific molecular weight distribution, as described above, includes: The first data acquisition module is used to acquire historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data.
[0043] The model building module is used to construct a process prediction molecular weight distribution model based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, using temporal convolutional networks and long short-term memory networks.
[0044] The second data acquisition module is used to acquire the enzymatic hydrolysis process data of the target protein peptide and the specific molecular weight distribution of the target protein peptide, and to standardize the enzymatic hydrolysis process data of the target protein peptide to obtain standardized enzymatic hydrolysis process data of the target protein peptide.
[0045] The distribution prediction module is used to input the standardized enzymatic hydrolysis process data of the target protein peptide into the process prediction molecular weight distribution model to obtain the predicted distribution of the specific molecular weight of the target protein peptide.
[0046] The control command generation module is used to input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model, and obtain the enzymatic hydrolysis process control command by comparing it with the specific molecular weight prediction distribution of the target protein peptide.
[0047] The adjustment feedback module is used to adjust the enzymatic hydrolysis process data according to control commands and feed it back to the process prediction molecular weight distribution model.
[0048] The present invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor calls the computer program in the memory, it implements the above-mentioned intelligent control method for the directional enzymatic hydrolysis of protein peptides with a specific molecular weight distribution.
[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution, characterized in that, Includes the following steps: S1. Obtain historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data; S2. Based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, construct a process prediction molecular weight distribution model according to temporal convolutional network and long short-term memory network. S3. Obtain the enzymatic hydrolysis process data and specific molecular weight distribution of the target protein peptide, and standardize the enzymatic hydrolysis process data of the target protein peptide to obtain standardized enzymatic hydrolysis process data of the target protein peptide. S4. Input the standardized enzymatic hydrolysis process data of the target protein peptide into the process prediction molecular weight distribution model to obtain the predicted molecular weight distribution of the target protein peptide. S5. Input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model, and obtain the enzymatic hydrolysis process control instructions by comparing it with the specific molecular weight prediction distribution of the target protein peptide. S6. Adjust the enzymatic hydrolysis process data according to the control instructions and feed it back to the process prediction molecular weight distribution model.
2. The method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 1, characterized in that, The historical enzymatic hydrolysis process data includes the amount of protease added, stirring rate, hydrolysis pH, hydrolysis temperature, and hydrolysis time.
3. The method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 1, characterized in that, The specific content of S2, which constructs a process prediction molecular weight distribution model based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, using temporal convolutional networks and long short-term memory networks, includes: Historical enzymatic hydrolysis process data were time-series aligned, missing values were removed, and standardized to obtain a historical standardized enzymatic hydrolysis process parameter sequence. By associating historical standardized enzymatic hydrolysis process parameter sequences with corresponding historical final product molecular weight distribution data, training and test sets are constructed. Based on the training and testing sets, using historical standardized enzymatic hydrolysis process parameter sequences as input and historical final product molecular weight distribution data as output, a temporal convolutional network and a long short-term memory network are trained. By minimizing the loss function between the predicted and actual molecular weight distributions of protein peptides, a process-predicted molecular weight distribution model is obtained.
4. The intelligent control method for targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 2, characterized in that, The expression for the loss function is: ； in, For loss function, This represents the actual molecular weight distribution of protein peptides. To predict the molecular weight distribution of protein peptides, The actual molecular weight distribution of protein peptides With the predicted molecular weight distribution of protein peptides cosine similarity, The actual molecular weight distribution of protein peptides With the predicted molecular weight distribution of protein peptides The mean square error, This represents the peak value of the actual protein peptide molecular weight distribution. To predict the peak molecular weight distribution of protein peptides, Peak value of actual protein peptide molecular weight distribution Compared with the predicted peak molecular weight distribution of protein peptides The average error, , and All are weighting coefficients.
5. The intelligent control method for directional enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 1, characterized in that, The standardization process described in S3 uses the Z-score standardization algorithm, and the calculation formula is as follows: ； in, This refers to the enzymatic hydrolysis process data for the target protein peptides. This represents the average of historical enzymatic hydrolysis process data. The standard deviation of historical enzymatic hydrolysis process data. Standardize the enzymatic hydrolysis process data for the target protein peptide.
6. The method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 1, characterized in that, In S5, the specific molecular weight distribution of the target protein peptide is input into the process prediction molecular weight distribution model. By comparing it with the predicted molecular weight distribution of the target protein peptide, the specific content of the enzymatic hydrolysis process control instructions is obtained, including: Input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model; The cosine similarity algorithm is used to calculate the similarity between the specific molecular weight distribution of the target protein peptide and the predicted specific molecular weight distribution of the target protein peptide; Based on the similarity, enzymatic hydrolysis process control instructions are generated; the enzymatic hydrolysis process control instructions include: adjusting the amount of protease added, stirring rate, enzymatic hydrolysis pH, enzymatic hydrolysis temperature, and enzymatic hydrolysis time.
7. The method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 6, characterized in that, Based on similarity, the specific content of the generated enzymatic hydrolysis process control instructions includes: When the similarity is greater than or equal to 95%, the current enzymatic hydrolysis process is determined to meet the requirements, and the control instruction is to maintain the existing process parameters. When the similarity is less than 95%, the distribution difference characteristic parameters are calculated, and the process adjustment instructions are generated based on the difference characteristic parameters; the distribution difference characteristic parameters include peak offset, half-peak width deviation and cumulative distribution area difference.
8. The method for intelligent control of targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution according to claim 7, characterized in that, The specific content of the process adjustment instructions generated based on the differential characteristic parameters includes: If the peak offset is greater than 50 Da, an instruction is generated to adjust the amount of protease added, with an adjustment range of ±5%-15%, and to fine-tune the enzymatic hydrolysis temperature by ±1℃-3℃. If the half-peak width deviation is greater than 20%, an instruction will be generated to adjust the stirring speed by ±10r / min-30r / min and the pH value by ±0.1-0.
3. If the cumulative distribution area difference is greater than 10%, an instruction to adjust the enzymatic hydrolysis time by ±10min-60min will be generated.
9. A smart control system for the targeted enzymatic hydrolysis of protein peptides with specific molecular weight distribution, characterized in that, A method for implementing the intelligent control method for the directional enzymatic hydrolysis of protein peptides with a specific molecular weight distribution as described in any one of claims 1-8 includes: The first data acquisition module is used to acquire historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data. The model building module is used to construct a process prediction molecular weight distribution model based on historical enzymatic hydrolysis process data and corresponding historical final product molecular weight distribution data, using temporal convolutional networks and long short-term memory networks. The second data acquisition module is used to acquire the enzymatic hydrolysis process data of the target protein peptide and the specific molecular weight distribution of the target protein peptide, and to standardize the enzymatic hydrolysis process data of the target protein peptide to obtain standardized enzymatic hydrolysis process data of the target protein peptide. The distribution prediction module is used to input the standardized enzymatic hydrolysis process data of the target protein peptide into the process prediction molecular weight distribution model to obtain the predicted distribution of the specific molecular weight of the target protein peptide. The control instruction generation module is used to input the specific molecular weight distribution of the target protein peptide into the process prediction molecular weight distribution model, and obtain the enzymatic hydrolysis process control instructions by comparing it with the specific molecular weight prediction distribution of the target protein peptide. The adjustment feedback module is used to adjust the enzymatic hydrolysis process data according to control commands and feed it back to the process prediction molecular weight distribution model.
10. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor, when calling the computer program in the memory, implements the content of the intelligent control method for directional enzymatic hydrolysis of protein peptides with specific molecular weight distribution as described in any one of claims 1 to 8.