Intelligent polypeptide synthesis process optimization method, equipment and medium

Abstract

The invention discloses an intelligent polypeptide synthesis process optimization method, equipment and a medium. The method comprises the following steps: 1, loading and preprocessing experimental data of polypeptide synthesis; 2, generating all experimental condition combinations according to experimental data to form a reaction space; 3, converting experimental data into vectorization features; 4, training a Gaussian process regression model based on vectorization features; 5, predicting a potential result corresponding to each group of experimental condition combination in the reaction space according to the Gaussian process regression model; candidate experiment parameters of the next round of experiment are screened out from the potential results according to a Bayesian optimization algorithm; 6, the candidate experiment parameters are applied to an actual experiment, a verification experiment is executed through a microreactor synthesis verification platform, and an experiment result is obtained; and 7, feeding back an experimental result to the Gaussian process regression model, updating model parameters, repeating the steps 5-7 for optimization until an experimental target is met, and outputting an optimal experimental result. According to the invention, efficient optimization of the polypeptide synthesis process can be realized.

Description

Technical Field

[0001] The present invention relates to the field of chemical engineering technology, and particularly to a method, equipment and medium for optimizing an intelligent polypeptide synthesis process. Background Art

[0002] Polypeptide synthesis is an important research direction in the fields of biotechnology and pharmaceuticals, and is widely used in multiple fields such as drug development, vaccine design and biocompatible material preparation. Traditional polypeptide synthesis methods mainly include solid-phase synthesis and liquid-phase synthesis, and the sequence and chemical properties of various amino acids determine the complexity of the synthesis process. During the polypeptide synthesis process, the reaction conditions of different amino acids (such as temperature, pH value, solvent type and reaction time, etc.) may vary significantly. Especially for some amino acids that are difficult to couple or have a low reaction rate, both the reaction efficiency and product purity in the synthesis process may be affected.

[0003] In the actual synthesis process, researchers usually need to optimize the reaction conditions through a large number of experiments, which not only consumes a lot of time and resources, but also may lead to errors and uncertainties in the experimental process. In addition, for the synthesis of polypeptides under complex conditions, there will also be interactions between multiple experimental variables (such as changing the temperature and then affecting variables such as viscosity and flow rate), making it more difficult to determine the optimal experimental parameters. Therefore, how to effectively explore the reaction parameter space with fewer experimental times and quickly find the optimal synthesis conditions is an important challenge in current polypeptide synthesis research.

[0004] In summary, the current defects are as follows: Low experimental efficiency: Traditional optimization methods often require a large number of experiments to determine the optimal reaction conditions, and the experimental design and implementation cycle are long, resulting in waste of resources.

[0005] Difficult to identify parameter interactions: During the polypeptide synthesis process, there are complex interactions between the reaction conditions, that is, among multiple reaction variables, changing one variable will also affect the remaining variables. Traditional experimental methods have not effectively identified this interaction relationship, and it is easy to lead to unsatisfactory experimental results.

[0006] Strong experience dependence: Current optimization strategies often rely on the experience and intuition of researchers, which not only limits the objectivity of the results, but also may lead to the omission of the optimal conditions.

[0007] Poor adaptability: Facing amino acids and polypeptide structures with different properties, existing traditional methods lack flexibility and are difficult to quickly adjust to meet new experimental requirements.

[0008] High trial - and - error cost: In actual synthesis, due to the exploration and trial - and - error of complex condition spaces, repeated experiments are often required, increasing the cost and time investment. Especially in the development of high - value or difficult - to - synthesize polypeptides, the trial - and - error cost is significantly increased.

[0009] In summary, the existing methods for optimizing the polypeptide synthesis process have obvious limitations in terms of efficiency, accuracy, and adaptability. A novel method and device combination are needed to solve these technical bottlenecks. Therefore, it is crucial to select appropriate methods and build models. Summary of the Invention

[0010] In view of this, the purpose of the present invention is to propose a method for optimizing intelligent polypeptide synthesis processes. Using the Bayesian optimization algorithm and a self - built micro - reactor synthesis verification platform, taking advantage of the easy - to - scale - up feature of the micro - reactor synthesis verification platform, through a small number of experiments, the optimal experimental parameters are accurately screened to achieve efficient optimization of the polypeptide synthesis process and provide strong support for large - scale production.

[0011] In order to achieve the above - mentioned technical objectives, the technical solutions adopted by the present invention are as follows: The present invention provides a method for optimizing intelligent polypeptide synthesis processes, including the following steps: Step 1: Load and pre - process the experimental data of polypeptide synthesis; Step 2: Generate all experimental condition combinations based on the experimental data to form a reaction space; Step 3: Convert the experimental data into vectorized features; Step 4: Train a Gaussian process regression model based on the vectorized features; Step 5: Predict the potential results corresponding to each experimental condition combination in the reaction space according to the Gaussian process regression model; Screen out the candidate experimental parameters for the next round of experiments from the potential results according to the Bayesian optimization algorithm; Step 6: Apply the candidate experimental parameters to actual experiments, perform verification experiments through the micro - reactor synthesis verification platform, and obtain experimental results; Step 7: Feed back the experimental results into the Gaussian process regression model, update the model parameters, and repeat steps 5 - 7 for optimization until the experimental objectives are met, and finally output the best experimental results.

[0012] Further, step 1 specifically includes: Step 11: Load the experimental data containing various experimental parameters from the experimental data file. The experimental data includes the substrates and reaction conditions of polypeptide synthesis; Step 12: Use the RDKit library to perform preliminary cleaning, correction, and formatting on the experimental data, and convert the chemical molecular formulas in the substrates into the SMILES format recognizable by the computer.

[0013] Further, step 2 specifically includes: Step 21: Perform operations using the Cartesian product according to different types and parameter ranges of reaction conditions to generate all possible combinations of reaction conditions. Each combination of reaction conditions and the substrate for polypeptide synthesis constitutes a set of experimental condition combinations, and the set of all experimental condition combinations constitutes the reaction space. Step 22: Store all experimental condition combinations as structured data.

[0014] Further, step 3 specifically includes: Step 31: Convert experimental data into vectorized features for use in a machine learning model. The vectorized features include the molecular fingerprint vector corresponding to the substrate and the discretized features corresponding to the reaction conditions. Step 32: Use the Morgan fingerprint algorithm to convert the SMILES structure of the substrate into a fixed-length molecular fingerprint vector. Step 33: Use the KBinsDiscretizer tool to discretize continuous numerical reaction conditions into discrete features.

[0015] Further, step 4 specifically includes: Step 41: Input the molecular fingerprint vector and the discretized features into the GPensemble model, which is a collection of multiple Gaussian process regression models. Step 42: Use multiple Gaussian process regression models to capture different data features and distributions respectively, and fit the relationship between experimental conditions and target results.

[0016] Further, step 5 specifically includes: Step 51: According to the Gaussian process regression model and in combination with the experimental objective, predict the potential results corresponding to each set of experimental condition combinations in the reaction space. Step 52: Through the Bayesian optimization algorithm, screen out the potential result with the largest value from multiple potential results as the target potential result. Step 53: Use the experimental condition combination corresponding to the target potential result as the candidate experimental parameters for the next round of experiments.

[0017] Further, the microreactor synthesis verification platform includes multiple reagent storage bottles, a waste liquid bottle, a multi-channel injection pump, and a microreactor. The multi-channel injection pump is respectively connected to multiple reagent storage bottles and the microreactor, and the waste liquid bottle is connected to the microreactor.

[0018] Further, after step 7, it also includes: Step 8: Organize and store the experimental results and the predicted data of the Gaussian process regression model to generate an experimental report and an optimization log; and save the finally optimized best experimental results as a structured file.

[0019] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the method for optimizing an intelligent polypeptide synthesis process as described above.

[0020] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the method for optimizing an intelligent polypeptide synthesis process as described above.

[0021] Adopting the above technical solution, compared with the prior art, the present invention has the following beneficial effects: 1. Optimization of experiment times and efficiency: By applying the Bayesian optimization algorithm, the present invention can predict and recommend the optimal experimental conditions based on the preliminary experimental data, significantly reducing the required number of experiments and thus improving the experimental efficiency.

[0022] 2. Intelligent experimental design: Through the Bayesian optimization algorithm, the system can automatically explore the optimal combination of experimental parameters in the multi-dimensional experimental space, ensuring that the experimental design is more systematic and intelligent, rather than the empirical adjustment in the traditional method. This advantage makes the experimental design more objective and scientific.

[0023] 3. Combination of the microreactor synthesis verification platform and the Bayesian optimization algorithm: The present invention combines the microreactor synthesis verification platform with the Bayesian optimization algorithm, and uses the high-throughput experimental ability of the microreactor synthesis verification platform to quickly verify under the optimal conditions recommended by the Bayesian optimization algorithm, thus accelerating the optimization iteration speed and greatly improving the efficiency of reaction condition screening.

[0024] 4. Closed-loop optimization process: The system can implement a closed-loop feedback mechanism between the experimental data and the Bayesian optimization algorithm. After each round of experiments, the new experimental results will be fed back into the Gaussian process regression model to further optimize the prediction parameters. Different from the traditional staged experiments, this closed-loop mechanism ensures the continuity and gradual progress of the optimization process, making each experiment provide the maximum information value for subsequent decisions.

[0025] 5. Cost savings: The Bayesian optimization algorithm can be used to reduce the number of experiments and improve the experimental efficiency, and the microreactor synthesis verification platform can reduce the consumption of experimental reagents, effectively reducing the trial-and-error cost. Especially in the development of high-value or difficult-to-synthesize polypeptides, it has obvious economic benefits.

[0026] In summary, through the combination of the Bayesian optimization algorithm and the microreactor synthesis verification platform, the present invention realizes the efficient conversion from experimental data to optimization decisions, greatly improves the efficiency and accuracy of experimental design, provides an innovative solution for the optimization process of polypeptide synthesis, and has significant advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0028] Figure 1 It is a flowchart of the execution of a method for optimizing an intelligent polypeptide synthesis process provided by an embodiment of the present invention.

[0029] Figure 2 It is a Bayesian synthesis optimization logic diagram provided by an embodiment of the present invention.

[0030] Figure 3 It is a schematic diagram of a SMILES format example provided by an embodiment of the present invention.

[0031] Figure 4 It is a schematic diagram of the structure of a microreactor synthesis verification platform provided by an embodiment of the present invention.

[0032] Figure 5 It is a fitting comparison diagram of the actual and predicted experimental results of a model provided by an embodiment of the present invention.

[0033] Figure 6 It is a schematic diagram of an electronic device provided by an embodiment of the present invention.

[0034] Figure 7 It is a schematic diagram of a computer-readable storage medium provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] The following will further describe the present invention in detail with reference to the drawings and embodiments. It should be particularly noted that the following embodiments are only used to illustrate the present invention, but do not limit the scope of the present invention. Similarly, the following embodiments are only partial embodiments of the present invention rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0036] Please refer to Figures 1 - 5 , a method for optimizing an intelligent polypeptide synthesis process of the present invention includes the following steps: Step 1: Load and preprocess the experimental data for peptide synthesis; In this embodiment, step 1 specifically includes: Step 11: Load the experimental data containing various experimental parameters from the experimental data file, supporting multiple data formats (such as CSV, JSON, EXCEL, etc.). The experimental data includes the substrates and reaction conditions for peptide synthesis; the substrates include amino acids, pre-chains, and coupling reagents, and the reaction conditions include temperature (20°C to 70°C), reaction time (2 min to 10 min), flow rate (100 to 400 ul / s), deprotection time, and feeding ratio. The corresponding parameters of the substrates and reaction conditions can also be increased and adjusted according to the specific circumstances of the experiment.

[0037] Step 12: Use the RDKit library (a cheminformatics library for Python, mainly used to process chemical molecule data) to perform preliminary cleaning, calibration, and formatting on the experimental data, and convert the chemical molecular formulas in the substrates into the SMILES format recognizable by the computer. That is, remove the data rows with obvious errors or omissions in the experimental conditions and standardize the data format (such as the data with non-standard SMILES format) to ensure the integrity and standardization of the data for subsequent algorithm input.

[0038] It is necessary to first convert the chemical molecular formula into the SMILES format because the computer cannot recognize ordinary chemical molecular formulas. Therefore, the description of chemical molecules uses the Simplified Molecular Input Line Entry System (SMILES), as Figure 3 shown. SMILES is a standard method for describing the chemical molecular structure through strings, which can accurately describe the atoms, bonds, and stereochemical information of molecules. Its advantages are simplicity, intuitiveness, and wide applicability in molecular calculations and modeling.

[0039] Using cheminformatics libraries (such as RDKit, Open Babel) or online tools (such as PubChem, ChemSpider) can automatically convert the molecular formula into SMILES to ensure accuracy.

[0040] Step 2: Generate all experimental condition combinations based on the experimental data to form a reaction space; In this embodiment, step 2 specifically includes: Step 21: Use the Cartesian product operation according to the different types and parameter ranges of the reaction conditions to generate all possible reaction condition combinations. Each set of reaction condition combinations and the substrates for peptide synthesis form a set of experimental condition combinations, and the set of all experimental condition combinations forms a reaction space; Based on the multi-factor parameter permutation, it is mainly to determine the value ranges of all parameters and enumerate all possible parameter combinations through the Cartesian product (which is an existing mathematical algorithm widely used in experimental design, permutation and combination, and parameter grid search in machine learning).

[0041] Suppose the reaction conditions include three different types: temperature, reaction time, and flow rate. The corresponding parameter ranges are: temperature (20°C to 70°C), reaction time (2 min to 10 min), and flow rate (100 to 400 ul / s). Several specific parameter values can be selected within this range. If there are two temperature parameters a and b, two reaction time parameters c and d, and two flow rate parameters e and f, then there are combinations (a, c, e), (a, c, f), (a, d, e), (a, d, f), (b, c, e), (b, c, f), (b, d, e), (b, d, f). The specific combinations are determined according to the experimental input parameter factors. The more input parameters, the more combination results are obtained.

[0042] Step 22: Store all experimental condition combinations as structured data in the format of CSV or JSON for unified processing, which is conducive to subsequent data processing.

[0043] Step 3: Convert the experimental data into vectorized features; In this embodiment, the specific content of Step 3 includes: Step 31: Convert the experimental data into vectorized features for use in the machine learning model. The vectorized features include the molecular fingerprint vector corresponding to the substrate and the discretized features corresponding to the reaction conditions; Step 32: Use the Morgan fingerprint algorithm to convert the SMILES structure of the substrate into a molecular fingerprint vector with a fixed length. The Morgan fingerprint algorithm is an algorithm for molecular description. It generates a fixed-length vector by calculating the topological features of the molecular structure (such as ring structure, functional groups, atomic connection methods, etc.) and is widely used in molecular similarity analysis and machine learning model training in the field of chemical informatics; Step 33: Use the KBinsDiscretizer tool to discretize the continuous numerical reaction conditions (such as temperature, reaction time, flow rate) into discrete features. KBinsDiscretizer (binning) is a tool for discretizing continuous numerical variables into categorical data. For example, for temperature (20~80) with a step size of 20, it is discretized into (20, 40, 60, 80).

[0044] Step 4: Train a Gaussian process regression model based on the vectorized features; Build the model: Define the kernel function, fit the training data, and optimize the model hyperparameters Training the model: The training process is to pass all the processed data mentioned above through multiple Gaussian process models to capture different features and distributions in the data, and it is trained in combination with the code.

[0045] After training is completed, the Gaussian process regression model will be able to make predictions. For new experimental conditions, the Gaussian process regression model will predict the corresponding target results.

[0046] In this embodiment, step 4 specifically includes: Step 41: Input the molecular fingerprint vector and discretized features into the GPensemble model. The GPensemble model is a collection of multiple Gaussian process regression (GPR) models. The Gaussian process regression model is a machine learning method based on probability theory used to solve regression problems and is particularly suitable for modeling and predicting small sample data sets. The GPensemble model is a collection of Gaussian process regression models and is a method to extend Gaussian processes. By using multiple sub-models to capture different data features and distributions respectively, it can more comprehensively fit the relationship between experimental conditions and target results (such as purity).

[0047] Step 42: Use multiple Gaussian process regression models to capture different data features and distributions respectively and fit the relationship between experimental conditions and target results.

[0048] For example: Using the initial experimental data, train the Gaussian process regression model to fit the relationship between experimental conditions and product performance (purity): Input: Experimental conditions (temperature, reaction time, flow rate, deprotection time, feed ratio) are processed by vectorization to form a training feature matrix.

[0049] Output: Experimental results (purity).

[0050] Using the GPensemble model, the model can predict the performance of un-explored experimental conditions in the reaction space and estimate the uncertainty of the prediction results Step 5: Predict the potential results corresponding to each group of experimental condition combinations in the reaction space according to the Gaussian process regression model; Screen out the candidate experimental parameters for the next round of experiments from the potential results according to the Bayesian optimization algorithm; In this embodiment, step 5 specifically includes: Step 51: According to the Gaussian process regression model and combined with the experimental objective (target purity value), predict the potential results corresponding to each group of experimental condition combinations in the reaction space; Step 52: Use the Bayesian optimization algorithm and select the potential result with the largest value from multiple potential results through the acquisition function as the target potential result; The Bayesian optimization algorithm is an algorithm that gradually collects data and updates the model to optimize the objective function, especially suitable for scenarios of expensive experiments or function evaluations, such as experimental optimization in polypeptide synthesis. The core of Bayesian optimization is to select the next experimental condition to maximize the information gain. In the present invention, a Gaussian process regression model is used as the surrogate model, and the optimal condition is selected through the acquisition function (expected improvement EI), which is a processing and output process of the computer system according to the code.

[0051] Step 53: Use the experimental condition combination corresponding to the target potential result as the candidate experimental parameters for the next round of experiments.

[0052] Step 6: Apply the candidate experimental parameters to the actual experiment, perform a verification experiment through the microreactor synthesis verification platform, and obtain the experimental results; In this embodiment, the microreactor synthesis verification platform includes multiple reagent storage bottles, a waste liquid bottle, a multi-channel injection pump, and a microreactor. The multi-channel injection pump is respectively connected to the multiple reagent storage bottles and the microreactor, and the waste liquid bottle is connected to the microreactor.

[0053] Step 7: Feed the experimental results back into the Gaussian process regression model, update the model parameters, and repeat Steps 5 - 7 for optimization until the experimental goal is met, and finally output the best experimental results. Quickly converge to the best candidate experimental parameters (combinations) by minimizing the number of experiments, and make full use of the existing data for learning and updating. Combine the experimental results, update the model parameters, and re-predict and optimize the experimental conditions for the next round of experiments to achieve the closed-loop optimization of the model-experiment. By repeating the cycle of training the model - predicting - verifying, gradually converge to the optimal parameters for polypeptide synthesis.

[0054] Step 8: Organize and store the experimental results and the prediction data of the Gaussian process regression model, generate an experimental report and an optimization log; and save the finally optimized best experimental results as a structured file (such as EXCEL) for easy analysis and sharing.

[0055] Figure 5 It is the fitting of the actual and predicted results, and the results show that the Gaussian process regression model obviously effectively uses the existing data for training and fits a reasonable prediction function. The Gaussian process regression model not only has high prediction accuracy but also can well capture the trend of the actual data.

[0056] Such as Figure 6As shown in the figure, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned method for optimizing an intelligent polypeptide synthesis process.

[0057] As Figure 7 shown in the figure, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the above-mentioned method for optimizing an intelligent polypeptide synthesis process.

[0058] In addition, in each embodiment of the present invention, each functional unit can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0059] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods in each embodiment of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0060] The above are only partial embodiments of the present invention, and thus do not limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A method for optimizing an intelligent polypeptide synthesis process, characterized in that It includes the following steps: Step 1: Load and preprocess the experimental data for polypeptide synthesis; Step 2: Generate all combinations of experimental conditions based on the experimental data to form a reaction space; Step 3: Convert the experimental data into vectorized features; Step 4: Train a Gaussian process regression model based on the vectorized features; Step 5: Predict the potential results corresponding to each combination of experimental conditions in the reaction space according to the Gaussian process regression model; Screen out the candidate experimental parameters for the next round of experiments from the potential results according to the Bayesian optimization algorithm; Step 6: Apply the candidate experimental parameters to actual experiments, perform verification experiments through a microreactor synthesis verification platform, and obtain experimental results; Step 7: Feed back the experimental results into the Gaussian process regression model, update the model parameters, and repeat Steps 5 - 7 for optimization until the experimental goal is met, and finally output the best experimental results.

2. The method for optimizing an intelligent polypeptide synthesis process according to claim 1, characterized in that, The specific content of Step 1 includes: Step 11: Load the experimental data containing various experimental parameters from the experimental data file, and the experimental data includes the substrates and reaction conditions for polypeptide synthesis; Step 12: Use the RDKit library to perform preliminary cleaning, correction, and formatting processing on the experimental data, and convert the chemical molecular formula in the substrate into the SMILES format recognizable by the computer.

3. The method for optimizing an intelligent polypeptide synthesis process according to claim 2, wherein The specific content of Step 2 includes: Step 21: Perform operations using the Cartesian product according to different types and parameter ranges of reaction conditions to generate all possible combinations of reaction conditions. Each combination of reaction conditions and the substrate for polypeptide synthesis constitutes a combination of experimental conditions, and the set of all combinations of experimental conditions forms a reaction space; Step 22: Store all combinations of experimental conditions as structured data.

4. The method for optimizing an intelligent polypeptide synthesis process according to claim 3, wherein, The specific content of Step 3 includes: Step 31: Convert the experimental data into vectorized features for use in machine learning models. The vectorized features include the molecular fingerprint vector corresponding to the substrate and the discretized features corresponding to the reaction conditions; Step 32: Use the Morgan fingerprint algorithm to convert the SMILES structure of the substrate into a molecular fingerprint vector with a fixed length; Step 33: Use the KBinsDiscretizer tool to discretize the continuous numerical reaction conditions and convert them into discrete features.

5. A method for optimizing an intelligent polypeptide synthesis process according to claim 1, characterized in that, The specific content of Step 4 includes: Step 41: Input the molecular fingerprint vector and discretized features into the GPensemble model, and the GPensemble model is a collection of multiple Gaussian process regression models; Step 42: Capture different data features and distributions through multiple Gaussian process regression models respectively, and fit the relationship between experimental conditions and target results.

6. The method for optimizing an intelligent polypeptide synthesis process according to claim 1, wherein The specific content of Step 5 includes: Step 51: According to the Gaussian process regression model and in combination with the experimental goal, predict the potential results corresponding to each combination of experimental conditions in the reaction space; Step 52: Through the Bayesian optimization algorithm, screen out the potential result with the largest value from multiple potential results as the target potential result; Step 53: Use the combination of experimental conditions corresponding to the target potential result as the candidate experimental parameters for the next round of experiments.

7. The method for optimizing an intelligent polypeptide synthesis process according to claim 1, wherein The microreactor synthesis verification platform includes multiple reagent storage bottles, a waste liquid bottle, a multi-channel injection pump, and a microreactor. The multi-channel injection pump is respectively connected to the multiple reagent storage bottles and the microreactor, and the waste liquid bottle is connected to the microreactor.

8. The method for optimizing an intelligent polypeptide synthesis process according to claim 1, wherein After the step 7, the following steps are further included: Step 8: Organize and store the experimental results and the predicted data of the Gaussian process regression model to generate an experimental report and an optimization log; and save the finally optimized best experimental result as a structured file.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it realizes a method for optimizing an intelligent polypeptide synthesis process according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it realizes a method for optimizing an intelligent polypeptide synthesis process according to any one of claims 1 to 8.

Images