Antibody expression and purification integration method based on algae cells

Abstract

The invention provides an antibody expression and purification integration method based on algae cells, and belongs to the technical field of computational biology and synthetic biology, the method comprises the following steps: selecting a proper algae cell strain through morphological analysis, culturing in a standard culture medium, and dynamically adjusting culture conditions by adopting a culture parameter optimization equation. Antibody expression is carried out by adding a protein expression inducer, then cells are collected, and optimal disruption conditions are determined by using a cell lysis optimization equation. The cracked sample is filtered and centrifugally separated, purification parameters are determined by adopting a chromatographic purification optimization equation, and affinity chromatographic purification is carried out. And finally, selecting a proper concentration condition by using a concentration optimization equation, and comprehensively evaluating the product through a quality judgment equation. In the whole process, through the synergistic effect of a plurality of optimization equations, the technical problems of low antibody yield, poor purity and low activity caused by uncoordinated parameter optimization in the algae cell antibody expression and purification process in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of computational biology and synthetic biology, and particularly relates to an antibody expression and purification integrated method based on algal cells. BACKGROUND

[0002] Antibodies are important therapeutic and diagnostic molecules in the field of biological medicine, and their production mainly relies on mammalian cell expression systems. In recent years, microalgae expression systems represented by Chlamydomonas reinhardtii have gradually become a new platform for antibody production due to their rapid growth, low culture cost, and complete post-translational modification of proteins. Traditional antibody expression and purification techniques mainly include cell culture, protein induction expression, cell disruption, chromatographic separation, and concentration steps. During the cell culture stage, the temperature, light, and pH value are usually controlled using empirical parameters; during the protein expression stage, a fixed concentration of inducer is mainly relied on; during the cell disruption process, the ultrasonic power and time are set based on empirical values; and during the chromatographic purification process, the binding time, elution conditions, and other parameters are often standardized.

[0003] However, this traditional parameter setting method has obvious shortcomings. First, the cell culture conditions cannot be dynamically adjusted according to the real-time growth state of the cells, resulting in uneven cell growth and large fluctuations in yield. Second, during the protein expression induction process, a fixed inducer concentration cannot adapt to the physiological state of different batches of cells, causing unstable expression efficiency. Third, the cell disruption parameters lack correlation with cell density, viability, and other factors, resulting in uneven disruption effects. Finally, the parameters in the chromatographic purification and concentration processes are independent of each other, and the influence of upstream and downstream processes is not considered, making it difficult to ensure the quality of the final product.

[0004] The root cause of these technical defects lies in the lack of a systematic parameter optimization method, which cannot achieve coordinated optimization of culture, expression, disruption, purification, and other aspects. In particular, during process scaling, the coupling relationship between parameters is more complex, and the traditional empirical parameter setting method cannot guarantee the uniformity and stability of the product. Therefore, how to establish an antibody expression and purification integrated method that can achieve coordinated optimization of all parameters in the whole process has become a key technical problem to be solved in this field. That is, the existing technology has the technical problem that the uncoordinated optimization of various parameters in the antibody expression and purification process of algal cells leads to low antibody yield, poor purity, and low activity. SUMMARY

[0005] Therefore, the present application provides an antibody expression and purification integrated method based on algal cells, which can solve the technical problem that the uncoordinated optimization of various parameters in the antibody expression and purification process of algal cells leads to low antibody yield, poor purity, and low activity in the prior art.

[0006] The application is achieved in that the application provides an antibody expression and purification integrated method based on algal cells, which comprises the following steps: performing morphological analysis on algal cell strains, and selecting algal cell strains with cell morphological parameters meeting conditions; inoculating the algal cell strains into a standard culture medium for culture; dynamically adjusting culture conditions by using a culture parameter optimization equation; adding a protein expression inducer for continuous culture; collecting the algal cells and resuspending them in a lysis buffer; determining ultrasonic lysis parameters by using a cell lysis optimization equation; filtering the lysis solution and collecting the supernatant by centrifugation; determining affinity chromatography parameters by using a chromatography purification optimization equation; transferring the incubation mixture into a chromatography column for purification; determining concentration parameters by using a concentration optimization equation; and performing quality evaluation on the purified product by using a quality determination equation.

[0007] The cell morphological parameters meet conditions, specifically, the cell sphericity is greater than 0.95, the cell diameter is 2-5 microns, the cell density is 1,000,000-2,000,000 per milliliter, and the growth cycle is 48-72 hours.

[0008] The standard culture medium has an initial temperature of 22-25°C, an initial hydrogen ion concentration value of 7.0-7.5, and an initial light intensity of 2,000-3,000 lux, and the culture is performed for 48-72 hours to record a cell growth curve.

[0009] The culture parameter optimization equation takes a cell growth rate vector, a cell photosynthetic rate vector, a cell metabolic rate vector, and a cell density vector as input parameters, and outputs optimal culture temperature, optimal light intensity, and optimal hydrogen ion concentration value.

[0010] The protein expression inducer has a concentration of 0.2-0.5 mg / mL, and the culture is continued for 24-36 hours to determine the cell viability.

[0011] The algal cells are collected by centrifugation at a speed of 5,000-6,000 rpm, washed with phosphate buffer three times, and resuspended in a cell lysis buffer.

[0012] The cell lysis optimization equation takes the cell density, the cell viability, lysis temperature, buffer ion strength, and the cell sphericity as input parameters, and outputs optimal ultrasonic power and optimal lysis time.

[0013] The lysis solution is filtered through a 0.45-micron filter, the filtrate is collected and separated by centrifugation at a speed of 10,000-12,000 rpm for 30-40 minutes, and the turbidity of the supernatant is determined.

[0014] The chromatography purification optimization equation takes the supernatant turbidity, temperature, hydrogen ion concentration value and protein concentration vector as input parameters, and the output parameters include optimal binding time, optimal binding temperature, optimal elution gradient and optimal flow rate.

[0015] The quality determination equation is used to evaluate the quality of the purified product, specifically to calculate a quality comprehensive score greater than 85 points and simultaneously satisfy the purity index greater than 95%, the activity retention rate greater than 90%, and the protein concentration greater than 1 mg / mL, to determine the product as qualified.

[0016] The concentration optimization equation takes the protein concentration, elution volume, target concentration multiple, protein isoelectric point and protein molecular weight as input parameters, and the output parameters include optimal membrane molecular weight cut-off and optimal concentration pressure; the quality determination equation takes the breakage efficiency, the supernatant turbidity, the binding rate, the elution curve peak shape parameter, the concentration multiple, the purity index, and the activity retention rate as input parameters; the quality determination equation calculates a quality comprehensive score through a coupling relationship between the input parameters, which includes a linear coupling term of the breakage efficiency and the binding rate, an exponential coupling term of the elution curve peak shape parameter and the purity index, and a Gaussian coupling term of the concentration multiple and the activity retention rate.

[0017] Further, the culture parameter optimization equation adopts a multi-layer perceptron neural network structure, the input layer includes 192 neurons, the hidden layer adopts a 3-layer structure, each layer includes 128, 64 and 32 neurons respectively, and the output layer includes 3 neurons; the network training adopts a back propagation algorithm, the learning rate is set to 0.001, the batch size is 32, and the training rounds are 1000 rounds.

[0018] Further, the chromatography purification optimization equation adopts a Box-Behnken design method, uses a least squares method to estimate model parameters, and filters significant variables through a stepwise regression method.

[0019] Further, the concentration optimization equation adopts an artificial neural network model, the hidden layer adopts a double hidden layer structure, uses a Bayesian regularization method to prevent overfitting, and the model optimization adopts a genetic algorithm.

[0020] Further, the quality determination equation adopts a fuzzy neural network model, combines fuzzy logic reasoning with neural network learning, the model training adopts a hybrid learning algorithm, and the model stability is evaluated through repeated sampling.

[0021] Compared with the prior art, the application provides an antibody expression and purification integrated method based on algal cells, and the application provides an algal cell antibody expression and purification integrated method based on multiple optimization equations. Through establishment of a culture parameter optimization equation, a cell lysis optimization equation, a chromatography purification optimization equation, a concentration optimization equation and a quality determination equation, synergistic optimization of all-process parameters is realized. The method first determines suitable algal cell strains through cell morphology screening, and establishes a culture condition dynamic regulation system based on multiple parameters such as cell growth rate, photosynthetic rate and metabolic rate, so as to ensure uniformity and stability of cell growth.

[0022] In the process implementation, the method of the application realizes precise control of parameters through linkage of the optimization equations. The culture parameter optimization equation dynamically adjusts temperature, illumination and pH value based on the real-time growth state of cells; the cell lysis optimization equation determines optimal ultrasonic crushing conditions according to cell density, viability and other factors; the chromatography purification optimization equation considers the influence of upstream processes and optimizes binding and elution parameters; and the concentration optimization equation selects suitable membrane materials and pressure conditions based on protein characteristics. Finally, the quality determination equation establishes a scientific product quality evaluation system by investigating the coupling relationship between parameters of various processes.

[0023] The method of the application successfully solves the problem of uncoordinated parameter optimization in traditional technologies. Through establishment of an optimization equation group based on multiple parameters, the limitations of traditional empirical parameter setting are overcome, the process is more precisely controllable, and finally high-purity and high-activity antibody products are obtained. This parameter optimization method based on mathematical models not only improves the stability of product quality, but also provides a reliable theoretical basis for process scale-up. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 The figure is a flowchart of the method of the application.

[0025] Figure 2 The figure is a cell growth parameter change trend graph with time in Example 2.

[0026] Figure 3 The figure is a graph of the relationship between crushing efficiency and temperature in the cell lysis process in Example 2.

[0027] Figure 4 The figure is an elution curve graph of the chromatography purification process in Example 2.

[0028] Figure 5 The figure is a graph of changes in protein concentration and activity in the concentration process in Example 2. DETAILED DESCRIPTION

[0029] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application.

[0030] As Figure 1 shown in the figure, the present application provides a flow chart of an antibody expression and purification integrated method based on algal cells, and the method comprises the following steps:

[0031] S01, morphological analysis is performed on algal cell strains, and an algal cell strain meeting the following conditions in terms of cell morphological parameters is selected: cell sphericity is greater than 0.95, cell diameter is 2-5 microns, cell density is 1,000,000-2,000,000 per milliliter, and growth cycle is 48-72 hours;

[0032] S02, the algal cell strain is inoculated into a standard culture medium, the initial temperature of the standard culture medium is 22-25°C, the initial hydrogen ion concentration value is 7.0-7.5, and the initial light intensity is 2,000-3,000 lux, and the culture is performed for 48-72 hours, and a cell growth curve is recorded;

[0033] S03, a pre-fitted culture parameter optimization equation is used, the culture parameter optimization equation takes a cell growth rate vector, a cell photosynthetic rate vector, a cell metabolic rate vector and a cell density vector as input parameters, the cell growth rate vector, the cell photosynthetic rate vector, the cell metabolic rate vector and the cell density vector all comprise continuous monitoring data from 0 to 72 hours, and the output parameters of the culture parameter optimization equation comprise an optimal culture temperature, an optimal light intensity and an optimal hydrogen ion concentration value, the culture temperature, the light intensity and the hydrogen ion concentration value are dynamically adjusted according to the output results of the culture parameter optimization equation;

[0034] S04, a protein expression inducer is added to the adjusted standard culture medium, the concentration of the protein expression inducer is 0.2-0.5 mg / mL, the culture is continued for 24-36 hours, and a cell viability is determined;

[0035] S05, the algal cells are collected by centrifugation, the centrifugal speed is 5,000-6,000 rpm, the algal cells are washed with phosphate buffer solution for 3 times, and are resuspended in a cell lysis buffer;

[0036] S06, a pre-fitted cell lysis optimization equation is used, the cell lysis optimization equation takes the cell density, the cell viability, a lysis temperature, a buffer ion strength and the cell sphericity as input parameters, and the output parameters of the cell lysis optimization equation comprise an optimal ultrasonic power and an optimal lysis time, the algal cells are lysed by using an ultrasonic crushing system according to the output results of the cell lysis optimization equation, and a crushing efficiency is recorded.

[0037] S07, filtering the lysate through a 0.45-micron filter membrane, collecting the filtrate and centrifuging the filtrate at a speed of 10,000 to 12,000 revolutions per minute for 30 to 40 minutes to obtain a supernatant;

[0038] S08, using a previously fitted chromatography purification optimization equation, the chromatography purification optimization equation taking the turbidity of the supernatant, temperature, the hydrogen ion concentration value, and a protein concentration vector as input parameters, wherein the protein concentration vector includes the concentration of target proteins and impurity proteins, the output parameters of the chromatography purification optimization equation including optimal binding time, optimal binding temperature, optimal elution gradient, and optimal flow rate, according to the output results of the chromatography purification optimization equation, adding affinity chromatography medium to the supernatant, incubating the supernatant at the optimal binding temperature for the optimal binding time, and determining the binding rate;

[0039] S09, transferring the incubation mixture into a chromatography column, purifying the mixture using the optimal elution gradient and the optimal flow rate, and recording the peak shape parameters of the elution curve;

[0040] S10, using a previously fitted concentration optimization equation, the concentration optimization equation taking the protein concentration, elution volume, target concentration multiple, isoelectric point of the protein, and molecular weight of the protein as input parameters, the output parameters of the concentration optimization equation including optimal membrane molecular weight cut-off and optimal concentration pressure, selecting an ultrafiltration membrane according to the output results of the concentration optimization equation, concentrating the mixture at the optimal concentration pressure, and determining the concentration multiple, purity index, and activity retention rate;

[0041] S11, using a previously fitted quality determination equation to evaluate the quality of the final purified product, the quality determination equation taking the fragmentation efficiency, the turbidity of the supernatant, the binding rate, the peak shape parameters of the elution curve, the concentration multiple, the purity index, and the activity retention rate as input parameters, the quality determination equation calculating a quality comprehensive score through a coupling relationship between the input parameters, the coupling relationship including a linear coupling term between the fragmentation efficiency and the binding rate, an exponential coupling term between the peak shape parameters of the elution curve and the purity index, and a Gaussian coupling term between the concentration multiple and the activity retention rate, and the final product is determined to be qualified when the quality comprehensive score is greater than 85 points and simultaneously satisfies the purity index being greater than 95%, the activity retention rate being greater than 90%, and the protein concentration being greater than 1 milligram per milliliter;

[0042] Wherein: the cell sphericity refers to the area ratio of the largest inscribed circle to the smallest circumscribed circle in the microscopic image of the algal cell; the cell growth rate refers to the amount of cell density growth per unit time; the cell photosynthetic rate refers to the amount of change in chlorophyll fluorescence intensity per unit time; the cell metabolic rate refers to the amount of glucose consumption in the culture medium per unit time; the fragmentation efficiency refers to the percentage of the degree of cell lysis after ultrasonic treatment; the supernatant turbidity refers to the absorbance value at a wavelength of 280 nm of the supernatant; the binding rate refers to the binding ratio of the affinity chromatography medium to the target protein; the elution curve peak shape parameter refers to the ratio of the half-peak width to the peak height in the elution curve; and the concentration multiple refers to the ratio of the protein concentration after concentration to that before concentration.

[0043] The antibody expression and purification integrated method is suitable for production of human cell surface marker antibodies, including hematopoietic cell differentiation antigen CD45, T lymphocyte common antigen CD3, helper T lymphocyte marker CD4, cytotoxic T lymphocyte marker CD8, natural killer cell marker CD16, neural cell adhesion molecule CD56, B lymphocyte marker CD19, complement regulatory protein CD55, complement regulatory protein CD59, interleukin 2 receptor alpha chain CD25, interleukin 7 receptor alpha chain CD127, programmed death receptor 1 (PD-1), T cell costimulatory molecule CD28, immunoglobulin Fc receptor CD64, complement receptor 3 (CD11b), complement receptor 1 (CD35), perforin; suitable algae include one or more of Chlamydomonas reinhardtii, Chlorella sp., Dunaliella salina, Prorocentrum minimum, Alexandrium sp., Platymonas sp., Scenedesmus sp., Euglena sp., Porphyridium sp., and Nephroselmis sp., preferably Chlamydomonas reinhardtii. The above-mentioned algal cell strains have been established through long-term laboratory domestication and genetic modification, and have established stable ribosome entry site sequences, chloroplast targeting sequences and endoplasmic reticulum localization sequences, and can achieve high-efficiency expression of the above-mentioned human cell surface marker antibodies. These optimized algal cell strains have the advantages of fast growth rate, simple culture conditions, high protein expression level, complete post-translational modification, etc., and through the expression and purification integrated method, high-purity and high-activity recombinant antibody products can be obtained, which exhibit good specificity and sensitivity in immunological detection methods such as flow cytometry, immunohistochemistry, and Western blotting, and meet the application requirements of clinical diagnosis and scientific research.

[0044] The specific implementation of step S01 is to perform morphological analysis using a cell analyzer equipped with a high-resolution microscope system that uses computer vision algorithms to quantitatively evaluate the morphological characteristics of algal cells. First, high-definition images of the cells are collected by the microscope system, with an image resolution of no less than 1920x1080 pixels and a sampling rate of no less than 30 frames per second. Then, an image segmentation algorithm is used to extract the cell contour, which is based on a combination of region growing and edge detection methods and can accurately identify the cell boundary. Then, the minimum circumscribed circle algorithm and the maximum inscribed circle algorithm are used to calculate the sphericity of the cells, with a sphericity threshold of 0.95, which is the optimal parameter obtained through a large amount of experimental data statistics. To ensure accurate determination of cell density, flow cytometry is used to dilute the cell suspension to an appropriate concentration with phosphate buffer, and a flow cytometer is used to count the cells, with a counting time of 60 seconds and a sampling volume of 100 microliters. The determination of the growth cycle uses a continuous culture method, with sampling and determination of cell density every 6 hours, drawing a growth curve, and determining the growth cycle through the curve inflection point. The purpose of this step is to screen the most suitable algal cell strain for antibody expression, ensuring the stability and efficiency of the subsequent expression process.

[0045] The specific implementation of step S02 is to use an intelligent bioreactor system to standardize the culture of algal cells. The preparation of the standard culture medium uses sterile operation, and all reagents are filtered through a 0.22 micron filter to remove bacteria. The temperature control of the culture medium uses a precision temperature control system with a control accuracy of ±0.1 degrees Celsius, and the temperature is uniformly distributed through water bath circulation. The automatic titration system is used to adjust the hydrogen ion concentration, equipped with a high-precision pH electrode with a response time of less than 10 seconds and a control accuracy of ±0.05. The light system uses an LED light source array with a spectral range of 400 to 700 nanometers, a light intensity that can be continuously adjusted from 0 to 10000 lux, and a uniformity error of less than 5%. During the culture process, automatic sampling is performed every 2 hours, and the cell density, culture medium pH, dissolved oxygen and other parameters are recorded by the online monitoring system. This step provides the optimal environment for cell growth by precisely controlling the culture conditions, while obtaining detailed growth data for subsequent optimization.

[0046] The specific implementation of step S03 is a machine learning-based culture parameter real-time optimization system. The system adopts a multi-layer perceptron neural network structure, the input layer contains 192 neurons corresponding to the cell growth rate, photosynthetic rate, metabolic rate and cell density data collected every hour within 72 hours. The hidden layer adopts a 3-layer structure, each layer contains 128, 64 and 32 neurons respectively, and the activation function uses the rectified linear unit function. The output layer contains 3 neurons corresponding to the optimal culture temperature, light intensity and hydrogen ion concentration value. The network training adopts the back propagation algorithm, the learning rate is set to 0.001, the batch size is 32, and the training rounds are 1000. The optimized culture temperature range is 22 to 25 degrees Celsius, the light intensity range is 2000 to 3000 lux, and the pH value range is 7.0 to 7.5. This step realizes intelligent regulation of culture parameters through deep learning method, improves culture efficiency.

[0047] The specific implementation of step S04 is an accurate protein expression induction process. The selection of protein expression inducer is based on the specificity of the target gene promoter, and the commonly used inducers include isopropyl thiogalactoside, methanol, copper ions, etc. The addition of inducer adopts micro-injection system, the injection precision is ±0.01 milliliter, and the injection speed can be adjusted in the range of 0.1 to 10 milliliters per minute. The determination of cell viability adopts trypan blue staining method, 100 microliters of cell suspension is mixed with equal volume of 0.4% trypan blue solution, blood cell counting plate is used to count under microscope, each sample is repeated 3 times, and the average value is taken. This step controls the induction conditions to ensure efficient expression of target protein.

[0048] The specific implementation of step S05 is the collection and pretreatment process of cells. Centrifugal collection adopts large-capacity refrigerated centrifuge, the rotor temperature is maintained at 4 degrees Celsius, the centrifuge tube adopts 500 milliliter specification, and the maximum capacity can reach 4 liters. The preparation of phosphate buffer solution adopts analytical pure reagent, the pH value is adjusted to 7.4, and the ionic strength is 150 millimoles per liter. During the washing process, the supernatant is discarded after each centrifugation, and the same volume of buffer solution is added for resuspension, which is repeated 3 times. After the last washing, cell lysis buffer containing protease inhibitor mixture is added, and the composition of lysis buffer includes 50 millimoles per liter Tris buffer, 150 millimoles per liter sodium chloride, 1% final concentration of detergent, and 1 millimole per liter dithiothreitol. This step prepares for subsequent cell lysis, and removes impurities in the culture medium that may affect subsequent purification.

[0049] The specific implementation of step S06 is to use an intelligent ultrasonic cell disruption system for cell lysis. The system uses a dual-frequency ultrasonic probe design, with a main frequency of 20 kHz and a secondary frequency of 40 kHz. The two frequencies can independently adjust the power output. The ultrasonic treatment uses an intermittent method, with a working time to intermittent time ratio of 2 to 1, and each cycle lasts for 3 minutes. Real-time temperature monitoring is used during the lysis process, and the sample temperature is maintained at 4 to 8 degrees Celsius through a circulating refrigeration system. The determination of the disruption efficiency uses a dual verification method of flow cytometry and microscopy, with sampling and analysis every 30 seconds until the disruption efficiency reaches the preset value. This step achieves efficient cell lysis while protecting the target protein from ultrasonic damage by precisely controlling the ultrasonic parameters.

[0050] The specific implementation of step S07 is a pretreatment process for separating and purifying the lysis products. First, a vacuum filtration system is used for filtration, using a 47 mm diameter polyvinylidene fluoride filter membrane, which is pre-soaked with phosphate buffer. The filtration speed is controlled at 10 to 15 milliliters per minute, and the constant filtration rate is maintained by adjusting the vacuum degree. The filtered sample is centrifuged using a high-speed refrigerated centrifuge, and the centrifuge tube is made of polypropylene material that can withstand high-speed centrifugation, with a capacity of 50 milliliters. The temperature is maintained at 4 degrees Celsius during centrifugation, and the centrifugal force reaches 12000 revolutions per minute, generating a centrifugal force of about 17000 times the acceleration of gravity. Immediately after centrifugation, the turbidity of the supernatant is measured, and the ultraviolet spectrophotometer is used to measure the absorbance at 280 nanometers, while the background absorbance is measured at 320 nanometers. The difference between the two is the actual turbidity data. The purpose of this step is to remove cell debris and insoluble impurities, preparing for subsequent chromatographic purification.

[0051] The specific implementation of step S08 is to use an intelligent affinity chromatography system to preliminarily separate the target protein. The system is equipped with an automatic sample loading device and a precise temperature control system, which can accurately control various parameters during the binding process. The affinity chromatography medium is a surface-modified agarose matrix, and the ligand is selected according to the specific binding site of the target antibody. The binding process is carried out in a constant temperature shaker, with a shaking speed of 150 to 200 revolutions per minute and a temperature control accuracy of ±0.5 degrees Celsius. The determination of the binding rate uses an indirect method, by measuring the concentration of unbound protein in the supernatant and comparing it with the total protein concentration. The protein concentration is determined by the improved Bradford method, using Coomassie Brilliant Blue G250 as the developer, and measuring the absorbance at 595 nanometers. This step achieves efficient capture of the target protein by optimizing the binding conditions.

[0052] The specific implementation of step S09 is fine purification using a fully automatic chromatography system. The chromatography column is selected as a low-pressure liquid chromatography column, a C-type quick connector is used for connection, and the ratio of column bed height to inner diameter is controlled between 8 to 1 and 12 to 1. The system is equipped with a four-channel gradient mixer, which can realize complex elution gradient control. The mobile phase is selected as a phosphate buffer system, and gradient elution is achieved by different concentrations of sodium chloride. Ultraviolet detector and conductivity detector are used for simultaneous monitoring during elution, and the data acquisition frequency is 10 Hz. The recording of elution curve adopts professional chromatography workstation software, which automatically calculates the peak shape parameters, including theoretical plate number, separation degree, and tailing factor. This step realizes high-purity separation of target protein by accurately controlling the chromatography conditions.

[0053] The specific implementation of step S10 is protein concentration using tangential flow filtration technology. The concentration system uses a fully automatic tangential flow filtration device, equipped with a pressure sensor and a flow meter to realize accurate pressure and flow rate control. The selection of ultrafiltration membrane is based on the molecular weight of the target protein, and the membrane material is polyether sulfone or regenerated cellulose with a molecular weight cutoff of 1 / 3 of the target protein. The concentration process uses constant pressure mode, the transmembrane pressure difference is adjusted by a back pressure valve, and the concentration temperature is maintained at 4 to 8 degrees Celsius. The concentration multiple is tracked in real time by an online concentration monitoring system, the purity index is determined by high-performance liquid chromatography, and the activity retention rate is evaluated by enzyme-linked immunosorbent assay. This step realizes efficient concentration of target protein under mild concentration conditions while maintaining its biological activity.

[0054] The specific implementation of step S11 is a quality control system based on multi-parameter comprehensive evaluation. This system uses fuzzy comprehensive evaluation method to establish an analytic hierarchy structure, and divides various quality indicators into three levels of process parameters, purification effect and product characteristics. The weight of the evaluation index is determined by the analytic hierarchy process, a judgment matrix is constructed, and the characteristic value and characteristic vector are calculated. The calculation of quality comprehensive score uses the weighted summation method, and each index is given different weights according to importance. Among them, the process parameters include crushing efficiency, supernatant turbidity and binding rate, the purification effect includes elution curve peak shape parameters, and the product characteristics include concentration multiple, purity index and activity retention rate. Finally, the product quality is judged by setting a threshold value to ensure the stability and reliability of the product quality. This step realizes comprehensive control of product quality through a scientific evaluation system.

[0055] The fitting process of the culture parameter optimization equation first needs to establish a culture parameter database containing cell growth data under different culture conditions. The orthogonal experiment method is used to design the culture condition combination in the data collection stage. Temperature, light intensity, and pH value are set at 5 levels respectively, and 125 groups of culture conditions are constructed. Each group of conditions is repeated for 3 times, and the cell growth rate, photosynthetic rate, metabolic rate, and cell density are monitored for 72 hours continuously. In the data preprocessing stage, the sliding average method is used to eliminate noise, and the standardized method is used to unify the data of different dimensions to the interval of 0 to 1. The equation fitting uses a deep neural network model, uses the root mean square error as the loss function, and uses the Adam optimizer to update the parameters with a learning rate of 0.001. The model verification uses the cross-validation method, and the data set is divided into training set and test set in the ratio of 8 to 2. Through iterative training, when the loss function value on the test set changes by less than 0.0001 for 10 consecutive rounds, the model is considered to have converged. The final optimization equation can predict the optimal culture parameters in real time according to the cell growth state.

[0056] The fitting process of the cell lysis optimization equation first needs to collect experimental data related to cell lysis. Single factor experiment method is used to study the effects of ultrasonic power, treatment time, sample temperature, etc. on lysis effect. High-throughput screening system is used for data collection, which can perform multiple parallel experiments simultaneously. Response surface method is used for experimental design, and central composite design scheme is constructed, which contains all factor levels combinations. Principal component analysis method is used in data analysis stage to extract the main factors affecting lysis effect and establish the correlation matrix between factors. Support vector regression algorithm is used for equation fitting, radial basis kernel function is used, and penalty parameter is optimized by grid search method. The model evaluation uses two indicators of determination coefficient and root mean square error. When the determination coefficient is greater than 0.95 and the root mean square error is less than 0.05, the fitting effect is considered good. The final optimization equation can accurately predict the optimal lysis conditions.

[0057] The fitting process of the chromatography purification optimization equation needs to establish a database containing all key parameters of the chromatography process. Batch experiment method is used to systematically study the effects of binding time, binding temperature, elution gradient, and flow rate on purification effect. Box-Behnken design method is used for experimental design to ensure that test points are evenly distributed in the test space. Automatic chromatography system is used for data collection to record parameters such as ultraviolet absorption, conductivity, and pressure during the whole process. Wavelet transform method is used for data processing stage to denoise and baseline correction of chromatogram, and peak shape characteristic parameters are extracted. Polynomial regression model is used for equation fitting, least squares method is used to estimate model parameters, and stepwise regression method is used to screen significant variables. Leave-one-out cross-validation is used for model verification to calculate the correlation coefficient of predicted value and measured value. When the correlation coefficient is greater than 0.9, the model is considered to have good prediction ability. The final optimization equation can quickly determine the optimal chromatography conditions.

[0058] The fitting process of the concentration optimization equation first needs to obtain experimental data under different concentration conditions. Gradient screening method is used to study the influence of membrane molecular weight cut-off, operating pressure, temperature and other factors on the concentration effect. The experimental design adopts uniform design method to uniformly distribute points in the test factor space, reducing the number of tests. The data collection uses online monitoring system to record the changes of pressure, flow, concentration and other parameters during the concentration process. The data analysis uses partial least squares method to process the data with multiple collinearity and extract the main influencing factors. The equation fitting uses artificial neural network model, the hidden layer uses double hidden layer structure, and the Bayesian regularization method is used to prevent overfitting. The model optimization uses genetic algorithm to find the optimal network parameters through multiple generations of evolution. The model evaluation uses average relative error, when the error is less than 5%, the model accuracy is considered to meet the requirements. The final optimization equation can accurately predict the optimal concentration conditions.

[0059] The fitting process of the quality determination equation needs to establish a complete quality evaluation system. Delphi method is used to collect expert opinions to determine the weight of each quality index. The data collection stage establishes a product quality history database containing the quality detection data of all batches of products. The data preprocessing uses outlier detection method to eliminate obviously deviating data points. The equation fitting uses fuzzy neural network model, which combines fuzzy logic reasoning and neural network learning. The model training uses hybrid learning algorithm, forward propagation to calculate membership, and backward propagation to optimize parameters. The model verification uses bootstrap method to evaluate the stability of the model by repeated sampling. When the prediction accuracy is more than 95%, the model is considered reliable. The final determination equation can accurately evaluate the product quality grade.

[0060] The equations or calculation processes involved in the present application are described in detail as follows.

[0061] 1. Cell sphericity calculation equation:

[0062] The specific calculation of the cell sphericity is as follows:

[0063]

[0064] In the formula, S is the cell sphericity; A in is the maximum inscribed circle area, with the unit of square microns; A out is the minimum circumscribed circle area, with the unit of square microns.

[0065] Wherein, the parameter acquisition method is:

[0066] A in Obtained by image processing, the steps include: 1) collecting cell microscopic image; 2) image binarization processing; 3) edge detection; 4) maximum inscribed circle fitting.

[0067] A out The step of obtaining by image processing includes: 1) collecting a cell microscopic image; 2) image binarization processing; 3) edge detection; and 4) minimum circumscribed circle fitting.

[0068] The equation uses an area ratio to represent the roundness of the cell shape. The closer the ratio is to 1, the closer the cell shape is to an ideal sphere. The area ratio is selected instead of the perimeter ratio because the area is more sensitive to shape changes. The derivation process of the equation is as follows:

[0069] First step: From the perspective of image analysis, the two-dimensional projection of a perfect sphere should be a perfect circle, at which time the inscribed circle and the circumscribed circle coincide;

[0070] Second step: Considering the deviation of the actual cell shape, the area ratio of the inscribed circle to the circumscribed circle is used to quantify the degree of deviation;

[0071] Third step: Through a large number of experimental data verification, 0.95 is determined as the critical value.

[0072] In practical applications, the equation has the following characteristics: the value range is between 0 and 1, which is convenient for standardization; it is sensitive to shape changes and can accurately reflect the morphological differences of cells.

[0073] 2. Cell growth rate calculation equation:

[0074] The specific calculation of the cell growth rate is as follows:

[0075]

[0076] In the formula, μ(t) is the cell growth rate at time t, with a unit of per hour; X(t) is the cell density at time t, with a unit of per milliliter; α is an initial growth rate adjustment factor, with a value range of 0.01-0.1; and β is a time decay coefficient, with a value range of 0.001-0.01.

[0077] Wherein, the parameter acquisition method is:

[0078] X(t) is obtained by flow cytometry, and sampling is performed every 2 hours.

[0079] The difference method is used for calculation:

[0080] The equation considers the specific growth rate and the time decay term, and the exponential decay term reflects the self-limiting effect in the cell growth process. The derivation process is as follows:

[0081] First step: Based on Monod growth kinetics, establish the basic specific growth rate expression

[0082] Second step: Introduce an exponential decay term αe considering the self-limiting effect of cell growth -βt ;

[0083] Third step: Determine the value range of parameters α and β through experimental data fitting.

[0084] Where the specific representation of the cell density vector X(t) is:

[0085]

[0086] In the formula, t1, t2, …, t n are sampling time points, with an interval of 2 hours.

[0087] 3. Cell photosynthetic rate calculation equation:

[0088] The specific calculation of the cell photosynthetic rate is as follows:

[0089]

[0090] In the formula, P(t) is the photosynthetic rate at time t, with a unit of relative fluorescence intensity per hour; I is the light intensity, with a unit of lux; K I is the light saturation constant, taking a value of 2000 lux; C is the carbon dioxide concentration, with a unit of milligrams per liter; K C is the carbon dioxide saturation constant, taking a value of 300 milligrams per liter; k1 is the maximum photosynthetic rate, taking a value range of 0.5-2; k2 is the time response coefficient, taking a value range of 0.1-0.5; γ is the basic photosynthetic rate, taking a value range of 0.05-0.1.

[0091] This equation is based on Michaelis-Menten kinetics, considering the double-substrate limitation of light and carbon dioxide, as well as the time response characteristics. The derivation process is as follows:

[0092] First step: Based on Michaelis-Menten kinetics, establish a double-substrate limitation model of light intensity and carbon dioxide concentration;

[0093] Second step: Introduce an exponential term considering the time delay of photosynthesis initiation process

[0094] Third step: Add a constant term γ considering the basic metabolic maintenance;

[0095] Fourth step: Determine the parameter values through the experimental data of light response curve and CO2 response curve.

[0096] 4. Cell metabolic rate calculation equation:

[0097] The specific calculation of the cell metabolic rate is as follows:

[0098]

[0099] where M(t) is the metabolic rate at time t, with unit of mg glucose per liter per hour; [Glu] is the glucose concentration, with unit of mg per liter; k3 is the basal metabolic coefficient, with value range of 0.01-0.05; k4 is the substrate promotion coefficient, with value range of 1-5; K m is the Michaelis constant, with value of 500 mg per liter; δ is the non-specific metabolism term, with value range of 0.1-0.5.

[0100] This equation combines first-order kinetics and Michaelis kinetics to describe the glucose consumption process of cells. The derivation process is as follows:

[0101] First step: Establish the basal glucose consumption model k3X(t);

[0102] Second step: Consider the substrate promotion effect and introduce the Michaelis kinetics term

[0103] Third step: Consider non-specific metabolism and add constant term δ;

[0104] Fourth step: Determine the parameter values through glucose consumption kinetics experiments.

[0105] 5. Culture parameter optimization equation:

[0106] The specific representation of the culture parameter optimization equation is as follows:

[0107]

[0108] where T opt is the optimal culture temperature, with unit of degrees Celsius; I opt is the optimal light intensity, with unit of lux; pH opt is the optimal hydrogen ion concentration value; is the 72-hour growth rate vector; is the 72-hour photosynthetic rate vector; is the 72-hour metabolic rate vector; is the 72-hour cell density vector; W1, W2 are weight matrices; b1, b2 are bias vectors; σ is the activation function; ε T , ε I , ε pH are error terms, all with a range of ±0.1.

[0109] This equation uses a deep neural network structure to realize the mapping from the cell growth state to the optimal culture parameters. The specific representation of the input vector is as follows:

[0110]

[0111] The dimension of the weight matrix is:

[0112]

[0113] where h is the number of hidden neurons and n is the length of time series 72. The derivation process is as follows:

[0114] First step: build a three-layer neural network structure;

[0115] Second step: select ReLU as the activation function σ;

[0116] Third step: use the backpropagation algorithm to optimize the weight matrix;

[0117] Fourth step: determine the network structure parameters through cross-validation.

[0118] 6. Cell lysis optimization equation:

[0119] The specific representation of the cell lysis optimization equation is as follows:

[0120]

[0121] where P is the optimal ultrasonic power, unit is watt; t is the optimal lysis time, unit is second; D, V, T, I, S are input feature vectors [D, V, T, I, S]; D is cell density; V is cell viability; T is lysis temperature; I is buffer ionic strength; S is cell sphericity; α is the coefficient of support vector; φ is the kernel function; b, b are bias terms. opt opt T i P t

[0122] The kernel function uses the Gaussian radial basis function:

[0123]

[0124] where σ is the kernel function parameter, determined by cross-validation.

[0125] This equation is based on support vector regression, and the kernel trick is used to realize nonlinear mapping, which improves the generalization ability of the model. The specific representation of the input feature vector is:

[0126]

[0127] The coefficient matrix of the support vector is represented as:

[0128] ​​​​​​​

[0129] The derivation process is as follows:

[0130] First step: build a feature space, select a suitable kernel function;

[0131] Second step: solve the dual optimization problem to obtain support vectors;

[0132] Third step: use the grid search method to optimize the kernel function parameter σ;

[0133] Fourth step: determine the optimal model parameters by cross-validation.

[0134] 7. Chromatography purification optimization equation:

[0135] The specific representation of the chromatography purification optimization equation is as follows:

[0136]

[0137] In the formula, t b is the optimal binding time, in minutes; T b is the optimal binding temperature, in degrees Celsius; is the optimal elution gradient vector; v is the optimal flow rate, in milliliters per minute; is the input feature vector A 280 is the supernatant turbidity; T is the temperature; pH is the hydrogen ion concentration value; is the protein concentration vector; β i is the polynomial coefficient matrix; ε t , ε T , ε v is the error term.

[0138] This equation uses a polynomial regression model and considers high-order interactions between variables. The specific representation of the input feature vector is:

[0139]

[0140] Where the protein concentration vector is:

[0141]

[0142] The specific representation of the gradient vector is:

[0143]

[0144] In the formula, k is the number of impurity protein species, and p is the number of gradient segments. The derivation process is as follows:

[0145] First step: establish a polynomial regression model framework;

[0146] Step 2: Significant variables are screened by stepwise regression method;

[0147] Step 3: Polynomial coefficients are estimated by least square method;

[0148] Step 4: Polynomial order is determined by leave-one-out cross validation.

[0149] 8. Concentration optimization equation:

[0150] The specific representation of the concentration optimization equation is as follows:

[0151]

[0152] In the formula, MWCO is the optimal membrane cut-off molecular weight, unit is Dalton; P opt is the optimal concentration pressure, unit is bar; is the input feature vector [C, V, N, pi, MW] T ; C is the protein concentration; V is the elution volume; N is the target concentration multiple; pi is the protein isoelectric point; MW is the protein molecular weight; W3, W4, W5 are weight matrices; is the bias vector; f is the output activation function; ε MW , ε P is the error term.

[0153] This equation adopts a feedforward neural network structure, realizing the nonlinear mapping from protein characteristic parameters to concentration process parameters. The specific representation of the input feature vector is as follows:

[0154]

[0155] The dimension of the weight matrix is:

[0156]

[0157] The derivation process is as follows:

[0158] Step 1: Construct a double-hidden layer neural network structure;

[0159] Step 2: Choose tanh as the hidden layer activation function σ;

[0160] Step 3: Choose a linear function as the output activation function f;

[0161] Step 4: Use Bayesian regularization to prevent overfitting;

[0162] Step 5: Optimize network parameters by genetic algorithm.

[0163] 9. Quality determination equation:

[0164] The specific representation of the quality determination equation is as follows:

[0165]

[0166] where Q is the quality comprehensive score; E is the breakage efficiency; B is the binding rate; P is the elution curve peak shape parameter; U is the purity index; N is the concentration fold; A is the activity retention rate; ω1, ω2, ω3 are weight coefficients, and ω1+ω2+ω3=1; k1, k2, k3 are coupling coefficients; σ is the Gaussian function parameter; ε Q is the error term.

[0167] This equation contains three coupling relationships: 1) linear coupling of breakage efficiency and binding rate; 2) exponential coupling of peak shape parameter and purity; 3) Gaussian coupling of concentration fold and activity, which comprehensively reflects the relationship between process and product quality. The derivation process is as follows:

[0168] First step: based on process experience, determine three main coupling relationships;

[0169] Second step: linear coupling term reflects the positive correlation between breakage efficiency and binding rate;

[0170] Third step: exponential coupling term reflects the nonlinear relationship between peak shape parameter and purity;

[0171] Fourth step: Gaussian coupling term reflects the optimal matching relationship between concentration fold and activity;

[0172] Fifth step: determine the weight coefficients by the analytic hierarchy process;

[0173] Sixth step: verify the reliability of the model through historical data.

[0174] Specifically, the principle of the present application is: the technical principle of the present application is based on system engineering and biological process control theory. First, by establishing a cell morphological parameter system including cell sphericity, diameter, density and other indicators, it is ensured that the selected algal cell strain has stable growth characteristics. These morphological parameters are closely related to the physiological function of the cell and affect the subsequent protein expression and purification effect.

[0175] During the cultivation process, the present application establishes an optimization equation based on multi-dimensional parameters. This equation takes the cell growth rate vector, photosynthetic rate vector, metabolic rate vector and cell density vector as input parameters, and calculates the optimal cultivation conditions through a mathematical model. The principle of this dynamic optimization method is that: the cell growth rate reflects the cell proliferation state, the photosynthetic rate represents the energy metabolism level, the metabolic rate indicates the material conversion efficiency, and the cell density reflects the overall growth condition. The synergistic effect of these parameters determines the physiological state of the cell and the protein expression capacity.

[0176] During the process implementation, there are logical correlations between the optimization equations. The cell lysis optimization equation considers the upstream parameters such as cell density and viability, to ensure that the lysis conditions match the cell state. The chromatography purification optimization equation takes into account factors such as supernatant turbidity and temperature to optimize the binding and elution conditions. The concentration optimization equation is based on the physicochemical properties of the protein to select appropriate operating parameters. Finally, the quality determination equation achieves full-process quality control by establishing coupling relationships between parameters, including linear coupling, exponential coupling, and Gaussian coupling. This parameter optimization principle based on mathematical models ensures the scientificity and controllability of the process.

[0177] A specific embodiment 1 of the present application is provided below, and the specific implementation of each step in embodiment 1 is described in detail as follows.

[0178] The specific implementation of step S01 is to use a high-resolution microscopic image analysis system for quantitative evaluation of cell morphology. The system is equipped with a high-speed camera and image analysis software. First, microscopic image acquisition is performed using a 60x objective lens, with an image resolution of 1920x1080 pixels, a sampling rate of 30 frames per second, and an exposure time of 0.1 milliseconds to ensure image clarity. Then, a composite segmentation algorithm based on region growing and edge detection is used to process the image. The algorithm first uses Gaussian filtering for noise reduction, uses the Sobel operator for edge detection with a threshold of 128, then connects the edges and fills the regions through region growing, and finally optimizes the boundaries through morphological operations. Next, the cell sphericity is calculated according to the formula where the maximum inscribed circle area A in The starting point is selected as the centroid, the step size is 0.1 pixels, and the minimum circumscribed circle area A out The least squares method is used for fitting, and the iteration termination condition is that the residual error is less than 0.01. The sphericity threshold is set to 0.95, which is obtained through statistical analysis of 1000 samples. The cell diameter measurement uses the equivalent circle diameter method, which equates the irregular cell profile to a circle, with an effective range of 2 to 5 microns. The cell density determination uses the flow cytometry counting method, with a sample dilution factor of 10, a counting volume of 100 microliters, a counting time of 60 seconds, and an average value obtained by repeating the measurement 3 times, with a density control of 1000000 to 2000000 per milliliter. The growth cycle determination uses the continuous culture method, with sampling and density determination every 6 hours, and the determination of the logarithmic growth phase through the density change curve, with an effective range of 48 to 72 hours. This step realizes quantitative screening of algal cell strains through morphological analysis, providing high-quality cell materials for subsequent expression processes.

[0179] The specific implementation of step S02 is to use an intelligent bioreactor system for standardized culture. The preparation of the standard culture medium is carried out by sterile operation, and all reagents are filtered through a 0.22 micron cellulose acetate membrane to remove bacteria. Temperature control uses a water bath circulation method, with a temperature control range of 22 to 25 degrees Celsius, a temperature control accuracy of ±0.1 degrees Celsius, and real-time monitoring using a platinum resistance temperature sensor. Hydrogen ion concentration control uses an intelligent pH adjustment system, equipped with a composite glass electrode, with a response time of less than 10 seconds, a control range of 7.0 to 7.5, a control accuracy of ±0.05, and automatic titration using hydrochloric acid and sodium hydroxide solution. Light control uses an LED array with a wavelength range of 400 to 700 nanometers, a light intensity adjustable range of 2000 to 3000 lux, a uniformity error of less than 5%, and real-time monitoring using a light quantum sensor. During the culture process, automatic sampling is performed every 2 hours to determine the cell growth rate μ(t), according to the formula , where α is in the range of 0.01 to 0.1, and β is in the range of 0.001 to 0.01. At the same time, the photosynthetic rate P(t) is measured, according to the formula , where k1 is in the range of 0.5 to 2, k2 is in the range of 0.1 to 0.5, and γ is in the range of 0.05 to 0.1. In addition, the cell metabolic rate M(t) is also measured, according to the formula , where k3 is in the range of 0.01 to 0.05, k4 is in the range of 1 to 5, and δ is in the range of 0.1 to 0.5. This step provides a data basis for establishing an optimization model by precisely controlling the culture conditions and real-time monitoring of growth parameters.

[0180] The specific implementation of step S03 is a deep learning-based real-time optimization system for culture parameters. The system uses a three-layer neural network structure, with 288 neurons in the input layer corresponding to 72 hours of growth rate, photosynthetic rate, metabolic rate, and cell density data. The hidden layer uses a three-layer structure, with 128, 64, and 32 neurons in each layer, and the activation function uses a rectified linear unit function. The output layer contains 3 neurons, corresponding to the optimal culture temperature, light intensity, and hydrogen ion concentration values. Network optimization uses the stochastic gradient descent method, with a learning rate of 0.001, a batch size of 32, and 1000 training rounds. According to the culture parameter optimization equation , the optimal parameters are calculated, where is the 72-hour monitoring data vector, W1, W2 are weight matrices, b1, b2 are bias vectors, and ε T , ε I , and ε pH are error terms, all with a range of ±0.1. The optimized culture temperature range is 22 to 25 degrees Celsius, the light intensity range is 2000 to 3000 lux, and the pH value range is 7.0 to 7.5. This step realizes intelligent regulation of culture parameters through deep learning methods.

[0181] The specific implementation of step S04 is the precise protein expression induction process. The selection of the inducer is based on the specificity of the target gene promoter, and isopropyl thiogalactoside is commonly used as the inducer, with a concentration range of 0.2 to 0.5 mg / mL. The inducer is added using a micro-injection system, with an injection accuracy of ±0.01 mL and an injection speed that can be adjusted within the range of 0.1 to 10 mL / min. The induction process is carried out in a constant-temperature shaker at a speed of 180 rpm and a temperature identical to the culture temperature. The induction time is 24 to 36 hours, during which the target protein expression is monitored every 4 hours. The cell viability is determined by trypan blue staining, 100 μL of cell suspension is mixed with an equal volume of 0.4% trypan blue solution, and the blood cell counting plate is used to count under a microscope, with 3 repeated counts for each sample and an average value taken, and the viability threshold is set to 90%. This step achieves efficient expression of the target protein by controlling the induction conditions.

[0182] The specific implementation of step S05 is the collection and pretreatment process of the cells. The centrifugal collection uses a large-capacity refrigerated centrifuge with a rotor temperature maintained at 4°C, 500 mL scale centrifuge tubes with a maximum capacity of 4 L, and a centrifugal speed of 5000 to 6000 rpm for 10 minutes. The phosphate buffer is prepared using analytical reagents, with a pH value adjusted to 7.4 and an ionic strength of 150 mM / L. The washing process uses the centrifugal sedimentation method, with an equal volume of buffer added to resuspend the cells after each collection, repeated 3 times. After the last washing, cell lysis buffer is added, which includes 50 mM / L Tris buffer, 150 mM / L NaCl, 1% final concentration of detergent, 1 mM / L dithiothreitol, and a mixture of protease inhibitors. This step prepares for subsequent cell lysis and removes impurities in the culture medium that may affect subsequent purification.

[0183] The specific implementation of step S06 is the use of an intelligent ultrasonic cell disruption system for cell lysis. The system uses a dual-frequency ultrasonic probe with a main frequency of 20 kHz and a secondary frequency of 40 kHz, with a power range of 50 to 500 W, and the two frequencies can independently adjust the power output. According to the cell lysis optimization equation The optimal lysis conditions are calculated, with the input feature vector including cell density, cell viability, lysis temperature, buffer ionic strength, and cell sphericity, and the kernel function uses Gaussian radial basis function The ultrasonic treatment is in an intermittent mode, the ratio of working time to intermittent time is 2:1, and each cycle lasts for 3 minutes. Real-time temperature monitoring is used during the lysis process, and the sample temperature is maintained in the range of 4 to 8 degrees Celsius through a circulating refrigeration system. The determination of the breaking efficiency adopts a double verification method of flow cytometry and microscopy, and the sample is analyzed every 30 seconds until the breaking efficiency reaches more than 95%. This step realizes efficient lysis of cells while protecting target proteins from ultrasonic damage by precisely controlling ultrasonic parameters.

[0184] The specific implementation of step S07 is a pretreatment process for separating and purifying the lysis product. First, a vacuum filtration system is used for filtration, a polyvinylidene fluoride filter membrane with a diameter of 47 mm and a pore size of 0.45 microns is used, and the filter membrane is soaked in phosphate buffer for 10 minutes in advance. The filtration speed is controlled at 10 to 15 milliliters per minute, the flow rate is controlled through a vacuum degree regulating valve, and the vacuum degree is maintained at 0.05 megapascals. The filtered sample is centrifuged using a high-speed refrigerated centrifuge, and the centrifuge tube is made of polypropylene material resistant to high-speed centrifugation, with a capacity of 50 milliliters. The centrifugation conditions are 10,000 to 12,000 revolutions per minute, the temperature is maintained at 4 degrees Celsius, and the time is 30 to 40 minutes. After centrifugation, the turbidity of the supernatant is measured, the ultraviolet spectrophotometer is used to measure the absorbance at 280 nanometers, and the background absorbance is measured at 320 nanometers, the difference between the two is taken as the actual turbidity data, and the turbidity threshold is set to 0.5. This step removes cell debris and insoluble impurities through efficient separation, preparing for subsequent chromatographic purification.

[0185] The specific implementation of step S08 is to use an intelligent affinity chromatography system to preliminarily separate the target protein. According to the chromatographic purification optimization equation the optimal purification parameters are calculated, where is the input feature vector, including the supernatant turbidity, temperature, pH value and protein concentration data, β i is a polynomial coefficient matrix. The affinity chromatography medium is selected from a surface-modified agarose matrix, and the ligand is selected according to the specific binding site of the target antibody. The binding process is carried out in a constant temperature shaker, the shaker speed is 150 to 200 revolutions per minute, and the temperature control accuracy is ±0.5 degrees Celsius. The determination of the binding rate adopts an indirect method, which is calculated by comparing the concentration of unbound protein in the supernatant with the total protein concentration, and the binding rate threshold is set to 85%. The protein concentration is determined by the improved Bradford method, using Coomassie Brilliant Blue G250 as the developer, and measuring the absorbance at 595 nanometers. This step realizes efficient capture of the target protein by optimizing the binding conditions.

[0186] The specific implementation of step S09 is fine purification using a fully automatic chromatography system. The chromatography column is selected as a low-pressure liquid chromatography column, a C-type quick connector is used, and the ratio of column bed height to inner diameter is controlled between 8:1 and 12:1. The system is equipped with a four-channel gradient mixer, and the mobile phase is selected as a phosphate buffer system, which is eluted by gradient elution through different concentrations of sodium chloride. The ultraviolet detector and conductivity detector are used to monitor the elution process simultaneously, and the data acquisition frequency is 10 Hz. The elution curve is recorded using professional chromatography workstation software, which automatically calculates the peak shape parameters, including theoretical plate number, separation degree, and tailing factor. The peak shape evaluation uses the ratio of half-peak width to peak height, and the smaller the value, the better the separation effect, with a threshold value of 0.2. This step realizes high-purity separation of the target protein by accurately controlling the chromatography conditions.

[0187] The specific implementation of step S10 is protein concentration using tangential flow filtration technology. According to the concentration optimization equation the optimal concentration parameters are calculated, where is the input feature vector, including protein concentration, elution volume, target concentration multiple, protein isoelectric point, and molecular weight data. The concentration system uses a fully automatic tangential flow filtration device equipped with a pressure sensor and flow meter. The selection of ultrafiltration membrane is based on the molecular weight of the target protein, and the membrane material is polyether sulfone or regenerated cellulose with a molecular weight cutoff of 1 / 3 of the target protein molecular weight. The concentration process uses a constant pressure mode, and the transmembrane pressure difference is adjusted by a back pressure valve, and the concentration temperature is maintained at 4 to 8 degrees Celsius. The concentration multiple is tracked in real time by an online concentration monitoring system, the purity index is determined by high performance liquid chromatography, and the activity retention rate is evaluated by enzyme-linked immunosorbent assay. This step realizes efficient concentration of the target protein under mild concentration conditions while maintaining its biological activity.

[0188] The specific implementation of step S11 is a quality control system based on multi-parameter comprehensive evaluation. According to the quality determination equation the quality is evaluated, where E is the crushing efficiency, B is the binding rate, P is the peak shape parameter of the elution curve, U is the purity index, N is the concentration multiple, A is the activity retention rate, and ω1, ω2, ω3 are weight coefficients. The weight of the evaluation index is determined by the analytic hierarchy process, a judgment matrix is constructed, and the eigenvalue and eigenvector are calculated. The final quality evaluation standard is: when the quality comprehensive score is greater than 85 points and at the same time the purity index is greater than 95%, the activity retention rate is greater than 90%, and the protein concentration is greater than 1 mg / mL, it is determined as qualified product. This step realizes comprehensive control of product quality through a scientific evaluation system.

[0189] The specific implementation of each equation fitting step will now be described in detail:

[0190] The fitting process of the culture parameter optimization equation is first to establish a culture parameter database. The orthogonal test method is used to design the culture conditions, the temperature is set to 5 levels, which are 22, 23, 24, 25 degrees Celsius and the intermediate temperature 23.5 degrees Celsius; the light intensity is set to 5 levels, which are 2000, 2250, 2500, 2750, 3000 lux; the pH value is set to 5 levels, which are 7.0, 7.1, 7.3, 7.4, 7.5, and 125 groups of culture conditions are constructed. Each group of conditions is repeated for 3 times, and the cell growth data is monitored for 72 hours continuously. The monitoring data includes cell growth rate vector photosynthetic rate vector metabolic rate vector and cell density vector The data preprocessing uses the moving average method to eliminate noise, the window size is 5 and the step is 1. The maximum and minimum value standardization method is used to unify the data of different dimensions to the interval of 0 to 1. The equation fitting uses a deep neural network model, the network structure contains 288 neurons in the input layer, three hidden layers with 128, 64, 32 neurons respectively, and 3 neurons in the output layer. The loss function uses the root mean square error, the optimizer selects the Adam algorithm, the learning rate is set to 0.001, the momentum parameters β1 is 0.9 and β2 is 0.999. The model verification uses the 5-fold cross-validation method, the data set is divided into training set and test set according to the ratio of 8:2. Through iterative training, when the loss function value on the test set changes less than 0.0001 for 10 consecutive rounds, the model is considered to be converged. The final optimization equation can predict the optimal culture parameters in real time according to the cell growth state.

[0191] The fitting process of the cell lysis optimization equation is first to collect experimental data related to lysis. Using single factor test method, the effect of ultrasonic power in the range of 50 to 500 watts, treatment time in the range of 1 to 10 minutes, and sample temperature in the range of 0 to 20 degrees Celsius on lysis effect is studied. The experimental design uses response surface method to construct a central composite design scheme, and the experimental points include cubic vertex points, axial points and center points. The data collection uses a high-throughput screening system equipped with an automatic sampling device and an online monitoring system. A data set is established with cell density, cell viability, lysis temperature, buffer ionic strength, and cell sphericity as independent variables, and ultrasonic power and lysis time as dependent variables. The data analysis uses principal component analysis method to extract the main influencing factors through eigenvalue decomposition, and the contribution rate threshold is set to 85%. The equation fitting uses support vector regression algorithm, and the kernel function selects Gaussian radial basis function. The kernel parameter σ is optimized by grid search method, and the search range is 0.1 to 10 with a step of 0.1. The optimization range of penalty parameter C is 1 to 100 with a step of 1. The model evaluation uses two indexes of determination coefficient and root mean square error. When the determination coefficient is greater than 0.95 and the root mean square error is less than 0.05, the fitting effect is considered good.

[0192] The fitting process of the chromatography purification optimization equation is first to establish a database containing key parameters of the chromatography process. Using batch experiment method, the effect of binding time in the range of 10 to 60 minutes, binding temperature in the range of 4 to 25 degrees Celsius, elution gradient in the range of 0 to 1 mole per liter of sodium chloride, and flow rate in the range of 0.1 to 2 milliliters per minute on purification effect is studied. The experimental design uses Box-Behnken design method, three factors and three levels, a total of 15 experimental points. The data collection uses an automated chromatography system to record the ultraviolet absorption, conductivity and pressure parameters of the whole process, and the sampling frequency is 10 hertz. The data processing uses wavelet transform method to denoise and baseline correct the chromatogram, selects Daubechies4 wavelet, and the decomposition layer is 3 layers. The peak shape characteristic parameters are extracted, including retention time, peak area, theoretical plate number, separation degree, tailing factor, etc. The equation fitting uses polynomial regression model, including linear term, quadratic term and interaction term, uses least squares method to estimate model parameters, and selects significant variables by stepwise regression method, and the significance level is set to 0.05. The model verification uses leave-one-out cross-validation to calculate the correlation coefficient of predicted value and measured value. When the correlation coefficient is greater than 0.9, the model is considered to have good prediction ability.

[0193] The fitting process of the concentration optimization equation is first to obtain experimental data under different concentration conditions. Gradient screening method is used to study the effect of membrane molecular weight cut-off in the range of 3 to 100 kilodaltons, operating pressure in the range of 0.05 to 0.5 megapascal, and temperature in the range of 4 to 25 degrees Celsius on the concentration effect. The experimental design uses uniform design method, and U n (q m ) type design table is selected, where n is the number of experiments, q is the number of levels, and m is the number of factors. Data collection uses online monitoring system to record pressure, flow, and concentration parameter changes during concentration process with 1 minute sampling interval. Data analysis uses partial least squares method to handle data with multiple collinearity, extract main influencing factors, and accumulate variance contribution rate greater than 85%. Equation fitting uses artificial neural network model, network structure is double hidden layer, first hidden layer uses sigmoid activation function, second hidden layer uses tanh activation function, and Bayesian regularization method is used to prevent overfitting. Regularization parameter is determined by maximum likelihood estimation. Model optimization uses genetic algorithm with population size of 100, crossover probability of 0.8, mutation probability of 0.1, and evolution number of 1000 generations. Model evaluation uses average relative error, and when the error is less than 5%, the model accuracy is considered to meet the requirements.

[0194] The fitting process of the quality determination equation is first to establish a complete quality evaluation system. Delphi method is used to collect evaluation opinions, and the evaluation group consists of 10 process developers, quality control personnel and application engineers. Through three rounds of questionnaires, the weights of each quality index are determined. In the data collection stage, a product quality historical database is established, which contains all the quality detection data of 100 batches of products. The data includes process parameters (crushing efficiency, supernatant turbidity, binding rate), purification effect (elution curve peak shape parameter), and product characteristics (concentration ratio, purity index, activity retention rate). Data preprocessing uses 3σ principle for outlier detection, and data points deviating obviously are removed. Equation fitting uses fuzzy neural network model, input layer uses fuzzy processing, uses Gaussian membership function, and fuzzy rule uses Mamdani reasoning system. Network training uses hybrid learning algorithm, forward propagation calculates membership, backward propagation optimizes parameters, and learning rate is dynamically adjusted. Model verification uses bootstrap method with 1000 times of repeated sampling, and calculates the confidence interval of model prediction. When the prediction accuracy is more than 95%, the model is considered reliable. The finally established determination equation can accurately evaluate the product quality grade and provide basis for product release. These equation fitting processes reflect systematicness and scientificalness, and through reasonable experimental design, data processing and model construction, the reliability and practicability of the optimization equation are ensured.

[0195] For better understanding and implementation of the present application, the following provides an embodiment 2 of a specific application scenario of the present application: during the production research of CD3 antibody, the research team adopts the antibody expression and purification integrated method of the present application, selects Chlamydomonas reinhardtii as the expression system, and the specific implementation process is as follows.

[0196] Firstly, cell morphology analysis is carried out, and the morphology parameters of the genetically modified Chlamydomonas reinhardtii cell strain are measured. High-resolution microscope system is used for image acquisition and analysis, and the cell morphology parameters are measured as shown in Table 1:

[0197] Table 1 Chlamydomonas reinhardtii cell morphology parameter measurement results

[0198] Parameter Measured value Standard deviation Cell sphericity 0.97 ±0.02 Cell diameter (microns) 3.5 ±0.3 Cell density (cells / mL) 1.5 x 10 6 ]]> ±0.2 x 10 6 ]]> Growth cycle (hours) 58 ±2

[0199] The screened algal cell strain is inoculated in the standard culture medium, and the initial culture conditions are: temperature 23.5 degrees Celsius, pH value 7.2, and light intensity 2500 lux. After 48 hours of culture, the recorded cell growth parameters are shown in Table 2:

[0200] Table 2 Cell growth monitoring data

[0201]

[0202] Figure 2 The change trend of cell growth parameters with time is shown. In the figure, the red solid line represents the growth rate (h-1), and the blue dotted line represents the cell density (×10 6 mL-1). The optimal culture conditions calculated according to the culture parameter optimization equation are: temperature 24.2 degrees Celsius, light intensity 2750 lux, and pH value 7.3. After adjusting the culture conditions to the optimal values, 0.3 milligrams per milliliter of isopropyl thiogalactoside is added as an inducer, and the culture is continued for 30 hours. The cell viability data during the induction period are shown in Table 3:

[0203] Table 3 Cell viability monitoring results during induction

[0204] Time (hours) Cell viability (%) 0 98.5 6 97.8 12 96.5 18 95.2 24 94.8 30 93.6

[0205] The cells are collected by centrifugation at 5500 revolutions per minute for 10 minutes, and washed with pH 7.4 phosphate buffer for 3 times. According to the cell lysis optimization equation, the optimal lysis conditions are calculated as: ultrasonic power 350 watts and lysis time 4 minutes. Under the above conditions, cell lysis is carried out, and the measured breakage efficiency data are shown in Table 4:

[0206] Table 4 Cell lysis process monitoring data

[0207] Time (minutes) Fragmentation efficiency (%) Temperature (degrees Celsius) 1 45.5 5.2 2 78.3 6.1 3 92.7 6.8 4 96.8 7.5

[0208] Figure 3 The relationship between the breakage efficiency and temperature during cell lysis is shown. The blue line and circles represent the breakage efficiency (%), and the red line and triangles represent the temperature (°C). The lysate was filtered through a 0.45 micron filter and centrifuged at 11,000 rpm for 35 minutes. The absorbance of the supernatant at 280 nm was measured to be 0.42. The optimal purification conditions calculated from the chromatography purification optimization equation are shown in Table 5:

[0209] Table 5. Optimal conditions for chromatography purification

[0210] Parameter Value Binding time (minutes) 45 Binding temperature (degrees Celsius) 8 Elution gradient (M NaCl) 0-0.5 Flow rate (mL / min) 0.8

[0211] Figure 4 The elution profile of the chromatography purification process is shown. The blue solid line represents the protein peak (UV 280 absorbance), and the red dashed line represents the conductivity change. The affinity chromatography purification was performed using the above conditions, and the binding rate was measured to be 88.5%. The peak shape parameter (half-peak width to peak height ratio) of the elution profile was 0.15. According to the concentration optimization equation, the optimal membrane molecular weight cut-off was calculated to be 30 kilodaltons, and the optimal concentration pressure was calculated to be 0.15 megapascals. The monitoring data of the concentration process are shown in Table 6:

[0212] Table 6. Monitoring data of the concentration process

[0213] Time (minutes) Protein concentration (mg / mL) Activity (%) 0 0.25 100 15 0.52 98.5 30 0.88 97.2 45 1.35 95.8 60 1.82 93.5

[0214] Figure 5 The changes in protein concentration and activity during the concentration process are shown. The blue line and circles represent the protein concentration (mg·mL-1), and the red line and triangles represent the activity retention rate (%). The quality evaluation indicators of the final product are shown in Table 7:

[0215] Table 7. Quality evaluation results of the product

[0216] Indicator Value Fragmentation efficiency (%) 96.8 Binding rate (%) 88.5 Peak shape parameter 0.15 Purity (%) 96.5 Concentration fold 7.3 Activity retention rate (%) 93.5 Quality comprehensive score 88.7

[0217] The overall quality score calculated from the quality judgment equation is 88.7, which meets the requirements of a purity greater than 95%, an activity retention rate greater than 90%, and a protein concentration greater than 1 milligram per milliliter, and is determined to be a qualified product.

[0218] Compared with the traditional antibody expression and purification method, the integrated method has significant advantages. The traditional method mainly uses mammalian cells (such as CHO cells) as the expression system, and the culture conditions are harsh, the culture cost is high, the culture period is long (usually 10-14 days are needed), and a complex culture medium formula and strict culture condition control are required. In the purification process, the traditional method mainly relies on the setting of empirical parameters, and lacks a systematic optimization method, resulting in unstable purification efficiency and large product quality fluctuations. Through the method of the present application, the culture period is shortened to 3-4 days, the culture cost is reduced by about 60%, the batch consistency of the product is significantly improved, the purity is increased by about 2 percentage points, and the activity retention rate is increased by about 5 percentage points. At the same time, the present application realizes the intelligent regulation of process parameters by establishing a series of optimization equations, greatly reduces the process development difficulty and operation requirement, and improves the production efficiency and product quality stability.

[0219] It should be noted that the variables involved in the present application are explained in detail as shown in Table 8.

[0220] Table 8: Variable Explanation Table

[0221]

[0222]

[0223] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. An integrated method for antibody expression and purification based on algal cells, characterized in that, include: Morphological analysis was performed on algal cell lines, and algal cell lines that met the requirements for cell morphology parameters were selected. The algal cell line was inoculated into a standard culture medium and cultured. Culture conditions were dynamically adjusted using a culture parameter optimization equation; protein expression inducers were added for continued culture. Collect algal cells and resuspend them in lysis buffer; The ultrasonic lysis parameters were determined using a cell lysis optimization equation; the lysate was filtered and the supernatant was collected by centrifugation; the affinity chromatography parameters were determined using a chromatography purification optimization equation. The incubation mixture was transferred to a chromatography column for purification; Concentration parameters were determined using a concentration optimization equation; the quality of the purified product was evaluated using a quality judgment equation.

2. The method according to claim 1, characterized in that, The cell morphology parameters must meet the following conditions: cell sphericity greater than 0.95, cell diameter between 2 and 5 micrometers, cell density between 1,000,000 and 2,000,000 per milliliter, and growth cycle between 48 and 72 hours.

3. The method according to claim 2, characterized in that, The standard culture medium was initially cultured at a temperature of 22 to 25°C, an initial hydrogen ion concentration of 7.0 to 7.5, and an initial light intensity of 2000 to 3000 lux for 48 to 72 hours, and the cell growth curve was recorded.

4. The method according to claim 3, characterized in that, The culture parameter optimization equation uses cell growth rate vector, cell photosynthetic rate vector, cell metabolic rate vector, and cell density vector as input parameters, and output parameters include optimal culture temperature, optimal light intensity, and optimal hydrogen ion concentration.

5. The method according to claim 4, characterized in that, The concentration of the protein expression inducer was 0.2 to 0.5 mg / mL, and the cells were cultured for another 24 to 36 hours before cell viability was measured.

6. The method according to claim 5, characterized in that, The algal cells were collected by centrifugation at a speed of 5000 to 6000 rpm, washed three times with phosphate buffer, and resuspended in cell lysis buffer.

7. The method according to claim 6, characterized in that, The cell lysis optimization equation takes the cell density, cell viability, lysis temperature, buffer ionic strength, and cell sphericity as input parameters, and outputs the optimal ultrasound power and optimal lysis time.

8. The method according to claim 7, characterized in that, The lysate was filtered through a 0.45-micron filter membrane, and the filtrate was collected and centrifuged at a speed of 10,000 to 12,000 revolutions per minute for 30 to 40 minutes. The turbidity of the supernatant was then measured.

9. The method according to claim 8, characterized in that, The chromatography purification optimization equation takes the turbidity of the supernatant, temperature, hydrogen ion concentration, and protein concentration vector as input parameters, and outputs the optimal binding time, optimal binding temperature, optimal elution gradient, and optimal flow rate.

10. The method according to claim 9, characterized in that, The quality assessment of the purified product using the quality judgment equation is specifically defined as follows: when the overall quality score is greater than 85 points and the purity index is greater than 95%, the activity retention rate is greater than 90%, and the protein concentration is greater than 1 mg / mL, the product is judged to be qualified.

Images