Control method of biological product chromatography process based on digital twinborn technology

Abstract

The invention relates to the technical field of protein chromatographic separation in the fields of biological pharmacy and biological engineering, in particular to a digital twinning system and a control method, system and device for a biological product chromatographic process based on the digital twinning technology. The method comprises the following steps: collecting real-time data and / or biological product property data of the operation process of the physical chromatography equipment; simulating and predicting the operation process of the physical chromatography equipment according to the real-time data through a pre-established virtual entity bidirectionally mapped with the physical chromatography equipment, determining a behavior strategy of the chromatography process and generating a behavior strategy instruction; and the physical chromatography equipment completes a corresponding operation process according to the received behavior strategy instruction. Digital twinning established based on the chromatography mechanism model is adopted, real-time closed-loop optimization of the chromatography process is guided, the robustness and efficiency of the chromatography process are improved, the control precision is improved, the whole process is fully automatic, and a large amount of manpower and material resources can be saved.

Description

Technical Field

[0001] The present invention relates to the field of protein chromatography separation technology in the fields of biopharmaceuticals and bioengineering, and more specifically, to a control method, system, and device for a biological product chromatography process based on digital twin technology, and a digital twin system for a biological product chromatography process. Background Art

[0002] Antibody drugs are the most important protein drugs, with good targeting, high efficacy, and low toxicity and side effects. In recent years, with the growth of upstream cell expression and the expansion of production scale, the economy and robustness of downstream separation processes have received more and more attention. The traditional downstream separation process of antibodies is crude separation by protein A affinity capture and refined separation by two-step chromatography.

[0003] The two-step polishing chromatography is divided into cationic chromatography, anionic chromatography, hydrophobic interaction chromatography and mixed mode chromatography according to different separation principles. The specific polishing chromatography combination is optimized and designed according to the impurities contained in the antibody feed solution.

[0004] The antibody protein purification process requires the separation and purification of relatively small amounts of impurities such as aggregates, fragments, charged heterogeneities, host cell proteins, DNA, etc., which places extremely high demands on the robustness and precision of the process.

[0005] Due to the uncertainty of the upstream culture process of biological samples, the feed solution generally has certain fluctuations in concentration and impurity content, which will bring certain challenges to the downstream separation process.

[0006] Taking cationic chromatography as an example, this chromatography step generally separates the target protein and impurities based on the charge difference of different proteins. The charge difference of the proteins is reflected in the gradient or isocratic elution process. Different proteins have different retention times. The purpose of purifying the target protein can be achieved by collecting the eluate within the retention time of the target protein.

[0007] Generally speaking, when the concentration fluctuations of the upstream feed liquid and the impurity content are small, a fixed retention time collection method can be adopted. However, when the fluctuations are large, the retention times of the target protein and impurities vary greatly. If fixed collection is still used, a certain loss of yield and purity will occur.

[0008] At this time, in order to cope with the fluctuations generated in the upstream culture process, there are generally three solutions (see reference Journal of Chemical Technology & Biotechnology, 90(2): 341-348).

[0009] The first method is to use a fixed UV absorption value to determine the collection start and end point, that is, to monitor the real-time UV absorption value of the eluted protein on an automated chromatograph, and start or end the collection when the specified collection UV absorption value is reached, thereby coping with the shift in protein retention time and peak shape caused by fluctuations in the upstream process to a certain extent. This method is the simplest and most commonly used, but the quality of the final product cannot be precisely controlled.

[0010] The second method is to adopt the method of collecting in separate bags, collect in separate bags throughout the elution process, and perform HPLC detection on each bag of samples to obtain accurate mass data to determine the peak distribution of various substances in the elution peak, and combine the purity and yield targets to obtain the collection interval. After determining the sample combination plan based on the collection interval, each bag of samples is combined. Although this method can cope with the fluctuations of the upstream process well and accurately control the purity and yield, it has the disadvantages of complex process, high production cost, and long detection time in actual application. For example, more disposable liquid storage bags are needed, which increases the production cost, and the detection time is relatively long, generally about 1 day. Subsequent production steps need to wait for the test results to be fed back before they can proceed, which greatly prolongs the production time and reduces production capacity.

[0011] The third method uses a technology based on real-time detection and statistical model (statistic model), for example, the literature of Rathore et al. (Biotechnology and Bioengineering, 2008, 100: 306-316) inputs the real-time online detection data of the elution process into a well-fitted statistical model (statistic model), calculates the collection interval by the statistical model, and controls the collection. For example, the literature of Brestrich et al. (Biotechnology and Bioengineering, 2015, 112 (7): 1406-1416) removes the elution peak type by the established PLS model, calculates the collection interval and collects. Although the above two methods can reduce a certain detection time and bag distribution cost, there are certain limitations in the statistical model, such as the need for a large amount of data to fit the model in the early stage, and the real-time feedback of experimental data is required during use, and the requirements for detection and calculation speed are relatively high. At the same time, due to the poor extension of the statistical model, only relatively simple systems can be processed, and each time the system is replaced, a large amount of experimental data is required to update the model, and practical application is relatively difficult.

[0012] Digital twins are virtual models of physical objects. This virtual model can interact with the real world in real time to perform process simulation, virtual experiments, real-time control, maintenance reminders, and other tasks, thereby guiding the actual process.

[0013] At present, digital twin technology is widely used in aerospace, intelligent manufacturing, and urban management. For example, Chinese invention patent CN110399642A discloses a digital twin for a production line and its construction method and application, and Chinese invention patent CN111230887A discloses a method for monitoring the operating status of an industrial glue coating robot based on digital twin technology. However, digital twin technology is not widely used in the field of biopharmaceuticals.

[0014] In the existing technology, the application of digital twin technology in the biopharmaceutical field is limited to the upstream fermentation process. Digital twins have not yet been maturely applied to the downstream separation chromatography and purification process of biological drugs. This is mainly because the downstream separation chromatography and purification process has many steps and parameters, which makes data collection difficult, the calculation cost is high, and the control is difficult. Summary of the invention

[0015] The purpose of the present invention is to provide a control method, system, device and digital twin system for a biological product chromatography process based on digital twin technology, so as to solve the problem of difficulty in controlling the separation and purification of biological samples in the prior art.

[0016] Another object of the present invention is to provide a biological product chromatography process control method, system, device and a digital twin system for a biological product chromatography process based on digital twin technology to solve the problem of low control accuracy of biological sample separation and purification in the prior art.

[0017] In order to achieve the above object, the present invention provides a control method for a biological product chromatography process based on digital twin technology, comprising:

[0018] Collect real-time data of the operation process of physical chromatography equipment and / or biological product property data;

[0019] Through the pre-established virtual entity that is bidirectionally mapped with the physical chromatography equipment, the operation process of the physical chromatography equipment is simulated and predicted according to real-time data, the behavior strategy of the chromatography process is determined and the behavior strategy instructions are generated;

[0020] The physical chromatography device completes the corresponding operation process according to the received behavior strategy instructions.

[0021] In one embodiment, the virtual entity, through a bidirectional mapping established with the physical chromatography device, establishes a model based on the operating data of the physical chromatography device and / or the biological product property data, simulates the operating process of the physical chromatography device, and performs simulation analysis and optimization on the operating process of the physical chromatography device.

[0022] In one embodiment, the virtual entity creation steps are:

[0023] Based on the real-time and / or historical operation data of the physical chromatography equipment and the property data of the biological product, the model dynamic parameters of the chromatography process are defined and the model is established.

[0024] In one embodiment, the virtual entity establishment step includes:

[0025] Based on at least two groups of real-time and / or historical operation data of the physical chromatography equipment operation process of the biological product chromatography process and the biological product property data, combined with the preset fitting target, define the model dynamic parameters of the chromatography process and establish the model; more specifically, in combination with the preset fitting target, fit and calculate the dynamic parameters of each model and establish the model;

[0026] In one embodiment, the control method further includes:

[0027] Based on the real-time data of the operation process of the physical chromatography equipment of the biological product chromatography process and the biological product property data, the dynamic parameters of the model are adjusted in real time to optimize the model through evaluation between the real-time data and the simulation prediction results.

[0028] In one embodiment, the model comprises a chromatographic mechanism model;

[0029] Preferably, the chromatography mechanism model includes an intra-column mechanism model and / or an adsorption mechanism model;

[0030] Wherein, the in-column mechanism model is a convection and diffusion mass transfer model of protein in the chromatography column;

[0031] The adsorption mechanism model is a model of protein adsorption and dissociation on the surface of the chromatographic medium.

[0032] In one embodiment, the column mechanism model includes an equilibrium model, an equilibrium diffusion model (EDM), a lumped kinetic model (LKM), a shrinking core model and a general rate model (GRM). In one embodiment, the adsorption mechanism model includes a linear driving force model (LDF), a Langmuir adsorption isotherm model, a Freundlich adsorption isotherm model, an Anti-Langmuir adsorption isotherm model, a steric mass action model (SMA), a multi-state steric mass action model (Multi-State Steric Mass Action Model) and a colloidal particle adsorption model (CPA).

[0033] In one embodiment, the virtual entity predicts the operation process of the physical chromatography equipment through real-time data simulation and / or adjusts the simulation prediction results in real time, and determines and / or optimizes the behavior strategy of the chromatography process according to the set single or multiple objectives, and / or constraint parameters;

[0034] Preferably, the single or multiple objectives include bioproduct component selection, bioproduct yield, and bioproduct distribution;

[0035] The constraint parameters include bioproduct yield constraint, impurity ratio constraint, bioproduct component constraint and component distribution constraint.

[0036] In one embodiment, the virtual entity predicts the operation process of the physical chromatography equipment through real-time data simulation and / or adjusts the simulation prediction results in real time, and determines and / or optimizes the behavior strategy of the chromatography process based on specific biological product components and their yields, impurity ratios, and biological product distribution constraint parameters.

[0037] In one embodiment, the behavior strategy includes node control parameters of the chromatography process;

[0038] Among them, preferably, the node control parameters of the chromatography process include sample loading time point, elution collection starting point, elution collection end point, and elution volume.

[0039] In one embodiment, the biological product includes antibody proteins, enzymes and peptides, fusion proteins, structural proteins, regulatory proteins, growth factors, hormones and cytokines, nucleic acid molecules, compounds, and compound-protein conjugates.

[0040] In order to achieve the above object, the present invention provides a digital twin system for a biological product chromatography process, comprising:

[0041] Physical chromatography equipment, which is a physical application scenario of the digital twin system, including physical equipment for bioproduct chromatography and / or the physical environment of the physical equipment;

[0042] A multi-sensor data acquisition module, which is disposed on or connected to the physical chromatography device and is used to obtain real-time operation data of the physical chromatography device and / or biological product property data;

[0043] A virtual entity that establishes a bidirectional mapping relationship with the physical chromatography device and determines a behavior strategy based on the real-time data of the operation process of the physical chromatography device and / or the biological product property data received by the virtual entity;

[0044] The control service module generates corresponding control instructions according to the behavior strategy determined by the virtual entity, and transmits the control instructions to the physical chromatography equipment to perform operation process control.

[0045] In one embodiment, the virtual entity includes:

[0046] The signal domain is used to receive and transform the operation data of the physical chromatography equipment and / or the property data of the biological product acquired by the multi-sensor data acquisition module, and provide a data basis for the simulation domain;

[0047] A simulation domain, which establishes a bidirectional mapping model with the physical chromatography device according to the operation data of the physical chromatography device and / or the property data of the biological product;

[0048] The optimization domain uses a model that is bidirectionally mapped with the physical chromatography equipment to optimize and determine the behavior strategy based on real-time data and / or physical goals set by the chromatography process;

[0049] and / or,

[0050] The storage domain stores the operation data, property data and the optimized data of the optimization domain in a database to facilitate access by the control service module.

[0051] In one embodiment, a model of the physical chromatography device is established in the simulation domain, and it is capable of simulating and predicting the physical chromatography device in real time according to real-time data and / or adjusting the simulation prediction results in real time, and determining the behavior strategy through the optimization domain.

[0052] In one embodiment, the model bidirectionally mapped with the physical chromatography device includes a chromatography mechanism model; preferably, the chromatography mechanism model includes an intra-column mechanism model and / or an adsorption mechanism model.

[0053] In one embodiment, the multi-sensor data acquisition module includes a high performance liquid chromatograph (HPLC), a multi-angle light scattering sensor (MALS), a UV-VIS absorption sensor, a fluorescence sensor, an infrared absorption sensor (IR), an attenuated total reflection-Fourier transform infrared spectroscopy sensor (ATR-FTIR), a refractive index (RI) sensor, a pH sensor, a temperature sensor, a conductivity sensor, a pressure sensor, a small angle X-ray scattering (SAXS) sensor and a redox sensor or a combination thereof.

[0054] In one embodiment, the real-time data includes loading time, collection time, collection volume, gradient starting point, gradient end point, elution volume, loading amount, conductivity, UV absorbance, and elution peak type change;

[0055] The biological product property data include the properties of the protein components themselves and / or the changes in the protein during the chromatography process.

[0056] In order to achieve the above object, the present invention provides a control system for a biological product chromatography process based on digital twin technology, comprising:

[0057] A server, for receiving relevant data collected by detection, and for executing the control method of the biological product chromatography process based on the digital twin technology as described in any one of the above items, and transmitting the determined behavior strategy instructions to the chromatography equipment;

[0058] The chromatography device is used to collect data of the biological product chromatography process and transmit it to the server, and is used for chromatographic separation of the biological product, and it performs biological product chromatography based on the behavioral strategy instructions transmitted by the server.

[0059] In one embodiment, the control system additionally includes sensor equipment for collecting other data and transmitting the data to the server.

[0060] In one embodiment, data is exchanged between the chromatography device and the server;

[0061] The interaction data includes: state transfer data, method start and stop control data, behavior strategy control data and adjustment data, and collection interval data.

[0062] In order to achieve the above object, the present invention provides a control device for a biological product chromatography process based on digital twin technology, comprising:

[0063] A memory for storing instructions and related data executable by the processor;

[0064] A processor is used to execute the instructions to implement any of the methods described above.

[0065] The present invention provides a control method, system, device and digital twin system for a biological product chromatography process based on digital twin technology. The digital twin established based on a chromatography mechanism model is used to guide real-time closed-loop optimization of the chromatography process, increase the robustness and process efficiency of the chromatography process, improve control accuracy, and make the entire process fully automated, which can save a lot of manpower and material resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The above and other features, properties and advantages of the present invention will become more apparent through the following description in conjunction with the accompanying drawings and embodiments, in which the same reference numerals always represent the same features, wherein:

[0067] Figure 1 A schematic diagram of the principle of a digital twin system for a biological product chromatography process according to an embodiment of the present invention is disclosed;

[0068] Figure 2 A step diagram of a control method for a biological product chromatography process based on digital twin technology according to an embodiment of the present invention is disclosed;

[0069] Figure 3 A graph showing the HPLC test results of a protein A sample solution according to an embodiment of the present invention is disclosed;

[0070] Figure 4a A graph showing the detection results of each component in Experiment 1 according to an embodiment of the present invention is disclosed;

[0071] Figure 4b A graph showing the test results of each component in Experiment 2 according to an embodiment of the present invention is disclosed;

[0072] Figure 5a A comparison result diagram of the fitting results and the actual detection results of each component of Experiment 1 according to an embodiment of the present invention is disclosed;

[0073] Figure 5b A comparison result diagram of the fitting results and actual test results of each component of Experiment 2 according to an embodiment of the present invention is disclosed;

[0074] Figure 6 A schematic diagram of protein elution curve simulation prediction according to an embodiment of the present invention is disclosed;

[0075] Figure 7 A principle block diagram of a control system for a biological product chromatography process based on digital twin technology according to an embodiment of the present invention is disclosed;

[0076] Figure 8aA first predicted elution result graph of a chromatography process control system according to an embodiment of the present invention is disclosed;

[0077] Figure 8b A first actual elution result diagram of a chromatography process control system according to an embodiment of the present invention is disclosed;

[0078] Figure 9a A second predicted elution result graph of a chromatography process control system according to an embodiment of the present invention is disclosed;

[0079] Figure 9b A second actual elution result diagram of a chromatography process control system according to an embodiment of the present invention is disclosed;

[0080] Fig.10 A principle block diagram of a control device for a biological product chromatography process based on digital twin technology according to an embodiment of the present invention is disclosed.

[0081] The meanings of the reference numerals in the figures are as follows:

[0082] 11Physical chromatography equipment;

[0083] 12 multi-sensor data acquisition modules;

[0084] 13 virtual entities;

[0085] 131 signal domain;

[0086] 132 simulation domain;

[0087] 133 optimized domain;

[0088] 134 storage domain;

[0089] 14 control service module;

[0090] 71 sensor equipment;

[0091] 72 servers;

[0092] 73 Chromatography equipment;

[0093] 101 processor;

[0094] 102 memory. DETAILED DESCRIPTION

[0095] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the invention and are not used to limit the invention.

[0096] Figure 1The schematic diagram of the principle of the digital twin system for the chromatography process of biological products according to one embodiment of the present invention is disclosed, such as Figure 1 As shown, the digital twin system for the chromatography process of biological products proposed by the present invention includes:

[0097] The physical chromatography equipment 11 is a physical application scenario of the digital twin system, including the physical equipment for the chromatography of biological products and / or the physical environment of the physical equipment;

[0098] The multi-sensor data acquisition module 12 is connected to the physical chromatography device 11 and is used to obtain real-time data of the operation data of the physical chromatography device 11 and / or biological product property data;

[0099] The virtual entity 13 establishes a bidirectional mapping relationship with the physical chromatography device 11, and determines a behavior strategy based on the received real-time data of the operation process of the physical chromatography device 11 and / or the biological product property data;

[0100] The control service module 14 generates corresponding control instructions according to the behavior strategy determined by the virtual entity 13, and transmits the control instructions to the physical chromatography device 11 to perform operation process control.

[0101] It should be noted that the control service module 14 and the virtual entity 13 can be mounted on the same physical server or on different physical servers without affecting the realization of the functions. The multi-sensor data acquisition module 12 can be set as an independent physical device; or, in some scenarios, the multi-sensor data acquisition module 12 and the physical chromatography device 11 are set in the same physical device.

[0102] The following is a detailed introduction to each component involved in the digital twin system.

[0103] The real-time data of the operation data of the physical chromatography device 11 include but are not limited to the loading time, collection time, collection volume, gradient starting point, gradient end point, elution volume, loading amount, conductivity, UV absorbance value and elution peak type change.

[0104] The biological product property data of the physical chromatography device 11 include but are not limited to the properties of the protein components themselves and / or the changes of the protein during the chromatography process.

[0105] The multi-sensor data acquisition module 12 includes but is not limited to high performance liquid chromatography (HPLC), multi-angle light scattering sensor (MALS), UV-VIS absorption sensor, fluorescence sensor, infrared absorption sensor (IR), attenuated total reflection-Fourier transform infrared spectroscopy sensor (ATR-FTIR), refractive index (RI) sensor, pH sensor, temperature sensor, conductivity sensor, pressure sensor, small angle X-ray scattering (SAXS) sensor and redox sensor or a combination thereof.

[0106] The virtual entity 13 further includes a signal domain 131 , a simulation domain 132 and an optimization domain 133 .

[0107] Furthermore, the virtual entity 13 also includes a storage domain 134 .

[0108] The signal domain 131 is used to receive and transform the operation data of the physical chromatography device 11 and / or the biological product property data acquired by the multi-sensor data acquisition module 12, and provide a data basis for the simulation domain;

[0109] The simulation domain 132 establishes a bidirectional mapping model with the physical chromatography device 11 according to the operation data of the physical chromatography device 11 and / or the property data of the biological product;

[0110] The optimization domain 133 uses a model bidirectionally mapped with the physical chromatography device 11 to optimize and determine the behavior strategy according to the real-time data and / or the entity target set in the chromatography process.

[0111] Furthermore, in the simulation domain 132 , a model of the physical chromatography device 11 is established, and the physical chromatography device 11 can be simulated and predicted in real time according to real-time data and / or the prediction result of the model can be adjusted in real time, and the behavior strategy can be determined through the optimization domain 133 .

[0112] The storage domain 134 stores the operation data, property data and the optimized data of the optimization domain in a database to facilitate access by the control service module 14 .

[0113] The model bidirectionally mapped with the physical chromatography device 11 includes a chromatography mechanism model;

[0114] Preferably, the chromatography mechanism model includes an intra-column mechanism model and / or an adsorption mechanism model;

[0115] Wherein, the in-column mechanism model is a convection and diffusion mass transfer model of protein in the chromatography column;

[0116] The adsorption mechanism model is a model of protein adsorption and dissociation on the surface of the chromatographic medium.

[0117] The essence of the chromatography process is that the protein transfers in the chromatography column and then adsorbs and elutes on the medium surface. Therefore, two models are needed to describe the chromatography process, one is the in-column mechanism model and the other is the adsorption mechanism model. The relationship between the in-column mechanism model and the adsorption mechanism model is parallel.

[0118] The column mechanism model includes but is not limited to:

[0119] The complexity of the models increases continuously, including equilibrium model (Equilibrium model), equilibrium diffusion model (Equilibrium dispersive model, EDM), lumped kinetic model (LKM), shrinking core model (Shrinking core model) and general rate model (GRM).

[0120] To fully describe the mass transfer and adsorption process of the chromatography column, it is necessary to consider the convective mass transfer, axial diffusion mass transfer, solid-liquid mass transfer and internal diffusion mass transfer of the particles. Different models make trade-offs for some influencing factors:

[0121] The simplest ideal equilibrium model only considers the convection-diffusion terms and static adsorption equilibrium in the column;

[0122] The equilibrium diffusion model takes into account the axial diffusion effect of the liquid flow in the column based on the ideal equilibrium model.

[0123] The general rate model takes all mass transfer into account.

[0124] The adsorption mechanism model includes but is not limited to:

[0125] Linear Driving Force (LDF), Langmuir adsorption isotherm model, Freundlich adsorption isotherm model, Anti-Langmuir adsorption isotherm model, Steric Mass Action Model (SMA), Multi-State Steric Mass Action Model and colloidal particle model (CPA).

[0126] Chromatographic mechanism models include but are not limited to all of the above models and combinations of models.

[0127] When choosing a model, it is generally necessary to be purpose-oriented and first consider whether the model can accurately describe the target system. If multiple models can describe the target system, it is advisable to choose a model with a simple calculation process and relatively few parameters.

[0128] Figure 2 A step diagram of a control method for a biological product chromatography process based on digital twin technology according to an embodiment of the present invention is disclosed, such as Figure 2 As shown, the present invention proposes a control method for a biological product chromatography process based on digital twin technology, comprising the following steps:

[0129] Step S1, collecting real-time data of the operation process of the physical chromatography equipment and / or biological product property data;

[0130] Step S2, through a pre-established virtual entity bidirectionally mapped with the physical chromatography device, the operation process of the physical chromatography device is simulated and predicted according to real-time data, the behavior strategy of the chromatography process is determined and the behavior strategy instruction is generated;

[0131] Step S3: The physical chromatography device completes the corresponding operation process according to the received behavior strategy instruction.

[0132] Combine the following Figure 1 The digital twin system for the chromatography process of a bioproduct is shown and further explained.

[0133] The present invention proposes a method for controlling a biological product chromatography process based on digital twin technology, further comprising the following steps:

[0134] Step S1, collecting real-time data of the operation process of the physical chromatography device 11 and / or biological product property data through the multi-sensor data acquisition module 12;

[0135] Step S2, transmitting the real-time data to a pre-established virtual entity 13 bidirectionally mapped with the physical chromatography device, simulating and predicting the operation process of the physical chromatography device 11 through real-time data and / or adjusting the simulation prediction result in real time, determining the behavior strategy of the chromatography process, generating a behavior strategy instruction and sending it to the physical chromatography device 11;

[0136] Step S3: After receiving the behavior strategy instruction, the physical chromatography device 11 completes the corresponding operation process.

[0137] Combination Figure 1 and Figure 2 In the illustrated embodiment, the virtual entity 13 establishes a bidirectional mapping with the physical chromatography device 11, and simulates the operation process of the physical chromatography device 11 based on a model established according to the operation data of the physical chromatography device 11 and / or the property data of the biological product, and performs simulation analysis and optimization on the operation process of the physical chromatography device 11.

[0138] The steps of establishing the virtual entity 13 are:

[0139] Based on the real-time and / or historical operation data of the physical chromatography device 11 and the property data of the biological product, the model dynamic parameters of the chromatography process are defined and the model is established.

[0140] Furthermore, the virtual entity 13 establishment step includes:

[0141] Based on at least two groups of real-time and / or historical operating data of the physical chromatography equipment 11 of the biological product chromatography process and biological product property data, combined with pre-set fitting targets, the model dynamic parameters of the chromatography process are defined and the model is established.

[0142] Furthermore, the model comprises a chromatography mechanism model;

[0143] Preferably, the chromatography mechanism model includes an intra-column mechanism model and / or an adsorption mechanism model;

[0144] The in-column mechanism model and the adsorption mechanism model have been discussed in detail in the above-mentioned digital twin system and will not be repeated here.

[0145] Furthermore, based on the real-time data of the operation process of the physical chromatography equipment 11 of the biological product chromatography process and the biological product property data, the dynamic parameters of the model are adjusted in real time through the evaluation between the real-time data and the simulation prediction results to optimize the model.

[0146] The biological products include, but are not limited to, antibody proteins, enzymes and peptides, fusion proteins, structural proteins, regulatory proteins, growth factors, hormones and cytokines, nucleic acid molecules, compounds, and compound-protein conjugates.

[0147] Furthermore, the virtual entity 13 predicts the operation process of the physical chromatography device 11 through real-time data simulation, and determines and / or optimizes the behavior strategy of the chromatography process according to the set single objective or multiple objectives, and / or constraint parameters;

[0148] Preferably, the single or multiple objectives include bioproduct component selection, bioproduct yield, and bioproduct distribution;

[0149] The constraint parameters include bioproduct yield constraint, impurity component ratio, bioproduct component constraint and component distribution constraint.

[0150] Furthermore, the virtual entity 13 predicts the operation process of the physical chromatography device 11 through real-time data simulation, and determines and / or optimizes the behavior strategy of the chromatography process according to specific biological product components and their yields, impurity ratios, and biological product distribution constraint parameters.

[0151] The behavior strategy includes node control parameters of the chromatography process;

[0152] Among them, preferably, the node control parameters of the chromatography process include sample loading time point, elution collection starting point, elution collection end point, and elution volume.

[0153] The control method of the biological product chromatography process based on digital twin technology proposed in the present invention detects real-time data of the operation process of the physical chromatography equipment and / or the property data of the biological product, and uses digital twin technology to accurately control the chromatography process in real time to cope with fluctuations in process parameters in the chromatography process.

[0154] In order to further illustrate the digital twin system for the biological product chromatography process proposed by the present invention and the control method of the biological product chromatography process based on the digital twin technology, the method steps proposed by the present invention are described in detail below by taking cation chromatography as an example. It should be understood that within the scope of the present invention, the above-mentioned technical features of the present invention and the technical features specifically described below (such as the embodiments) can be combined with each other and interrelated to form a preferred technical solution.

[0155] Step S1, collecting real-time data of the operation process of the physical chromatography equipment and / or biological product property data.

[0156] In this embodiment, during the operation of the physical chromatography equipment, the biological product property data involved in the chromatography process, namely, protein-related parameters, are detected in real time through the multi-sensor data acquisition module.

[0157] The protein-related parameters include, but are not limited to, protein composition, protein concentration, and the content of each substance in the protein composition.

[0158] The protein composition includes, but is not limited to, charge heterogeneities (acidic peaks, basic peaks) generated due to changes in protein groups, aggregates, fragments, host cell proteins and other impurities present in the loaded sample.

[0159] Step S2: by using a pre-established virtual entity that is bidirectionally mapped with the physical chromatography device, the operation process of the physical chromatography device is simulated and predicted according to real-time data, the behavior strategy of the chromatography process is determined and a behavior strategy instruction is generated.

[0160] The model dynamic parameters of the chromatography process are defined by combining at least two sets of experimental data of the physical chromatography equipment operation process of the biological product chromatography process with the pre-set fitting target, wherein "definition" means determining the phase model dynamic parameters by calculation and / or fitting method.

[0161] In this embodiment, the model dynamic parameters of the chromatography process are determined by combining at least two sets of experimental data of the physical chromatography equipment operation process of the biological product chromatography process with the preset fitting target, further comprising the following steps:

[0162] The distribution of elution peaks during chromatography was determined by two elution experiments under different conditions.

[0163] The experimental results are fitted with the minimum residual between the predicted value of the mechanism model and the experimental results as the fitting target;

[0164] Obtain model dynamic parameters of the tomographic process.

[0165] More specifically, combining the fitting goal of the minimum residual between the predicted value of the mechanism model and the experimental result, the dynamic parameters of the model are fitted and calculated, the fitting function is determined, and the model is established;

[0166] Based on the real-time data of the operation process of the physical chromatography equipment of the biological product chromatography process and the biological product property data, the dynamic parameters of the model are adjusted in real time to optimize the model through comparison between the real-time data and the simulation prediction results.

[0167] Preferably, the sample loading capacity and elution gradient of two elution experiments with different conditions are different. The distribution of each protein component in the elution peak is determined by the method of separate collection and high performance liquid chromatography detection, and the minimum root mean square error between the predicted value of the mechanism model and the experimental result is used as the evaluation index to optimize the model dynamic parameters of the chromatography process.

[0168] Furthermore, genetic algorithm is used for parameter fitting, and the parameters of the tomographic mechanism model are obtained by fitting the experimental results.

[0169] Genetic algorithm refers to a method of solving unconstrained and constrained nonlinear optimization problems based on the natural selection process that imitates biological evolution. It is currently a relatively common optimization algorithm.

[0170] The genetic algorithm used to solve nonlinear optimization problems is basically the same and is widely used in many parameter nonlinear optimization problems.

[0171] At each optimization step, the genetic algorithm randomly selects individuals from the current population and uses them as parents to generate the next generation of children. After generations of iterations, the final population is the optimal solution to the problem.

[0172] The chromatography mechanism model further includes an intra-column mechanism model and / or an adsorption mechanism model:

[0173] In this embodiment, the column mechanism model adopts the equilibrium diffusion model, and the corresponding equation (1) is:

[0174]

[0175] Among them, c i is the concentration of the i-th protein in the chromatography column, in mol / L;

[0176] D app is the apparent axial diffusion coefficient in the column, in m 2 / s;

[0177] z is the axial distance inside the column, in m;

[0178] u is the empty tower velocity, in m / s;

[0179] ε t is the total void fraction in the column;

[0180] q i is the solid phase protein concentration, in g / L;

[0181] t is the running time in seconds.

[0182] In this embodiment, the adsorption mechanism model is the spatial mass action SMA model, and the corresponding equation (2) is:

[0183]

[0184] Among them, k kin,i , k eq,i , v i and σ i are the adsorption kinetic constants of the ith protein (in sM v ), equilibrium constant, characteristic charge number and shielding factor;

[0185] ∧ is the ion exchange capacity of the medium, in mol / L, c s is the salt concentration in mol / L.

[0186] In addition to ∧ being a medium-related parameter, k kin,i , k eq,i , v i and σ i They are all parameters related to protein properties, that is, different protein components in the elution system have different parameter values, but the ∧ is the same.

[0187] The obtained model dynamic parameters are transferred to the chromatography mechanism model, and the behavior strategy for optimizing the chromatography process, namely the elution and collection behavior strategy, is determined through the chromatography mechanism model optimization calculation.

[0188] In this embodiment, the protein-related parameters detected in step S1 are fed back to the chromatography mechanism model of the virtual entity, and the behavior strategy of the chromatography process is determined by calculation through the chromatography mechanism model to obtain the optimized elution and collection behavior strategy.

[0189] The purity of the target protein is used as a parameter and the maximum protein yield is used as the objective function for optimization. The behavior strategy of the chromatography process is optimized to determine the optimal isocratic or gradient elution method, collection interval, and maximum protein yield that meets the purity requirements.

[0190] The protein yield Yield is calculated as follows:

[0191]

[0192] Among them, m load is the total protein loading, in mol / L;

[0193] RT is the protein elution retention time, in seconds;

[0194] t start The collection start time, in seconds;

[0195] t end The end time of collection, in seconds;

[0196] c i (t) is the outlet concentration of the i-th protein at time t;

[0197] In formula (3), the numerator is the sum of all protein masses in the collected protein.

[0198] Furthermore, the protein yield before and after optimization can be predicted by model simulation or obtained through experiments. Comparing the predicted value with the experimental value can better evaluate the simulation prediction ability of the model. Generally speaking, the experiment mainly obtains the c value of each protein at each time point through the method of separate collection and high-performance liquid chromatography detection. i (t), substitute it into equation (3) to calculate.

[0199] Purity of the i-th protein component i , the corresponding calculation formula is formula (4):

[0200]

[0201] Among them, t start To start collecting time;

[0202] t end To end the collection time;

[0203] ci (t) is the outlet concentration of the i-th protein at time t.

[0204] In formula (4), the numerator is the mass of the i-th protein component in the collected protein, and the denominator is the sum of the masses of all proteins in the collected protein. The division of the two is the purity of component i.

[0205] Furthermore, the purity before and after optimization can be predicted by model simulation or obtained through experiments. Comparing the predicted value with the experimental value can better evaluate the simulation prediction ability of the model. Generally speaking, the experiment mainly obtains the c value of each protein at each time point through the method of separate collection and high-performance liquid chromatography detection. i (t), and substitute it into equation (4) to calculate.

[0206] The method for optimizing the collection interval further comprises:

[0207] Traverse all collection starting points and collection end points to obtain the corresponding protein purity and yield;

[0208] The data points that meet the purity requirements are screened, and the maximum value among the screened data points is the maximum yield that meets the protein purity requirements;

[0209] Determine the collection start point and collection end point corresponding to the maximum yield.

[0210] Furthermore, an exhaustive method is used to traverse all collection starting points and collection end points to obtain the corresponding protein purity and yield.

[0211] In this embodiment, the optimization method of the isocratic or gradient elution method further comprises:

[0212] Predict protein elution peak shape under isocratic or gradient conditions through chromatography mechanism model simulation;

[0213] Optimize the collection interval by protein elution peak shape to obtain the maximum yield that meets protein purity requirements;

[0214] The optimization was performed with the maximum yield as the optimization goal, and the isocratic or gradient elution conditions at the maximum yield were obtained.

[0215] Furthermore, taking the maximum yield as the optimization goal, a genetic algorithm is used for optimization to obtain the isocratic or gradient elution conditions at the maximum yield.

[0216] The genetic algorithm used to solve nonlinear optimization problems is basically the same and will not be repeated here.

[0217] Step S3: The physical chromatography device completes the corresponding operation process according to the received behavior strategy instruction.

[0218] In this embodiment, the physical chromatography device receives the behavior strategy instruction and controls the chromatography process in real time to achieve the target purity or yield requirement.

[0219] The model optimization results include the process parameters of the chromatography equipment. The virtual entity uses the chromatography mechanism model to optimize and determine the behavior strategy of the chromatography process, and uses remote communication to transmit the behavior strategy instructions to the physical chromatography equipment to control the physical chromatography equipment in real time.

[0220] Although the above methods are illustrated and described as a series of actions for simplicity of explanation, it should be understood and appreciated that these methods are not limited by the order of the actions, because according to one or more embodiments, some actions may occur in a different order and / or concurrently with other actions from those illustrated and described herein or not illustrated and described herein but understandable to those skilled in the art.

[0221] The control method of the biological product chromatography process based on the digital twin technology proposed in the present invention is described in detail through the specific experiment of Example 1.

[0222] Example 1

[0223] The dynamic parameters of the chromatography mechanism model were obtained through fitting of protein chromatography experiments.

[0224] Poros XS 50 filler from Thermo Fisher Scientific was used, and an Omnifit chromatography column was used. The inner diameter of the chromatography column was 1 cm, the column packing height was 18.5 cm, and the column volume was 14.5 mL. Monoclonal antibody protein A was used to perform two sets of protein chromatography experiments.

[0225] Figure 3 The HPLC test result diagram of the protein A sample solution according to one embodiment of the present invention is disclosed, as shown in FIG. Figure 3 As shown, according to HPLC (high performance liquid chromatography) analysis, protein A roughly includes the following components: acid peak-1 (acidic peak), acid peak-2 (acidic peak), main peak, base peak-1 (alkaline peak), base peak-2 (alkaline peak).

[0226] The protein chromatography experiment consists of two groups. The parameters of the chromatography mechanism model are determined through experiments one and two. The basic steps of experiments one and two are divided into the following four steps:

[0227] 1) Pre-equilibrium, the equilibrium solution is phosphate buffer A, the composition is 6.28mM NaH2PO4, 3.74mM Na2HPO4 and 70.3mM NaCl, pH is 6.59, the conductivity is 8.538mS / cm, the equilibrium volume is 3CV, and the retention time is 5min;

[0228] 2) Loading the sample, and the concentration of the sample solution was detected to be 10.7 g / L;

[0229] In experiment 1, the sample loading was 10 g / L, and the calculated sample volume was 13.5 mL;

[0230] In experiment 2, the sample loading was 30 g / L, and the calculated sample volume was 40.5 mL;

[0231] The sample retention time was 5 min.

[0232] 3) Post-equilibrium, buffer A; equilibrium volume 3CV, retention time 5min;

[0233] 4) Elution: Use equilibrium buffer A to elute phase A, and use buffer B to elute phase B. Buffer B consists of 5 mM NaH2PO4, 5 mM Na2HPO4 and 256.6 mM NaCl, with a pH of 6.578 and a conductivity of 26.549 mS / cm;

[0234] Experiment 1: elution gradient was 40%-80% buffer B, 10CV;

[0235] Experiment 2: elution gradient was 40%-80% buffer B, 15 CV;

[0236] The eluate was collected in separate tubes, 2.4 mL in each tube.

[0237] 5) Use HPLC to detect the protein component ratio of each tube of collected liquid. Collect data such as Figure 2 shown.

[0238] like Figure 3 As shown, since acid peak-1 and acid peak-2 are difficult to distinguish in the high performance liquid chromatography detection spectrum, the two acid peaks need to be added together.

[0239] Figure 4a The results of the detection of each component in Experiment 1 according to an embodiment of the present invention are disclosed. Figure 4b The detection results of each component of Experiment 2 according to an embodiment of the present invention are disclosed. Points of different shapes represent different protein components. The spatial mass action (SMA) model is used for calculation and fitting. Figure 4a and Figure 4b The component detection curve of the model equation is:

[0240]

[0241] The genetic algorithm is used for fitting, and the minimum residual is used as the fitting target. The fitting results are as follows: Figure 5a and Figure 5b shown.

[0242] The model parameters of the fitting curve are shown in Table 1:

[0243] Table 1 shows the model parameter fitting results for different protein components

[0244]

[0245] Figure 5a A comparison result diagram of the fitting results of each component and the actual detection results of Experiment 1 according to an embodiment of the present invention is disclosed. Figure 5b The comparison result diagram of the fitting results and the actual detection results of each component of Experiment 2 according to an embodiment of the present invention is disclosed, such as Figure 5a and Figure 5b As shown, the line is the fitting result, and the points are the actual detection results. The fitting root mean square error of Experiment 1 is 1.5%, and the fitting root mean square error of Experiment 2 is 1.7%. The root mean square errors are both less than 5%. Therefore, it is believed that the above model parameters can be used for subsequent process optimization and process control, so as to obtain the chromatography mechanism model parameters as shown in Table 1.

[0246] Protein elution collection condition optimization and elution condition optimization.

[0247] The chromatography mechanism model parameters obtained above are applied to the chromatography mechanism model, and the protein-related parameters are obtained and transmitted to the chromatography mechanism model for optimization of protein elution and collection conditions.

[0248] The behavior strategy determined by model optimization is the collection interval, that is, the collection start point and collection end point of elution. The optimization method of the collection interval has been introduced in detail.

[0249] In this Example 1, protein yield and acid peak purity are used as optimization targets of the elution strategy. For example, the acid peak purity is required to be <13%, so it is necessary to simulate the peak distribution of each protein component under fixed elution conditions.

[0250] Figure 6 A schematic diagram of protein elution curve simulation prediction according to an embodiment of the present invention is disclosed. Figure 6 The protein elution curve simulation prediction of the elution condition optimization experiment shown is shown, and the dotted lines are the sample collection start point and collection end point. The solid lines of different colors are the predicted peak distribution of different component proteins. Using the chromatography mechanism model parameters obtained in step S1, the simulation is performed under the conditions of gradient elution of 10%-80% B, 15CV, and a sample load of 25g / L, and the peak distribution of all components in the elution peak can be predicted and calculated.

[0251] The protein yield and acid peak purity were calculated for all combinations of collection start and end points. For example, at 35 min from the collection start point and 50 min from the collection end point, the total protein yield was calculated by formula (3):

[0252]

[0253] At the same time, the purity of the acidic protein can be calculated according to the peak shape by formula (4):

[0254]

[0255] At this time, the purity of the acidic protein is 18.1%, which is greater than 13%, so it does not meet the separation requirements. According to the same steps to calculate the collection conditions of 40 minutes at the collection start point and 60 minutes at the collection end point, the total protein yield is 42% and the purity is 10%, which meets the acid peak purity requirements.

[0256] According to the method described above, the points with acid peak purity <13% are screened out, and the point with maximum yield is selected among these points, which corresponds to the optimized collection interval.

[0257] exist Figure 3 In the embodiment shown, through the optimization process, the start collection time is 42 minutes, the end collection time is 65 minutes, and the maximum yield corresponding to the acid peak requirement of less than 13% is 68%.

[0258] The above collection interval optimization method can be used to optimize the elution conditions: the optimization parameters are B% at the elution starting point and B% at the elution end point, and the optimization objective function is the maximum protein yield that meets the purity requirements.

[0259] The process of solving the maximum yield under this gradient is as follows:

[0260] The peak distribution of each protein component in the elution peak under the gradient was simulated by the model, and the elution collection conditions were optimized to obtain the maximum yield of the target protein under the gradient.

[0261] The optimal elution strategy for protein A at a sample loading of 25 g / L was obtained through genetic algorithm optimization, and the maximum yield was 70%.

[0262] After optimization, the maximum yield of 47% under the original elution gradient of 40%-80% B was increased by 49%.

[0263] Experiments have shown that when the protein loading capacity is 25 g / L, the contents of the acid peak, main peak, base peak-1 and base peak-2 of each protein component are 18%, 42%, 21% and 11% respectively, and the optimized gradient is 14%-67% B. The eluate is collected and tested according to the optimized elution interval, and the acid peak purity is 10%, the yield is 67%, the acid peak meets the collection standard and is only 3% different from the predicted maximum yield.

[0264] In summary, Example 1 verifies the control method of the biological product chromatography process based on digital twin technology proposed in the present invention. The behavioral strategies of the elution interval and the collection interval obtained by optimizing the chromatography mechanism model have high feasibility and reliability.

[0265] In order to realize the control method of the biological product chromatography process based on the digital twin technology, the present invention proposes a control system for the biological product chromatography process based on the digital twin technology, and the control system for the biological product chromatography process is a set of combined equipment.

[0266] A control system for a biological product chromatography process based on digital twin technology according to an embodiment of the present invention comprises a server 72 connected to a chromatography device 73, for receiving data collected by the chromatography device 73, and executing the above-mentioned control method for a biological product chromatography process based on digital twin technology, and transmitting the determined behavior strategy instructions to the chromatography device 73;

[0267] The chromatography device 73 is used to collect data of the biological product chromatography process and transmit it to the server, and is used for chromatographic separation of the biological product, and performs chromatography of the biological product based on the behavioral strategy instructions transmitted by the server 72.

[0268] Another embodiment of the present invention is a control system for a biological product chromatography process based on digital twin technology, such as Figure 7 As shown, the present invention proposes a control system for a biological product chromatography process based on digital twin technology, comprising:

[0269] A sensor device 71 connected to a server 72 for collecting data of a biological product chromatography process and transmitting the collected data to the server 72;

[0270] The server 72 is connected to the chromatography device 73, and is used to receive the data transmitted by the sensor device 71, and execute the control method of the biological product chromatography process based on the digital twin technology, and transmit the determined behavior strategy instructions to the chromatography device 73;

[0271] The chromatography device 73 is used for chromatographic separation of biological products, and performs chromatography of biological products based on the behavioral strategy instructions transmitted by the server 72.

[0272] The data of the biological product chromatography process include, but are not limited to, loading time, collection time, collection volume, gradient starting point, gradient end point, elution volume, loading amount, conductivity, UV absorbance, elution peak shape change, and biological product property data (protein properties and protein changes, such as alkaline peak, acidic peak, main peak, etc.).

[0273] In this embodiment, the sensor device 71 can be a high performance liquid chromatograph (HPLC), which can realize automatic online sampling, online dilution, online detection, and real-time data processing (integration) and transmission of the fully automatic chromatograph. In other embodiments, the sensor device is a multi-angle light scattering sensor (MALS), a UV-VIS absorption sensor, a fluorescence sensor, an infrared absorption sensor (IR), an attenuated total reflection-Fourier transform infrared spectroscopy sensor (ATR-FTIR), a refractive index (RI) sensor, a pH sensor, a temperature sensor, a conductivity sensor, a pressure sensor, a small angle X-ray scattering (SAXS) sensor, and a redox sensor, as well as a device comprising a combination of these sensors.

[0274] The sensor device 71 receives the control command from the server 72, performs the start and stop control of the HPLC through the single chip microcomputer, transmits data through the network cable, and feeds back the result data to the server 72.

[0275] In this embodiment, the chromatography device 73 is a fully automatic chromatography instrument. pure(Cytiva life science).

[0276] The chromatography equipment 73 can be remotely controlled by the server 72 to modify the process parameters, automatically realize the chromatography operation, and provide real-time feedback of the current chromatography instrument status and instrument detection values ​​to the server 72. The instrument detection values ​​include but are not limited to ultraviolet absorption value, conductivity value and pH value.

[0277] The server 72 is equipped with a chromatography mechanism model and can perform data collection, mathematical calculation optimization, and process control of the chromatography equipment 73 and the sensor equipment 71 .

[0278] The server 72 controls the operation parameters to change the chromatography device 73 , and receives the chromatography status fed back by the chromatography device 73 .

[0279] The chromatography device 73 and the server 72 can exchange data via the standard communication protocol OPC (OLE for process control), and the exchanged data includes but is not limited to state transfer, method start and stop control data, behavior strategy control data, and collection interval data.

[0280] The server 72 uses the optimized behavior strategy and remote communication to adjust the process parameters of the chromatography equipment 73 in real time.

[0281] The chromatography process includes but is not limited to a sample loading process, a washing process, an elution process and a product collection process, and the process parameters include but are not limited to collection time, collection volume, gradient start, gradient end point and elution volume.

[0282] The control system for the chromatography process of biological products based on digital twin technology proposed in the present invention builds a digital twin system including chromatography equipment, sensor equipment and server, introduces a chromatography process mechanism model, and realizes the closed-loop control logic of online detection-model prediction-online control through communication between equipment. The chromatography process does not require bagging and waiting for detection, and the optimal yield is obtained under the purity conditions. There is no need to conduct a large number of experiments for fitting in the early stage, and there is no need for a large number of experiments to update the model after replacing the system, and the detection and calculation speed requirements are relatively low.

[0283] The following two embodiments (Example 2 and Example 3) further illustrate the operation example of the control system of the biological product chromatography process based on digital twin technology proposed in the present invention in the A protein ion exchange chromatography (CEX) separation.

[0284] Example 2

[0285] Here, the gradient 14%-67% B obtained by optimization in Example 1 was selected as the model dynamic parameter, a chromatography mechanism model was established, and a control system for the chromatography process of biological products based on digital twin technology was operated.

[0286] The protein loading capacity is 30 g / L, the contents of acid peak, main peak, base peak-1, and base peak-2 are unknown, and the operation process of the chromatography process control system is as follows:

[0287] Start the chromatograph via the server pure's CEX method automatically controls the detection equipment HPLC to perform online sampling, online dilution and running of the detection method during the sample loading process.

[0288] when pure is still running in the loading-post-equilibrium process, HPLC completes the detection and automatically integrates the detection results to obtain the contents of the acid peak, main peak, base peak-1 and base peak-2. In this embodiment, the contents of the four components are 25%, 38%, 21% and 11%, respectively, and the data is transmitted to the server.

[0289] The data show that the acid and base peak contents of the sample and the sample used for model parameter fitting have changed significantly.

[0290] After receiving the data, the server substitutes the content of each protein component into the model.

[0291] Figure 8a The first predicted elution result diagram of the chromatography process control system according to one embodiment of the present invention is disclosed. The straight line is the peak type distribution of each component, and the dotted line is the total peak type prediction (sum of the peak types of each component). Figure 8a The predicted elution peak shape is shown and the optimal collection interval is calculated.

[0292] The calculated collection start point is 47.5 min and the end point is 75 min. When the linear gradient elution starts, the collection instruction is fed back to In pure, pure collects the eluate according to the collection interval.

[0293] Figure 8b The first actual elution result diagram of the chromatography process control system according to one embodiment of the present invention is disclosed, which is the actual elution UV detection peak type, such as Figure 8b The actual elution profile and collection interval shown are similar to the predicted overall peak shape.

[0294] The yield and purity of the collected samples were tested and compared with the predicted results.

[0295] The predicted results are: acid peak purity 13%, yield 26.8%, and the actual detection results are acid peak purity 11.7%, yield 24.8%. The purity and yield errors are less than 2%, and the digital twin runs successfully, indicating that the control system for the chromatography process of biological products based on digital twin technology proposed in the present invention meets the precise control requirements of the chromatography process.

[0296] Example 3

[0297] Here, the gradient 14%-67% B obtained by optimization in Example 1 was selected as the model dynamic parameter, a chromatography mechanism model was established, and a control system for the chromatography process of biological products based on digital twin technology was operated.

[0298] The protein loading is 40 g / L, the contents of acid peak, main peak, base-1, and base-2 are unknown, and the operation process of the chromatography process control system is as follows:

[0299] Start the chromatograph via the server pure's CEX method automatically controls the detection equipment HPLC to perform online sampling, online dilution and running of the detection method during the sample loading process.

[0300] when Pure is still running in the loading-post-equilibrium process, HPLC completes the detection, and automatically integrates the detection results to obtain the contents of acid peak, main peak, base peak-1 and base peak-2. In this embodiment, the contents of the four components are: 29%, 35%, 19% and 10%, respectively, and the data is transmitted to the server. From the data, it can be seen that the acid and base peak contents of the sample and the sample used for model parameter fitting and Example 2 have changed significantly. And the loading capacity is higher than the two.

[0301] After receiving the data, the server substitutes the content of each protein component into the model.

[0302] Figure 9aThe second predicted elution result diagram of the chromatography process control system according to one embodiment of the present invention is disclosed. The straight line is the peak type distribution of each component, and the dotted line is the total peak type prediction (sum of the peak types of each component). Figure 9a The predicted elution peak shape was used to calculate the optimal collection interval.

[0303] The calculated collection start point is 49.5 min and the end point is 75 min. When the linear gradient elution starts, the collection instruction is fed back to In pure, pure collects the eluate according to the collection interval.

[0304] Figure 9b The second actual elution result diagram of the chromatography process control system according to one embodiment of the present invention is disclosed, which is the actual elution UV detection peak type, such as Figure 9b The actual elution profile and collection interval shown are similar to the predicted overall peak shape.

[0305] The yield and purity of the collected protein were tested and compared with the predicted results.

[0306] The predicted results are: acid peak purity 13%, yield 17%; the actual detection results are acid peak purity 12.2%, yield 15%, the purity and yield errors are less than 2%, and the digital twin runs successfully, indicating that the control system for the chromatography process of biological products based on digital twin technology proposed in the present invention meets the precise control requirements of the chromatography process.

[0307] As can be seen from the above Examples 2 and 3, the control system for the biological product chromatography process based on digital twin technology proposed in the present invention can predict the peak shape change when the sample loading and the ratio of each protein component change, and then automatically optimize the collection interval and control the fully automatic chromatograph. pure automatically collects the eluted samples that meet the separation requirements.

[0308] The present invention also provides a control device that can realize the above-mentioned biological product chromatography process based on digital twin technology.

[0309] Fig.10 The principle block diagram of a control device for a biological product chromatography process based on digital twin technology according to an embodiment of the present invention is disclosed, such as Fig.10 As shown, the control device for the biological product chromatography process based on the digital twin technology proposed by the present invention includes at least a processor 101 and a memory 102.

[0310] The processor 101 can make judgments and issue prompts. In some embodiments, the processor 101 can be composed of one or more processors.

[0311] The memory 102 stores various data files used for computing and / or communication, as well as possible programs executed by the processor 101, and also stores related data, including acquisition data (such as sensor acquisition data), real-time operation data and historical operation data of the chromatography operation process, biological product property data, model data (such as model prediction data, model parameter correction data) and behavior strategy data, etc.

[0312] The processor 101 executes these instructions to implement the main part of the method.

[0313] For example, the implementation process file of the above-mentioned control method of the biological product chromatography process based on digital twin technology can be a computer program, which is stored in a memory and can be recorded in a processor for execution to implement the method of the present application.

[0314] When the implementation process file of the control method of the biological product chromatography process based on digital twin technology is a computer program, it can also be stored in a computer-readable storage medium as a product. For example, a computer-readable storage medium may include, but is not limited to, a magnetic storage device (e.g., a hard disk, a floppy disk, a magnetic strip), an optical disk (e.g., a compact disk (CD), a digital versatile disk (DVD)), a smart card, and a flash memory device (e.g., an electrically erasable programmable read-only memory (EPROM), a card, a stick, a key drive). In addition, the various storage media described herein can represent one or more devices and / or other machine-readable media for storing information. The term "machine-readable medium" may include, but is not limited to, wireless channels and various other media (and / or storage media) that can store, contain and / or carry code and / or instructions and / or data.

[0315] The present invention proposes a control method, system, device for a biological product chromatography process based on digital twin technology, and a digital twin system for a biological product chromatography process, which specifically have the following beneficial effects:

[0316] 1) Use digital twins based on chromatography mechanism models to guide real-time closed-loop optimization of the chromatography process, flexibly respond to fluctuations in upstream process parameters, and increase the robustness and efficiency of the chromatography process;

[0317] 2) By organically combining chromatography equipment, sensor equipment and servers, the accurate prediction and optimization methods of the chromatography mechanism model are applied to actual production practice, and the entire process is fully automated, saving a lot of manpower and material resources;

[0318] 3) The control method is universal and applicable to the chromatography process in downstream separation and can be extended to the chromatography separation process of non-protein substances.

[0319] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0320] Those skilled in the art will appreciate that information, signals, and data may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips cited throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or optical particles, or any combination thereof.

[0321] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. The technician may implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention.

[0322] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or executed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration.

[0323] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read and write information from / to the storage medium. In an alternative, a storage medium may be integrated into a processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside in a user terminal as discrete components.

[0324] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0325] The above embodiments are provided for persons familiar with the art to implement or use the present invention. Personnel familiar with the art can make various modifications or changes to the above embodiments without departing from the inventive concept of the present invention. Therefore, the protection scope of the present invention is not limited to the above embodiments, but should be the maximum scope of the innovative features mentioned in the claims.

Claims

1. A control method for a biological product chromatography process based on digital twin technology, characterized in that: include: Collect real-time data of the operation process of physical chromatography equipment and / or biological product property data; Through the pre-established virtual entity bidirectionally mapped with the physical chromatography equipment, the operation process of the physical chromatography equipment is simulated and predicted according to the real-time data, the behavior strategy of the chromatography process is determined and the behavior strategy instructions are generated; The physical chromatography device completes the corresponding operation process according to the received behavior strategy instructions.

2. The method for controlling the chromatography process of biological products based on digital twin technology according to claim 1, characterized in that: The virtual entity establishes a bidirectional mapping with the physical chromatography device, establishes a model based on the operation data of the physical chromatography device and / or the property data of the biological product, simulates the operation process of the physical chromatography device, and performs simulation analysis and optimization on the operation process of the physical chromatography device.

3. The control method of the biological product chromatography process based on digital twin technology according to claim 1 or 2, characterized in that: The steps of establishing the virtual entity are: Based on the real-time and / or historical operation data of the physical chromatography equipment and the property data of the biological product, the model dynamic parameters of the chromatography process are defined and the model is established.

4. The method for controlling the chromatography process of biological products based on digital twin technology according to any one of claims 1 to 3, characterized in that: The virtual entity establishment step includes: Based on at least two groups of real-time and / or historical operation data of the physical chromatography equipment operation process of the biological product chromatography process and the biological product property data, combined with the pre-set fitting target, the model dynamic parameters of the chromatography process are defined and the model is established.

5. The method for controlling the chromatography process of biological products based on digital twin technology according to any one of claims 1 to 4, characterized in that: Also includes: Based on the real-time data of the operation process of the physical chromatography equipment of the biological product chromatography process and the biological product property data, the dynamic parameters of the model are adjusted in real time to optimize the model through evaluation between the real-time data and the simulation prediction results.

6. The method for controlling the chromatography process of biological products based on digital twin technology according to any one of claims 1 to 5, characterized in that: The model includes a chromatography mechanism model; The chromatography mechanism model includes an intra-column mechanism model and / or an adsorption mechanism model; Wherein, the in-column mechanism model is a convection and diffusion mass transfer model of protein in the chromatography column; The adsorption mechanism model is a model of protein adsorption and dissociation on the surface of the chromatographic medium.

7. The method for controlling the chromatography process of biological products based on digital twin technology according to claim 6, characterized in that: The in-column mechanism models include an equilibrium model (Equilibrium model), an equilibrium diffusion model (Equilibriumdispersive model, EDM), a lumped kinetic model (Lumped kinetic model, LKM), a shrinking core model (Shrinking core model) and a general rate model (General rate model, GRM).

8. The method for controlling the chromatography process of biological products based on digital twin technology according to claim 6, characterized in that: The adsorption mechanism models include linear driving force model (LDF), Langmuir adsorption isotherm model, Freundlich adsorption isotherm model, Anti-Langmuir adsorption isotherm model, steric mass action model (SMA), multi-state steric mass action model (Multi-State StericMass Action Model) and colloidal particle model (CPA).

9. The method for controlling the chromatography process of biological products based on digital twin technology according to any one of claims 1 to 8, characterized in that: The virtual entity predicts the operation process of the physical chromatography equipment through real-time data simulation, and determines and / or optimizes the behavior strategy of the chromatography process according to the set single objective or multiple objectives, and / or constraint parameters; The single or multiple objectives include bioproduct component selection, bioproduct yield, and bioproduct distribution; The constraint parameters include bioproduct yield constraint, impurity ratio constraint, bioproduct component constraint and component distribution constraint.

10. The method for controlling the chromatography process of biological products based on digital twin technology according to any one of claims 1 to 9, characterized in that: The virtual entity predicts the operation process of the physical chromatography equipment through real-time data simulation and / or adjusts the simulation prediction results in real time, and determines and / or optimizes the behavior strategy of the chromatography process according to specific biological product components and yields, impurity ratios, and biological product distribution constraint parameters.

11. The method for controlling a biological product chromatography process based on digital twin technology according to any one of claims 1 to 10, characterized in that: The behavior strategy includes node control parameters of the chromatography process; The node control parameters of the chromatography process include the loading time point, the elution collection starting point, the elution collection end point, and the elution volume.

12. The method for controlling a biological product chromatography process based on digital twin technology according to any one of claims 1 to 11, characterized in that: The biological products include antibody proteins, enzymes and peptides, fusion proteins, structural proteins, regulatory proteins, growth factors, hormones and cytokines, nucleic acid molecules, compounds, and compound-protein conjugates.

13. A digital twin system for a biological product chromatography process, characterized in that: include: Physical chromatography equipment is the physical application scenario of the digital twin system, including the physical equipment for bioproduct chromatography and / or the physical environment of the physical equipment; A multi-sensor data acquisition module, which is disposed on or connected to the physical chromatography device and is used to obtain real-time operation data of the physical chromatography device and / or biological product property data; The virtual entity establishes a bidirectional mapping relationship with the physical chromatography device, and determines a behavior strategy based on the received real-time data of the operation process of the physical chromatography device and / or the biological product property data; The control service module generates corresponding control instructions according to the behavior strategy determined by the virtual entity, and transmits the control instructions to the physical chromatography equipment for operation process control.

14. The digital twin system for the chromatography process of biological products according to claim 13, characterized in that: The virtual entity includes: The signal domain is used to receive and transform the operation data of the physical chromatography equipment and / or the property data of the biological product acquired by the multi-sensor data acquisition module, and provide a data basis for the simulation domain; A simulation domain, establishing a bidirectional mapping model with the physical chromatography device according to the operation data of the physical chromatography device and / or the property data of the biological product; The optimization domain uses a model that is bidirectionally mapped with the physical chromatography equipment to optimize and determine the behavior strategy based on real-time data and / or physical goals set by the chromatography process; and / or, The storage domain stores the operation data, property data and the optimized data of the optimization domain in the database to facilitate the access of the control service module.

15. The digital twin system for the chromatography process of biological products according to claim 13 or 14, characterized in that: In the simulation domain, a model of the physical chromatography device is established, and the physical chromatography device is simulated and predicted in real time according to real-time data and / or the simulation prediction is adjusted in real time, and the behavior strategy is determined through the optimization domain.

16. The digital twin system for the chromatography process of biological products according to any one of claims 13 to 15, characterized in that: The model bidirectionally mapped with the physical chromatography device includes a chromatography mechanism model; preferably, the chromatography mechanism model includes an intra-column mechanism model and / or an adsorption mechanism model.

17. The digital twin system for a bioproduct chromatography process according to any one of claims 13 to 16, characterized in that: The multi-sensor data acquisition module includes a high performance liquid chromatograph (HPLC), a multi-angle light scattering sensor (MALS), a UV-VIS absorption sensor, a fluorescence sensor, an infrared absorption sensor (IR), an attenuated total reflection-Fourier transform infrared spectroscopy sensor (ATR-FTIR), a refractive index (RI) sensor, a pH sensor, a temperature sensor, a conductivity sensor, a pressure sensor, a small angle X-ray scattering (SAXS) sensor and a redox sensor or a combination thereof.

18. The digital twin system for the chromatography process of biological products according to any one of claims 13 to 17, characterized in that: The real-time data includes loading time, collection time, collection volume, gradient starting point, gradient end point, elution volume, loading amount, conductivity, UV absorbance, and elution peak type change; The biological product property data include the properties of the protein components themselves and / or the changes in the protein during the chromatography process.

19. A control system for a biological product chromatography process based on digital twin technology, characterized in that: include, A server, configured to receive the collected relevant data, and execute the control method for the chromatography process of a biological product based on the digital twin technology according to any one of claims 1 to 12, and transmit the behavior strategy instructions determined according to the collected relevant data to the chromatography equipment; Chromatography equipment, used to collect data of the biological product chromatography process and transmit it to the server, and to separate the biological product by chromatography, and to perform the biological product chromatography based on the behavioral strategy instructions transmitted by the server; Alternatively, the control system additionally includes a sensor device for collecting data of other biological product chromatography processes and transmitting the data to the server.

20. The control system for the biological product chromatography process based on digital twin technology according to claim 19, characterized in that: Data exchange between chromatography equipment and server; The interaction data includes: state transfer data, method start and stop control data, behavior strategy control data and adjustment data, and collection interval data.

21. A control device for a biological product chromatography process based on digital twin technology, characterized in that: include A memory for storing instructions and related data executable by the processor; A processor, configured to execute the instructions to implement the method according to any one of claims 1 to 12.

Images