Manufacturing condition management method, manufacturing condition management program, and manufacturing condition management system

Abstract

Provide are a manufacturing condition management method and similar with which a highly accurate design design space can be efficiently created while taking various factors and parameters into consideration. This manufacturing condition management method has: a first extraction step (S12) for extracting quality characteristic candidates or similar pertaining to quality in the manufacture of an objective finished good; a first designation step (S13) for designating a quality characteristic or similar under consideration, on the basis of evidence with regard to the extracted quality characteristic candidates or similar; a discrimination step (S14) for discriminating between a normal state and an abnormal state pertaining to a manufacturing step by which a product is manufactured; a selection step (S15) for selecting a timing at which to create a design space (DS) within the manufacturing step; a second extraction step (S16) for extracting candidates for an important step parameter that is a cause of the designated quality characteristic or similar; a second designation step (S17) for designating an important step parameter from among the candidates for the important step parameter; a creation step (S18) for determining a range for the important step parameter and creating the DS; and a determination step (S19) for determining a manufacturing condition on the basis of the DS.

Description

Methods for managing manufacturing conditions, programs for managing manufacturing conditions, and systems for managing manufacturing conditions

[0001] This invention relates to a method for managing manufacturing conditions, a program for managing manufacturing conditions, and a system for managing manufacturing conditions.

[0002] This invention relates to the control of manufacturing conditions in the manufacturing processes of pharmaceuticals, foods, cosmetics, chemicals, and fuels using culture vessels, bioreactors, reaction vessels, and reaction apparatus. Methods for culturing cells such as plants, microorganisms, and animals to produce useful substances are used in various industries, including pharmaceuticals, foods, cosmetics, chemicals, and fuels. For example, biopharmaceuticals, including antibody drugs, contain substances produced by animal cells as their main component. Such substances can also be obtained by culturing animal cells, plants, or microorganisms and separating and purifying the target substance accumulated in the cells or secreted into the culture medium. Furthermore, human mesenchymal stem cells, for example, are used as therapeutic agents when cultured.

[0003] For any of the aforementioned products to be profitable and commercially viable, it is necessary to identify a range of manufacturing conditions that meet a certain level of quality and productivity, and to manage production within that range. Here, manufacturing conditions refer to the specifications of the manufacturing equipment and apparatus (capacity, structure, function, performance, etc.) and operating conditions (temperature, pH, pressure, osmotic pressure, aeration rate, stirring speed, fluid force, mixing capacity, etc. in the reactor). Traditionally, to determine the range of manufacturing conditions, an exploratory or test facility on a small scale is used to search for operating conditions that meet the desired quality and productivity. After appropriate operating conditions are found, the specifications and operating conditions of the manufacturing equipment and apparatus are then searched for to maintain the desired quality and productivity on a commercial scale (scale-up).

[0004] In scale-up, methods based on similarity principles are frequently used. For example, in manufacturing processes involving stirring, the process is generally carried out using geometric similarity (the ratio of the dimensions of the stirring device's structure and arrangement is all similar) and a criterion of keeping the stirring power per unit volume of liquid constant. Specifically, the tank is designed so that performance requirements such as stirring power, power consumption per unit volume of liquid, impeller rotation speed, impeller diameter, liquid discharge speed, liquid circulation speed within the reactor, impeller tip speed, and Reynolds number remain constant before and after scale-up. However, it is difficult to scale up while simultaneously satisfying all performance requirements. Therefore, it is possible that the most important performance requirements for stirring (performance requirements that affect productivity and quality) are met, but the manufacturing conditions after scale-up do not meet the desired quality and productivity.

[0005] In contrast, in recent years, manufacturing process development based on Quality by Design (QbD) has been recommended in pharmaceutical manufacturing. QbD is a strategic approach for development and manufacturing processes, and is a design concept to ensure that the final product reliably meets planned values ​​in terms of both purity and efficacy. As part of the QbD strategy, the design space is defined in the ICH Q11: Development and Manufacturing of Active Pharmaceutical Ingredients (APIs) Guidelines of the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) of Japan, the US, and the EU, which defines it as "a multidimensional combination and interaction of input variables (such as the properties of raw materials) and process parameters that have been proven to ensure quality in pharmaceutical manufacturing," and states that "operating within this design space is not considered a change." The design space examines the interaction of multiple critical process parameters, which are manufacturing conditions corresponding to critical quality characteristics that affect the quality of pharmaceuticals, and clarifies the area in which it is possible to manufacture.

[0006] Attempts are being made to apply this concept to scale-up. These attempts involve first determining the design space necessary to ensure sufficient quality on a small scale using experiments and theoretical formulas, and then using that design space to determine the minimum quality requirements for the scaled-up version. Generally, the culturing range in large-scale production tanks is narrower than that in small-scale tanks. This is equivalent to the difficulty of scaling up to simultaneously satisfy all performance requirements. To ensure minimum quality after scale-up, it is necessary to confirm the culturing range after scale-up at the small-scale stage. This method can be applied not only to pharmaceutical manufacturing but also to the production of other useful substances.

[0007] Figure 1 is an explanatory diagram illustrating the concept of scaling up a culture vessel using design space, one of the conventional technologies. Scaling up using design space involves evaluating the culturing range in laboratory experiments, focusing on the optimal operating conditions (optimal point) obtained in small-scale laboratory experiments, as shown in the left diagram of Figure 1. Subsequently, as shown in the right diagram of Figure 1, a vessel for the desired large-scale production is designed. However, a reduction in the culturing area is unavoidable, making it necessary to understand the appropriate conditions at the small-scale stage.

[0008] A technology utilizing design space is described, for example, in Patent Document 1. The technology described in Patent Document 1 aims to provide a real-time release test that can consistently guarantee quality and a product quality test method using the real-time release test. The technology described in Patent Document 1 is a method for designing a design space for the quality characteristics of an in-process product and / or a final product, characterized in that the design space is designed using only material properties as input variables.

[0009] International Publication No. 2013 / 008733

[0010] However, conventional design space creation methods for managing manufacturing conditions, including the technology described in Patent Document 1, have the following challenges: Various factors and parameters are involved in creating a design space, and creating a design space that comprehensively considers these factors relies on the know-how of skilled workers. There was a need for a method that could efficiently create a highly accurate design space while considering various factors and parameters.

[0011] The present invention has been made in view of the above circumstances. The object of the present invention is to provide a method for managing manufacturing conditions, a program for managing manufacturing conditions, and a management system for manufacturing conditions that can efficiently create a highly accurate design space while taking into account various factors and parameters.

[0012] The manufacturing condition management method according to the present invention, which solves the aforementioned problems, is characterized by comprising: a first extraction step of extracting candidate quality characteristics related to the quality in the manufacturing of a target product and / or candidate productivity characteristics related to the productivity in the manufacturing of a product; a first identification step of identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or productivity characteristics; a determination step of determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection step of selecting any one time within the manufacturing process to create a design space if it is in a steady state, and selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction step of extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification step of identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation step of determining the range of the identified important process parameters and creating a design space; and a determination step of determining manufacturing conditions based on the design space.

[0013] According to the present invention, it is possible to provide a method for managing manufacturing conditions, a manufacturing condition management program, and a manufacturing condition management system that can efficiently create a highly accurate design space while considering various factors and parameters. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments. Further features related to the present invention will become apparent from the description of this specification and the accompanying drawings.

[0014] It is an explanatory diagram for explaining the concept of scale-up of a culture tank by a design space, which is one of the prior arts. It is an explanatory diagram showing the procedure for managing manufacturing conditions. dP ｎ / dP ｍ When it is very large, the product quality index Q ０ A graph showing the relationship of a plurality of manufacturing conditions (P ｍ , P ｎ ) to satisfy. P ｎ =P ｍ ２ When, the inverse function P ｍ ≒±√P ｎ In, since P ｍ has two values, P ｍ cannot be said to be the inverse function of P ｎ (left figure), by setting the range of values of the inverse function to P ｍ >0, for P ｎ , the real number P ｍThis graph shows that it is possible to determine a unique value (right figure). This is an explanatory diagram illustrating the concept of inverse mapping. This is an explanatory diagram illustrating an example of an optimized solution with multiple variables. This is an explanatory diagram illustrating an example of the correlation between critical quality characteristics 171 and critical process parameters 173a to 173e in a conventional design space. This is an explanatory diagram illustrating an example of the correlation between critical quality characteristics 71, critical productivity characteristics 72, secondary critical process parameters 82, and primary critical process parameters 81. This is an explanatory diagram illustrating the relationship between the secondary design space 92 and the primary design space 91. This is an explanatory diagram illustrating an example of a correlation in cell culture. This is a flowchart illustrating a specific embodiment of a manufacturing condition management method according to one embodiment of the present invention. This is a flowchart illustrating the contents of a manufacturing condition management program 120 and a computer-readable recording medium 121 on which the manufacturing condition management program 120 is recorded according to one embodiment of the present invention. This is a schematic diagram illustrating a manufacturing condition management system 1300 according to one embodiment of the present invention. This is a schematic diagram illustrating a specific embodiment of a manufacturing condition management system 1300 according to one embodiment of the present invention. This is a flowchart illustrating an overview of the manufacturing process for producing antibody proteins. This is a graph showing an example of a growth curve and the change in antibody over time. This is a correlation diagram showing the correlation of each parameter in the example. This is a correlation diagram showing the correlation of secondary important step parameter 82 and primary important step parameter 81 in the example. This is a schematic diagram showing the configuration of a 1L culture device. This is a graph showing an overview of the change in DO over time when determining the oxygen consumption rate of cells. This is a schematic diagram showing the configuration of a shear stress effect evaluation device equipped with a flow chamber type culture device. This is an explanatory diagram showing an example of a design space created in the example. This is an explanatory diagram showing an example of a design space before and after scale-up. This is an explanatory diagram showing an example of a design space when the cell number density is different. This is an explanatory diagram showing the metabolic map of Corynebacterium.

[0015] To achieve the aforementioned objectives, a manufacturing condition management method, a manufacturing condition management program, and a manufacturing condition management system according to one embodiment of the present invention have the following configuration. First, the manufacturing condition management method according to one embodiment of the present invention will be described with reference to the drawings. The object, features, advantages, and ideas of the present invention will be apparent to those skilled in the art from the description herein, and those skilled in the art will be able to easily reproduce the present invention from the description herein. The specific embodiments of the invention described below are examples of preferred embodiments of the present invention and are provided for illustrative or explanatory purposes only, and do not limit the present invention to them. It will be apparent to those skilled in the art that various modifications and modifications can be made based on the description herein, within the intent and scope of the present invention as disclosed herein.

[0016] The embodiments described below are illustrative examples for illustrating the present invention, and have been omitted and simplified as appropriate for clarity of explanation. The present invention can be implemented in various other forms. Unless otherwise specified, each component may be singular or plural. The positions, sizes, shapes, and ranges of the components shown in the drawings may not represent the actual positions, sizes, shapes, and ranges, in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the positions, sizes, shapes, and ranges disclosed in the drawings.

[0017] When there are multiple components with the same or similar function, they may be described using the same symbol but with different subscripts. Furthermore, if it is not necessary to distinguish between these multiple components, the subscript may be omitted. Duplication of explanations for components with the same or similar function may be omitted.

[0018] In embodiments, processing performed by executing a program may be described. Here, the computer executes the program using a processor (e.g., a central processing unit (CPU), a graphics processing unit (GPU)) and performs processing defined by the program using memory resources (e.g., memory) and interface devices (e.g., communication ports). Therefore, the main entity performing the processing by executing the program may be the processor. Similarly, the main entity performing the processing by executing the program may be a controller, device, system, computer, or node having a processor. The main entity performing the processing by executing the program may be an arithmetic unit, and may include a dedicated circuit that performs a specific processing. Here, a dedicated circuit is, for example, an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), or a CPLD (Complex Programmable Logic Device).

[0019] The program may be installed on a computer from a program source. The program source may be, for example, a program distribution server or a computer-readable storage medium. If the program source is a program distribution server, the program distribution server includes a processor and storage resources for storing the program to be distributed, and the processor of the program distribution server may distribute the program to other computers. In addition, in some embodiments, two or more programs may be implemented as a single program, or one program may be implemented as two or more programs.

[0020] (Challenges in Creating a Design Space) The creation of a design space for managing the manufacturing conditions described above presents the following challenges: (Challenge 1) Products require both quality and productivity, but the design space created by QbD is a method for ensuring product quality and does not consider methods for ensuring productivity. (Challenge 2) In creating a design space, the range of manufacturing conditions (critical process parameters) that must be satisfied for critical quality characteristics is considered, but if there are many manufacturing conditions that correlate with critical quality characteristics, and there are also correlations within the manufacturing conditions themselves, it becomes necessary to examine the correlation with critical quality characteristics for all combinations of performance requirements of the manufacturing conditions, resulting in enormous time and cost for exploring the correlations. (Challenge 3) If the environment of the culture (reaction field) in the manufacturing process changes over time, it becomes necessary to create a design space for each change over time, which requires enormous time and cost. (Challenge 4) Since the manufacturing conditions to be considered span a large number of performance requirements, the number of correlation equations (theoretical equations, empirical equations) between the dependent variable and explanatory variables (critical process parameters) required for creating a design space becomes large, and deriving these correlation equations requires enormous time and cost. (Problem 5) When deriving the boundary of the design space from empirical and theoretical formulas, certain manufacturing conditions may have a large error compared to other manufacturing conditions, resulting in poor accuracy of the design space boundary. (Problem 6) When deriving the boundary of the design space, if the range of change or accuracy of one explanatory variable (critical process parameter) is significantly larger than that of other explanatory variables (critical process parameters), the solver or discretization calculation of the analysis software may not converge numerically, and it may not be possible to derive the boundary from the correlation formula.

[0021] (Procedure for Managing Manufacturing Conditions) In order to achieve these problems 1 to 6 which relate to the problems of the present invention, a method for managing manufacturing conditions according to one embodiment of the present invention (hereinafter sometimes referred to as "this management method") consists of the following steps (see Figure 2). Figure 2 is an explanatory diagram showing the procedure for managing manufacturing conditions. As shown in Figure 2, in step S1, a target product characteristic is set. In step S2, candidates for important quality characteristics and / or important productivity characteristics that may correlate with the target product characteristic are extracted. Important quality characteristics and important productivity characteristics will be described later. In step S3, important quality characteristics and / or important productivity characteristics are identified based on evidence such as experiments. In step S4, candidates for important process parameters that may correlate with important quality characteristics and / or important productivity characteristics are extracted. In step S5, important process parameters are identified based on evidence such as experiments, theoretical analysis, and simulations. In step S6, the process or unit operation group (for example, the first important process parameter and the second important process parameter described later) for creating the design space in the target manufacturing process is determined. For example, in step S6, the time in the manufacturing process in which the design space should be created is determined. Specifically, in step S6, the process determines whether the manufacturing process is in a steady state or a transient state, determines the time within the manufacturing process when a design space should be created if it is in a transient state, and determines one or more arbitrary time points if it is in a steady state. In step S7, the design space for the target manufacturing process or unit operation group is created. In step S8, the manufacturing conditions are determined based on the design space. Step S6 may be performed between steps S1 and S5. For example, step S6 may be performed between steps S3 and S4.

[0022] This control method sets the product characteristic targets as quality and productivity when implementing each of these steps (addressing Problem 1). Furthermore, this control method includes a step of selecting critical process parameters that correlate with quality and productivity (addressing Problem 2). This control method includes a step of selecting a group of unit operations that create a design space within the manufacturing process (addressing Problem 3). This control method includes a step of selecting and removing redundant critical process parameters (addressing Problem 4). In addition, this control method includes a boundary line using an inverse mapping in the design space (addressing Problems 5 and 6). The terms used in the description of this embodiment are defined as follows. The details of these steps are described below.

[0023] <Definition of Terms> Product Characteristics Objectives: Product characteristics objectives are the quality and productivity that a product must meet in the manufacturing process. The term "product" here is not limited to the finished product after manufacturing is complete, but also includes intermediates in the manufacturing process. In the case of intermediates, it refers to the quality and productivity of the intermediate required to proceed to the next process. Examples of product characteristics objectives include the function of the product (efficacy, side effects, reaction efficiency, etc.), the yield of the product, and the amount of yield produced.

[0024] Critical Quality Characteristics (Quality Characteristics): Critical quality characteristics refer to the physical, chemical, biological, and microbiological properties or characteristics of various components within the manufacturing process, including the product, that must be within appropriate limits, ranges, and distributions in order to guarantee the required quality of the product (quality in the product characteristic objectives). In other words, critical quality characteristics are characteristics related to the quality of a product. Critical quality characteristics are also called quality characteristics. Examples of critical quality characteristics include the purity of the product, the content of the product, the percentage of product variants, the stability of the product, and the percentage of impurities.

[0025] Key Productivity Characteristics (Productivity Characteristics): Key productivity characteristics refer to the physical, chemical, biological, and microbiological properties or characteristics of the manufacturing process environment (reaction field, culture environment, etc.) that must be within appropriate limits, ranges, and distributions in order to guarantee the productivity required for the product (productivity in the product characteristic target). In other words, key productivity characteristics are characteristics related to the productivity of product manufacturing. Key productivity characteristics are also called productivity characteristics. They are also called key production characteristics or production characteristics. Examples of key productivity characteristics include antibody concentration, antibody quantity, protein concentration, and protein quantity.

[0026] Critical Process Parameters: Critical process parameters are parameters that correlate with critical quality characteristics and / or critical productivity characteristics. In other words, critical process parameters are parameters that cause critical quality characteristics and / or critical productivity characteristics. Critical process parameters may also be parameters whose influence on critical quality characteristics and / or critical productivity characteristics meets a predetermined criterion. Examples of critical process parameters include parameters related to the specifications and operating conditions of manufacturing equipment. Examples of critical process parameters include culture medium state parameters related to the state of the culture medium, fluid parameters related to the state of the fluid, cell state parameters related to the state of the cells and the environment surrounding the cells, setting parameters related to the manufacturing conditions set in the production of the product, and design parameters related to the design specifications of the equipment used in the production of the product.

[0027] Secondary critical process parameters: Secondary critical process parameters are critical process parameters that directly correlate with critical quality characteristics and / or critical productivity characteristics.

[0028] Primary critical process parameters: Primary critical process parameters are parameters that are directly involved in the design, setup, and operation of critical process parameters, and that are indirectly correlated with critical quality characteristics and / or critical productivity characteristics through secondary critical process parameters.

[0029] Culture (reaction field) environment: The culture (reaction field) environment includes the physical and chemical factors to which the product (including intermediates) is exposed during the manufacturing process. In manufacturing equipment involved in culture, these include culture medium state parameters (concentration of culture medium components, dissolved oxygen concentration, dissolved carbon dioxide concentration, temperature distribution, pH distribution, etc.), cell state parameters (growth rate, antibody production rate, metabolic rate, gene expression level, etc.), and fluid parameters (shear stress, Kolmogorov scale, oxygen transfer capacity coefficient (kLa), mixing time constant).

[0030] Manufacturing conditions: Manufacturing conditions refer to the values ​​of parameters related to the specifications and operating conditions of the manufacturing equipment in the manufacturing process. For manufacturing equipment involved in culture, these include setting parameters (pH, dissolved carbon dioxide concentration, dissolved oxygen concentration, etc.) and design parameters (culture tank capacity, impeller shape, culture tank aspect ratio, tank volume, aeration of the liquid in the tank (bubble diameter, aeration rate), etc.).

[0031] (Step S1: Setting Product Characteristics Targets) In Step S1, product characteristics targets are set, which are the quality (function) and / or productivity required for the product to meet its purpose. Examples of quality (function) include activity, efficacy, and side effects. Examples of productivity include yield per unit volume, yield rate, and yield per input cost.

[0032] (Step S2: Extraction of candidate key quality characteristics and key productivity characteristics) In Step S2, factors that have a significant impact on the quality (function) set in Step S1 are designated as key quality characteristics. Also in Step S2, factors that have a significant impact on the productivity set in Step S1 are designated as key productivity characteristics. Examples of key quality characteristics include the molecular structure of the product, various chemical modifications to the product and their proportions, the proportion of impurities (contaminants), the proportion of variants, and the physical properties of the solvent (pH, temperature). Examples of key productivity characteristics include cell (enzyme) concentration and product production rate. The relationship between product characteristic targets such as function and productivity and key quality characteristics and / or key productivity characteristics differs depending on the product and manufacturing method. Therefore, each time a manufacturing process is constructed, the relationship between product characteristic targets and key productivity characteristics and / or key quality characteristics is measured and investigated, and product characteristics that may have an influence (correlation) are extracted. In the investigation, correlation analysis is used to find correlations from data obtained through experiments and other means. Furthermore, product characteristics are extracted as candidates for important quality characteristics or important productivity characteristics based on similar examples from publicly available literature.

[0033] (Step S3: Identification of Key Quality and Productivity Characteristics Based on Evidence) The candidate key quality and / or key productivity characteristics extracted in Step S2 are not necessarily correlated with the target product and manufacturing process. Therefore, in Step S3, the key quality and / or key productivity characteristics that actually correlate with the target product characteristics are identified in the target product and manufacturing process. These are generally identified by experimentally changing the values ​​of the candidate key quality and key productivity characteristics and examining how much each performance requirement of the target product characteristics changes. Those with large changes can be identified as key quality and / or key productivity characteristics. If the theory is well established for the system, a method of examining the correlation by calculation using in silico may be adopted. In this way, key quality and / or key productivity characteristics are identified using evidence-based methods.

[0034] (Step S4: Extraction of Candidate Critical Process Parameters) In Step S4, candidate critical process parameters that have a significant impact on the critical quality characteristics and / or critical productivity characteristics identified in Step S3 are extracted. The relationship between critical quality characteristics and / or critical productivity characteristics and critical process parameters differs depending on the product and manufacturing method. Therefore, each time a manufacturing process is constructed, the relationship between critical quality characteristics and / or critical productivity characteristics and critical process parameters is investigated, and performance requirements that may have an impact (correlation) are extracted. In the investigation, correlations are found using correlation analysis from data obtained through experiments and other means. In addition, critical quality characteristics and / or critical productivity characteristics are extracted as candidate critical process parameters based on similar examples from publicly available literature. Since extracting candidate critical process parameters from a vast amount of publicly available literature requires a great deal of time and effort, it is advisable to use generative artificial intelligence (AI). By using generative AI, candidate critical process parameters can be extracted efficiently.

[0035] (Step S5: Identification of critical process parameters based on evidence) The candidate critical process parameters extracted in Step S4 are not necessarily correlated with the target product and manufacturing process. Therefore, in Step S5, critical process parameters that actually affect the target product characteristics are identified in the target product and manufacturing process. This identification is generally done by investigating, through experiments, how much the values ​​of critical process parameters affect the performance requirements of each critical quality characteristic and / or critical productivity characteristic. Those with a large impact can be identified as critical process parameters. If the system can be numerically modeled, a method of investigating correlations through simulation may also be used. In this way, critical process parameters are identified using an evidence-based method.

[0036] For example, in step S5, n types of important quality characteristics and / or important productivity characteristics Q ｉ and m types of important process parameters (P １ , P ２ , , , P ｍ ) If Q is identified, ｉ and (P １ , P ２ , , , Pｍ The relationship between the two can be expressed as shown in Equation 1 through experiments and theoretical analysis. ｉ = f(P) １ , P ２ , , , P ｍ i = 1, 2, ..., n ... (Equation 1)

[0037] Equation 1 can be derived from theory alone or from experiment alone, but often the function is determined by deriving the general shape of the function theoretically and then determining the coefficients of the function experimentally. In this case, the error of the critical process parameter is the critical quality characteristic and / or critical productivity characteristic Q. ｉ The magnitude of the influence is given by each term in Equation 2 (∂f(P) / ∂P ｉ ΔP ｉ ) can be evaluated, and if this term is large, it indicates a critical quality characteristic and / or a critical productivity characteristic Q ｉ It can be seen that it has a significant impact. ΔQ = ∂f(P) / ∂P １ ΔP １ +∂f(P) / ∂P ２ ΔP ２ +∂f(P) / ∂P ３ ΔP ３ +... (Formula 2)

[0038] Fluid simulations can be used to theoretically derive the general shape of a function. Using fluid simulations, based on information about the culture vessel structure used in the experiment and the culture vessel structure after scale-up, it is possible to calculate critical quality attributes (CQAs) such as shear stress, dissolved oxygen distribution, and dissolved carbon dioxide distribution in the vessel using fluid analysis with explanatory variables such as stirring speed and empty column aeration velocity. By deriving CQAs for various explanatory variables, the general shape of the function can be investigated.

[0039] When determining the coefficients of a function through experimentation, it is necessary to consider the validity of the experimental values ​​to be input. For example, when using dissolved oxygen concentration as the input value, the dissolved oxygen concentration in the culture tank in a culture experiment is distributed within the tank, so a single point measured by a sensor may not necessarily represent the appropriate value in the experiment. Therefore, the dissolved oxygen distribution within the culture tank is calculated by fluid simulation, and this information is used to correct the measured values ​​(or to position the sensor appropriately based on the fluid simulation results) to input appropriate experimental values.

[0040] As described above, when the manufacturing process in which the product is produced includes cultivation in a culture tank, critical process parameters may be identified based on fluid simulations within the culture tank. Also, as described above, when the manufacturing process in which the product is produced includes cultivation in a culture tank, critical process parameters may be identified based on experimental values ​​for candidate critical process parameters and the results of fluid simulations within the culture tank for the candidate critical process parameters. Furthermore, critical process parameters may be identified based on the results of correcting experimental values ​​for candidate critical process parameters using the results of fluid simulations within the culture tank for the candidate critical process parameters.

[0041] To create a design space boundary, N-1 measurements are required for every N measurements (P). To create a boundary that minimizes ΔQ, use ∂f(P) / ∂P ｉ ΔP ｉ You can create a boundary line using the measurement items excluding the measurement i with the largest value.

[0042] (Step S6: Determining the time in the manufacturing process for which design space should be created) All stages of the manufacturing process are controlled to meet the product characteristics targets. It is possible to create design spaces for each stage of the process sequentially to determine the manufacturing conditions, but in that case, creating the design spaces would be extremely time-consuming and labor-intensive. Therefore, in Step S6, we will decide which point (time) in the manufacturing process for which design spaces should be created. There may be more than one point (time) for which design spaces should be created.

[0043] First, determine whether the manufacturing process in question is in a steady state (steady operation or a steady environment) or a transient state (transient operation or a transient environment). If it is in a steady state, the culture (reaction field) environment is the same at any given time, so a design space can be created at one or more points in time to determine the range of manufacturing conditions. If it is in a transient state, the culture (reaction field) environment changes over time. Therefore, critical quality characteristics and / or critical productivity characteristics Q ｉ To verify that all performance requirements are met at all times, we would need to create a design space for all times. However, this is too laborious, so it is preferable to adopt the following method.

[0044] Key quality characteristics and / or key productivity characteristics Q ｉ For each product characteristic, the timing at which critical process parameters have the greatest impact within the manufacturing process is measured and examined individually. Then, the time at which the design space for that performance requirement should be created is identified. There may be multiple times. After determining the time for each performance requirement and creating a design space for each time, the design spaces for all performance requirements are superimposed on a common critical process parameter coordinate axis, thereby determining the manufacturing conditions that must be met throughout the process duration.

[0045] (Step S7: Creating the design space) Key quality characteristics and / or key productivity characteristics Q identified in Step S3 ｉ The range of key process parameters (range of explanatory variables) identified in step S5 is determined so that the range satisfies the product characteristic target (range that the dependent variable must satisfy). This area that must be satisfied is the design space, and it is the range in which the manufacturing conditions are controlled. The design space is created using the following procedure.

[0046] In step S5 of the manufacturing process, as described above, n types of important quality characteristics and / or important productivity characteristics Q ｉ and m types of important process parameters (P １ , P ２ , , , P ｍ ) If Q is identified,ｉ and (P １ , P ２ , , , P ｍ The relationship between ) can be expressed as shown in Equation 1 through experiments and / or theoretical analysis. Equation 1 is shown again. Q ｉ = f(P) １ , P ２ , , , P ｍ i = 1, 2, ..., n ... (Equation 1)

[0047] A certain quality standard above which a key quality characteristic and / or key productivity characteristic Q ｉ Key process parameters (P) required to achieve this １ , P ２ ..., P ｍ The range of the minimum value Q is ｉ０ In the above cases, it can be expressed by Equation 3. The boundary of the design space is the minimum quality standard Q. ｉ０ The important process parameter (P) at that time １ , P ２ , , , P ｍ Because of the value of ), it can be expressed by equation 4. Q ｉ０ ≤ f(P) １ , P ２ , , , P ｍ ) i = 1, 2, ..., n ... (Equation 3) Q ｉ０ = f(P) １ , P ２ , , , P ｍ i = 1, 2, ..., n ... (Equation 4)

[0048] To create the design space boundary, any (P) that satisfies Equation 4 is used. １ , P ２ , , , P ｍ ) should be determined. One of the important process parameters is P. ｉ For this, transform equation 4 into equation 5, and any (P １ ,・・P ｉ－１ , P ｉ＋１ , , , P ｍ For the value of P ｉBy obtaining it, evaluation using mathematical formulas and figures becomes easy. P ｉ = g(P １ , ···P ｉ－１ , P ｉ＋１ , ···, P ｍ ) …(Equation 5)

[0049] On the other hand, there are cases where the equation cannot be easily transformed as in Equation 5. In such cases, the solution set of (P １ , P ２ , ···, P ｍ ) can be obtained through numerical analysis. Numerical analysis can use solvers attached to spreadsheet software, etc.

[0050] When calculating using a solver, the following problems may occur. As an example, for a certain product quality index Q ０ , there are two types of important process parameters (P ｍ , P ｎ ) and can be expressed as in Equation 6. When the change in P ｍ corresponding to the change in P ｎ is very large (see Figure 3), that is, when dP ｎ / dP ｍ is very large (for example, 10 times), a small change in P ｍ may greatly affect P ｎ , and there is a possibility of being overly affected by experimental errors in P ｍ . Note that Figure 3 is a graph showing the relationship between multiple manufacturing conditions (P ｎ , P ｍ ) to satisfy the product quality index Q ０ when dP ｍ , P ｎ ) is very large. Q ０ = f(P １ , P<000t07>) …(Equation 6)

[0051] To avoid being overly affected by experimental errors in P<00001t08> / dP ｍ when it is very large, P ｍ ≒ f ｍ ​－１ (P ｎ ) like P ｍ P ｎ It is best to express it as the inverse function (inverse mapping) of . If we can express it in this way, then on this inverse function (inverse mapping), dP ｍ / dP ｎ It becomes 1 or less and small, P ｍ This avoids being overly affected by experimental errors and other factors. Here, P ｍ P ｎ To express it using its inverse function (inverse mapping), we need a real number P within the tolerance range. ｎ For a real number P ｍ Only one of them must be determined. For example, P ｎ ≒P ｍ ２ In this case, as shown in Figure 4, the inverse function P ｍ ≈±√P ｎ In P ｍ Since it has two values, P ｍ P ｎ It cannot be said that it is the inverse function of . In such cases, the range of this inverse function is P ｍ By setting > 0, P ｎ For a real number P ｍ It is possible to determine a unique value. Note that Figure 4 shows P ｎ = P ｍ ２ When this happens, the inverse function P ｍ ≈±√P ｎ In P ｍ Since it has two values, P ｍ P ｎ It cannot be said that it is the inverse function of (left figure), and the range of the inverse function is P ｍ By setting > 0, P ｎ For a real number P ｍ This graph shows that it is possible to determine the value in a way that is unique (see diagram on the right).

[0052] As described above, when drawing the curve of the design space boundary, if it is difficult to express the function and its inverse function with simple mathematical formulas or if the formulas cannot be transformed, it can be easily done using a solver in spreadsheet software. Even in this case, if the rate of change of multiple manufacturing conditions is extremely different or the conditions for the inverse function to hold are not met, the calculation by the solver will often diverge and will not converge, and the method described in this embodiment is effective. The product quality index Q is obtained by drawing the curve using these methods. ０ A manufacturing condition range that satisfies these conditions can be defined, that is, the boundary of the design space as shown in the right-hand diagrams of Figures 3 and 4.

[0053] The above is P ｍ and P ｎ The above describes the case with two variables, but in the case of multiple variables, the concept of inverse mapping shown in Figure 5 is effective. In this case, multiple solutions may exist, and by changing the initial values ​​in various ways using an optimization method (see Figure 6), multiple candidate solutions (local optima of ΔQ, such as local minimums and global minimums) can be obtained. By setting these initial values ​​close to the experimental conditions, the actual solution can be found relatively easily. Figure 5 is an explanatory diagram illustrating the concept of inverse mapping. Figure 6 is an explanatory diagram showing an example of an optimized solution in the case of multiple variables.

[0054] The design space represents the range of explanatory variables when the objective variable satisfies defined conditions. Here, Figure 7 is an explanatory diagram showing an example of the correlation between critical quality characteristics 171 and critical process parameters 173a to 173e in a conventional design space. As shown in Figure 7, the conventional design space sets the objective variable as critical quality characteristics 171, 171 that are correlated with the product characteristic target 170. The conventional design space also sets the explanatory variables as critical process parameters 173a to 173e that are correlated with the critical quality characteristics 171, 171. However, depending on the situation, the number of parameters in the correlation equation required when creating the design space can be reduced by setting the objective variable as a secondary critical process parameter and the explanatory variables as a primary critical process parameter.

[0055] When there are two explanatory variables, the design space becomes a planar region shown on a two-dimensional graph, and when there are three, it becomes a spatial region shown on a three-dimensional graph. When there are four or more variables, it becomes impossible to visually represent the design space. Therefore, when there are four or more variables, two or three of the explanatory variables are selected, and the remaining variables are fixed to arbitrary values. The graph is then generated based on the selected explanatory variables (a two-dimensional graph for two variables, and a three-dimensional graph for three variables), and by changing the fixed values ​​and observing the changes in the design space, the overall picture of the design space can be imagined. One way to implement this is to have a "selection button" to determine the explanatory variables to be selected, a "slider" to freely change the fixed values ​​of the unselected explanatory variables, and a Graphical User Interface (GUI) that displays a two-dimensional or three-dimensional graph based on the above information.

[0056] As described above, after creating the design space, it is preferable that this management method includes a display step (not shown) that displays the design space on a display unit as a graph with multiple important process parameters as explanatory variables. In this display step, the important process parameters selected from among the identified important process parameters can be used as variables in the graph, and the important process parameters other than the selected important process parameters can be set as fixed values ​​to create the graph.

[0057] Figure 8 shows an example of reducing the number of parameters in the correlation equation required when creating a design space by setting the dependent variable as a secondary critical process parameter and the explanatory variables as primary critical process parameters. Figure 8 is an explanatory diagram showing an example of the correlation between critical quality characteristics 71, critical productivity characteristics 72, secondary critical process parameters 82, and primary critical process parameters 81. As shown in Figure 8, secondary critical process parameters 82 are items that are closely correlated with critical quality characteristics 71 and / or critical productivity characteristics 72. Primary critical process parameters 81 are items that are far from critical quality characteristics and / or critical productivity characteristics 72, even if they are close to design and operation parameters. Examples of secondary critical process parameters 82 include culture medium state parameters, cell state parameters, and fluid parameters. Examples of primary critical process parameters 81 include setting parameters and design parameters.

[0058] In this case, the design space is positioned as shown in Figure 9. Figure 9 is an explanatory diagram showing the relationship between the secondary design space 92 and the primary design space 91. As shown in Figure 9, the secondary design space 92 includes important quality characteristics 71 and / or important productivity characteristics 72 that are correlated with the product characteristic target 70, and secondary important process parameters 82. These secondary important process parameters 82 include important process parameters 73a to 73e that are correlated with the important quality characteristics 71 and important productivity characteristics 72, respectively. The primary design space 91 includes the secondary important process parameters 82 and the primary important process parameters 81. These primary important process parameters 81 include important process parameters 73d and 73e that are correlated with the important process parameters 73a to 73c of the secondary important process parameters 82, respectively.

[0059] In particular, in cell culture, as shown in Figure 10, there are numerous parameters and they have complex correlations. Therefore, dividing and modularizing important process parameters to create a design space is effective in reducing the number of parameters that need to be considered. Figure 10 is an explanatory diagram showing an example of correlations in cell culture. In the example shown in Figure 10, the numerous parameters include quality parameters related to important quality characteristics and productivity parameters related to important productivity characteristics. In addition, there are culture medium state parameters, fluid parameters, and cell state parameters that are correlated with each of the quality parameters and productivity parameters, and these are classified as secondary important process parameters 82. Furthermore, there are setting parameters and design parameters that are correlated with each of the culture medium state parameters, fluid parameters, and cell state parameters, and these are classified as primary important process parameters 81.

[0060] Quality parameters include sugar chain structure, polymers, host-derived proteins (HCPs), host-derived DNA, degradation products, and monomers. Productivity parameters include antibody concentration, antibody production per unit of glucose consumption, and antibody production per culture vessel. Culture medium state parameters include ammonia concentration, glucose concentration, lactate concentration, glutamine concentration, alanine concentration, temperature distribution differences, pH distribution differences, dissolved oxygen concentration distribution differences, and dissolved carbon dioxide distribution differences. Fluid parameters include mixing time constant, shear stress, Kolmogorov scale, and oxygen transport capacity coefficient (kLa). Cell state parameters include growth rate, glucose consumption rate, antibody ratio production rate, ammonia secretion rate, lactate secretion rate, glutamine consumption rate, viable cell density, cell morphology, cell diameter distribution differences, cell viability, and gene expression level. Setting parameters include cell separation method, perfusion rate, pH, temperature, stirring speed, and dissolved oxygen concentration. Design parameters include the capacity of the culture vessel (vessel volume), the shape of the stirring blades (blade shape), the aspect ratio of the culture vessel (vessel aspect), the amount of air permeation in the liquid, and air bubbles.

[0061] (Step S8: Determination of Manufacturing Conditions) In the design space created in Step S7, the area in which the explanatory variables keep the dependent variable within the range they must satisfy is the control area for manufacturing conditions. Since the boundary of the design space includes errors associated with experiments, the errors that occur in experiments may be calculated, and the area inside the boundary, excluding the part where errors are expected, may be considered the controlled manufacturing conditions. The error can be calculated as the variance σ of the errors of the explanatory variables in the design space. If the errors follow a normal distribution, the true boundary will be within a 1σ width from the boundary with a 68% probability, within a 2σ width with a 95% probability, and within a 3σ width with a 99.7% probability. The width can be determined by considering the risk of deviation from the true boundary and the severity of the deviation.

[0062] The control value can be within the boundary line considering the above error, but a more appropriate control value is the point (center) in the design space that is furthest from all boundary lines. Since the design space is usually a multidimensional region, to find its center, if the multidimensional region is represented as a set of points, the center is calculated by calculating the average of the coordinates in each dimension, and if the multidimensional region is defined as a function, the centroid is found by using the integral of the function. 1. In the case of a set of points: The center can be found by adding up the coordinates of each point and dividing by the number of points. For example, in the case of a set of points in two-dimensional space, the coordinates of the center can be determined by finding the average of the x coordinates and y coordinates of each point. Similarly, in the case of a general multidimensional space, the center can be found by calculating the average of the coordinates in each dimension. 2. In the case of a function: The center can be found by using the integral of the function. First, the region is used as the integration range, and the coordinate variables of each dimension are integrated. Next, the center can be found by dividing the result by the volume (or area) of the region.

[0063] …(Equation 7) V: Closure of the design space

[0064] For example, if one of the explanatory variables is temperature, the center of the design space can be determined using the method described above, and if the temperature value is 36°C, that value becomes the control value.

[0065] The above control values ​​are for a specific time point in the manufacturing process. If the design space within the manufacturing process changes constantly, the control values ​​will also change over time. For example, if the control value at time t is T°C and the control value at time t+Δt is T+ΔT°C, then the temperature is controlled to change with a gradient of ΔT / Δt. However, the above temperature gradient is adjusted as appropriate to ensure that the temperature does not deviate from the design space during the temperature change process.

[0066] This management method may determine appropriate control values ​​based on the design space and send control instructions based on those control values ​​to the equipment in the manufacturing process. This makes it possible to automatically manufacture the target product.

[0067] As described above, this management method may include a control process (not shown) that creates control values ​​determined based on the design space and control instructions based on manufacturing conditions based on those control values, and transmits those control instructions to the equipment in the manufacturing process in order to have the equipment in the manufacturing process execute them.

[0068] Furthermore, as described above, this management method may also be configured to create control instructions that change over time based on information about control values ​​and design space ranges at different points in time during the control process.

[0069] Furthermore, this management method may also display information relating appropriate control values ​​to the created design space during the control process. In this case, the design space information displayed on the display unit may be a graph in which multiple important process parameters are used as explanatory variables, similar to the example of the display process described above. In this case, the design space information displayed on the display unit during the control process may be a graph in which selected important process parameters from the identified important process parameters are used as variables in the graph, and important process parameters other than the selected important process parameters are set as fixed values. Furthermore, in this management method, as an example of how to display information relating appropriate control values ​​to the created design space during the control process, information indicating the position of appropriate control values ​​may be overlaid on the graph representing the design space. The position of the control value may be, for example, the center, centroid, or coordinates mentioned above. Furthermore, this management method may also display control instructions and design space information that change over time during the control process. This makes it possible for this management method to easily monitor the manufacturing process while performing optimal control based on the created design space.

[0070] As described above, this management method may also involve associating information regarding appropriate control values ​​with information from the created design space during the control process and displaying it on the display unit.

[0071] Furthermore, the design space information displayed on the display unit during the control process may be a graph in which the design space uses multiple critical process parameters as explanatory variables. This graph may be created by using selected critical process parameters from among the identified critical process parameters as variables in the graph, and setting critical process parameters other than the selected critical process parameters as fixed values.

[0072] Furthermore, in this management method, information indicating the appropriate control value position may be overlaid and displayed on a graph representing the design space during the control process.

[0073] Furthermore, in this management method, during the control process, information on control instructions and design space that changes over time, created based on information on control values ​​and the range of the design space at different points in time, may be displayed on a graph representing the design space.

[0074] By performing steps S1 through S8 described above, the manufacturing conditions to be controlled can be determined. Note that the order of steps S3 and S4 may be reversed. Also, steps S4, S5, and S6 may be in the order of S6, S4, and S5.

[0075] (An Embodiment of Manufacturing Condition Management Method) A specific embodiment of the management method described above is as follows. Figure 11 is a flowchart illustrating a specific embodiment of the manufacturing condition management method according to one embodiment of the present invention. As shown in Figure 11, the management method includes a first extraction step S12, a first identification step S13, a discrimination step S14, a selection step S15, a second extraction step S16, a second identification step S17, a creation step S18, and a determination step S19. Furthermore, as shown in Figure 11, the management method may also include a setting step S11 before the first extraction step S12. Figure 11 includes the setting step S11. These steps are generally carried out in the order of the steps listed above, but the order can be changed depending on the situation. In this embodiment, the steps S1 to S8 described above are carried out in the order of steps S1, S2, S3, S6, S4, S5, S7, and S8.

[0076] In the setting step S11, product characteristic targets 70 are set in order to manufacture the target product. This corresponds to step S1 described above. In the first extraction step S12, important quality characteristics 71 (quality characteristics) that may correlate with the product characteristic targets 70 described above are extracted, i.e., candidates for quality characteristics related to the quality in the manufacture of the target product and / or candidates for important productivity characteristics 72 (productivity characteristics) related to the productivity in the manufacture of the product. This corresponds to step S2 described above. In the first identification step S13, the target important quality characteristics 71 and / or important productivity characteristics 72 are identified based on evidence from the extracted candidates for important quality characteristics 71 and / or important productivity characteristics 72. This corresponds to step S3 described above. In the determination step S14, it is determined whether the manufacturing process in which the product is manufactured is in a steady state or a transient state. This corresponds to part of step S6 described above. In the selection step S15, if it is in a steady state, any one time within the manufacturing process to create a design space is selected, and if it is in a transient state, one or more times within the manufacturing process to create a design space are selected. This also corresponds to part of step S6 described above. In the second extraction step S16, candidates for important process parameters 73a to 73e (73) that cause the important quality characteristics 71 and / or important productivity characteristics 72 identified above are extracted. This corresponds to step S4 described above. In the second identification step S17, based on the evidence, the important process parameter 73 is identified from the candidates for important process parameter 73 extracted above. This corresponds to step S5 described above. In the creation step S18, the range of the important process parameter 73 identified above is determined and a design space is created. This corresponds to step S7 described above. In the determination step S19, the manufacturing conditions are determined based on the design space created above. This corresponds to step S8 described above.

[0077] This control method performs the setting process S11 to the determination process S19 (steps S1 to S8), particularly the first extraction process S12 to the determination process S19 (steps S2 to S8), and as explained above, it can efficiently create a highly accurate design space while considering various factors and parameters. Furthermore, this control method efficiently selects important process parameters 73 that affect the product characteristic target 70 and grasps the appropriate range with fewer experiments, thereby reducing unexpected fluctuations in productivity and quality before and after scale-up or when changing manufacturing conditions. In addition, this control method classifies important process parameters 82 and important process parameters 81, and reduces redundant parameters by performing sensitivity analysis on each parameter, thereby reducing the time and cost required to collect necessary experimental data by creating a design space in a part of the manufacturing process. As an example of a use case using this control method, the behavior analysis of important process parameters and the creation of a corresponding design space can be performed for each different equipment condition (for example, a combination of conditions such as tank shape, agitator shape, and agitator specifications). This makes it possible to search for and identify equipment conditions with similar or equivalent parameter behavior, even under different equipment conditions, and to create corresponding design spaces. The preferred embodiment of this management method is as follows:

[0078] In the first extraction step S12, both candidates for important quality characteristics 71 and important productivity characteristics 72 are extracted. In the first identification step S13, both important quality characteristics 71 and important productivity characteristics 72 are identified based on evidence from both candidates for important quality characteristics 71 and important productivity characteristics 72. Then, in the second extraction step S16, candidates for important process parameters 73 that cause both important quality characteristics 71 and important productivity characteristics 72 are extracted. In this way, this control method can create a more accurate design space that simultaneously considers both important quality characteristics 71 and important productivity characteristics 72.

[0079] In this management method, it is preferable to classify the aforementioned critical process parameters 73 into primary critical process parameters 81 that have a distant correlation with critical quality characteristics 71 and / or critical productivity characteristics 72, and secondary critical process parameters 82 that have a close correlation with critical quality characteristics 71 and / or critical productivity characteristics 72. In this way, this management method can divide and modularize the critical process parameters 73 to create a design space, thereby reducing the number of parameters that need to be considered.

[0080] In this control method, it is preferable to create a design space using the secondary critical process parameter 82 as the objective variable and the primary critical process parameter 81 as the explanatory variable. This reduces the number of parameters in the correlation equation required when creating the design space.

[0081] In this control method, it is preferable that the product characteristic target 70, which is the quality and productivity that the product must satisfy in the manufacturing process described above, is one or more selected from the group of product function, product yield, and yield. In this way, this control method can create a highly accurate design space that takes these into particular consideration.

[0082] In this control method, it is preferable that the aforementioned important quality characteristics 71 are one or more selected from the group consisting of sugar chain structure, host-derived DNA, monomer, polymer, and degradation product. In this way, this control method can create a highly accurate design space that takes these into particular consideration as important quality characteristics 71.

[0083] In this management method, it is preferable that the aforementioned important productivity characteristics 72 are one or more selected from the group consisting of antibody concentration, antibody production per unit of glucose consumption, and antibody production per culture vessel. In this way, this management method can create a highly accurate design space that takes these into particular consideration as important productivity characteristics 72.

[0084] In this control method, it is preferable that the aforementioned critical process parameter 73 is one or more selected from the group consisting of culture medium state parameters, temperature distribution differences, dissolved carbon dioxide distribution differences, dissolved oxygen concentration distribution differences, fluid parameters, cell state parameters, setting parameters, and design parameters. In this way, this control method can create a highly accurate design space that takes these into particular consideration as critical process parameters 73.

[0085] In this control method, it is preferable that the aforementioned important process parameters 73 are one or more selected from the group consisting of ammonia concentration, glucose concentration, lactate concentration, alanine concentration, glutamine concentration, mixing time constant, shear stress, Kolmogorov scale, oxygen transfer capacity coefficient (kLa), growth rate, cell shape, cell diameter distribution difference, cell viability, glutamine consumption rate, glucose consumption rate, antibody ratio production rate, ammonia secretion rate, lactate secretion rate, viable cell number density, gene expression level, pH, temperature, stirring speed, dissolved oxygen concentration, cell separation method, perfusion rate, tank capacity, blade shape, tank aspect ratio, liquid aeration rate, and bubble diameter. In this way, this control method can create a highly accurate design space that takes these into particular consideration as important process parameters 73.

[0086] In this control method, it is preferable that the secondary critical step parameter 82 is one or more selected from the group consisting of culture medium state parameters, temperature distribution differences, dissolved carbon dioxide distribution differences, dissolved oxygen concentration distribution differences, fluid parameters, and cell state parameters. In this way, this control method can create a highly accurate design space that takes these into particular consideration as secondary critical step parameters 82.

[0087] In this control method, it is preferable that the secondary critical step parameter 82 is one or more selected from the group consisting of ammonia concentration, glucose concentration, lactate concentration, alanine concentration, glutamine concentration, mixing time constant, shear stress, Kolmogorov scale, kLa, proliferation rate, cell morphology, cell diameter distribution difference, cell viability, glutamine consumption rate, glucose consumption rate, antibody ratio production rate, ammonia secretion rate, lactate secretion rate, viable cell number density, and gene expression level. In this way, this control method can create a highly accurate design space that takes these into particular consideration as secondary critical step parameters 82.

[0088] In this management method, it is preferable that the primary critical process parameter 81 is one or more selected from the group of setting parameters and design parameters. In this way, this management method can create a highly accurate design space that takes these into particular consideration as primary critical process parameters 81.

[0089] In this control method, it is preferable that the primary critical process parameter 81 is one or more selected from the group consisting of pH, temperature, stirring speed, dissolved oxygen concentration, cell separation method, perfusion rate, tank capacity, blade shape, tank aspect ratio, liquid aeration rate, and bubble diameter. In this way, this control method can create a highly accurate design space that takes these into particular consideration as primary critical process parameters 81.

[0090] This control method involves the above-mentioned manufacturing process S18, where multiple manufacturing conditions are met by the product quality index Q. ０ In the manufacturing process that affects the process, when determining the manufacturing conditions necessary to achieve a certain quality standard, it is preferable to treat the interrelationship between two or more of the aforementioned manufacturing conditions as a function and / or inverse function, so that each of the two or more manufacturing conditions is uniquely determined. In this way, this control method avoids the excessive influence of experimental errors and other factors that can occur when there are two or more manufacturing conditions and one of them changes very large, thereby preventing excessive influence. As a result, this control method can create a more accurate design space.

[0091] This management method is based on the aforementioned product quality index Q ０ The relationship between this and the aforementioned manufacturing conditions, that is, the product quality index Q ０ It is preferable to determine multiple manufacturing conditions to satisfy (critical quality characteristic 71 and / or critical productivity characteristic 72) through experimentation and / or theoretical analysis. In this way, this control method can create a more reliable, simpler, and accurate design space.

[0092] In this control method, when determining the manufacturing conditions in the aforementioned determination step S19, it is preferable to change the domain and / or range of these (i.e., the interrelationship between two or more manufacturing conditions) so that each of the two or more manufacturing conditions is uniquely determined. In this way, this control method can avoid being overly affected by experimental errors and other factors when a change in one manufacturing condition is very large, thereby preventing the control method from significantly influencing other manufacturing conditions. Therefore, this control method can create a more accurate design space.

[0093] In this control method, when determining the manufacturing conditions in the determination step S19 described above, it is preferable that one of the correlations between the two or more manufacturing conditions is strictly monotonically increasing or strictly monotonically decreasing. Strictly monotonically increasing means that the value obtained increases as the value of the variable increases, and a function that behaves in this way is called a strictly monotonically increasing function. Strictly monotonically decreasing means that the value obtained decreases as the value of the variable increases, and a function that behaves in this way is called a strictly monotonically decreasing function. In addition, in this embodiment, it is also preferable that when the value of the two or more manufacturing conditions with the larger rate of change is determined to one, the value of the condition with the smaller rate of change is also determined to one (see the right diagram in Figures 3 and 4). As long as these definitions hold, the strictly monotonically increasing function and the strictly monotonically decreasing function may be linear or quadratic functions. In this way, this control method makes it easy to determine the manufacturing conditions.

[0094] In this control method, it is preferable that the manufacturing process described above is carried out in a culture device. The culture device can be any commonly used device and is not particularly limited. In this way, this control method can create a highly accurate design space for the culture device.

[0095] In this management method, the product quality index Q of the product is as described above. ０ If there are two or more items, each product quality index Q ０ It is preferable to select an explanatory variable for each and create a different design space for each. In this way, this control method allows the product quality index Q to be controlled. ０ Having a dedicated design space allows for the determination of more appropriate manufacturing conditions.

[0096] It is preferable to create the design space for each manufacturing apparatus of different scales or for each manufacturing apparatus with different cell number densities using this control method. In this way, the control method can create a more accurate design space for various manufacturing conditions.

[0097] (Manufacturing Conditions Management Program) Next, a manufacturing conditions management program 120 (hereinafter sometimes referred to as "the management program 120") according to one embodiment of the present invention will be described. Figure 12 is a flowchart illustrating the contents of the manufacturing conditions management program 120 according to one embodiment of the present invention and a computer-readable recording medium 121 on which the manufacturing conditions management program 120 is recorded. The management program 120 causes the computer 1310 (see Figure 14) to execute the management method described above.

[0098] As shown in Figure 12, specifically, the management program 120 causes the computer 1310 to execute the first extraction step S22, the first identification step S23, the discrimination step S24, the selection step S25, the second extraction step S26, the second identification step S27, the creation step S28, and the determination step S29. Furthermore, as shown in Figure 12, the management program 120 may have a setting step S21 before the first extraction step S22. In this setting step S21, the management program 120 receives the product characteristic target 70 entered by the user in order to manufacture the target product and sets it in the computer 1310. Figure 12 includes this setting step S21 in its illustration.

[0099] The execution of the first extraction step S22, first identification step S23, discrimination step S24, selection step S25, second extraction step S26, second identification step S27, creation step S28, and decision step S29 in this management program 120 is the same as the first extraction step S12, first identification step S13, discrimination step S14, selection step S15, second extraction step S16, second identification step S17, creation step S18, and decision step S19 in the management method described above. This management program 120 can be recorded on a program distribution server or on a computer-readable recording medium 121 such as a CD or DVD. The management program 120 is then installed on a computer 1310.

[0100] (Computer-readable recording medium recording a manufacturing condition management program) In one embodiment, the management program 120 can be provided to the user in the form of a computer-readable recording medium 121 on which the management program 120 is recorded. The management program 120 recorded on the computer-readable recording medium 121 is read by a computer 1310 (see Figure 14) and installed on the computer 1310, causing the computer 1310 to execute the above-mentioned steps, namely the setting step S21, the first extraction step S22, the first identification step S23, the discrimination step S24, the selection step S25, the second extraction step S26, the second identification step S27, the creation step S28, and the determination step S29 (see Figure 12).

[0101] The management program 120 and the computer-readable recording medium 121 on which the management program 120 is recorded cause the computer 1310 to execute the setting process S21 to the determination process S29 (steps S1 to S8), in particular the first extraction process S22 to the determination process S29 (steps S2 to S8), so that, similar to the management method described above, a highly accurate design space can be efficiently created while considering various factors and parameters. Furthermore, the management program 120 and the computer-readable recording medium 121 on which the management program 120 is recorded can efficiently select important process parameters 73 that affect the product characteristic target 70 and grasp the appropriate range with fewer experiments, thereby reducing unexpected fluctuations in productivity and quality before and after scale-up or when manufacturing conditions are changed. Furthermore, in the management program 120 and the computer-readable recording medium 121 on which the management program 120 is recorded, the parameters are classified into secondary critical process parameters 82 and primary critical process parameters 81. By performing sensitivity analysis on each parameter, redundant parameters can be reduced, and by creating a design space in a part of the manufacturing process, the time and cost required to collect necessary experimental data can be reduced. It goes without saying that the preferred embodiments described in the management method described above can also be applied to the management program 120 and the computer-readable recording medium 121 on which the management program 120 is recorded. It also goes without saying that the same effects as the example use case described in the management method described above can be obtained in the management program 120 and the computer-readable recording medium 121 on which the management program 120 is recorded.

[0102] (Manufacturing Condition Management System) Next, a manufacturing condition management system 1300 (hereinafter sometimes referred to as "this management system 1300") according to one embodiment of the present invention will be described. Figure 13 is a schematic diagram illustrating the manufacturing condition management system 1300 according to one embodiment of the present invention. This management system 1300 executes the steps S2 to S8 described above. That is, this management system 1300 executes the first extraction step S12, the first identification step S13, the discrimination step S14, the selection step S15, the second extraction step S16, the second identification step S17, the creation step S18, and the determination step S19 in this management method. Furthermore, this management system 1300 executes step S1 before step S2 described above. That is, this management system 1300 executes the setting step S11 before the first extraction step S12 in this management method.

[0103] As shown in Figure 13, the management system 1300 consists of a data input device 1301, a fluid analysis device 1302, a literature search device 1303, a database 1304, a correlation analysis device 1305, a laboratory experiment device 1306, and a design space creation device 1307. The data input device 1301 inputs culture conditions, analysis, and measurement instrument values ​​from laboratory experiments (past data). The fluid analysis device 1302 analyzes the results of fluid simulations related to laboratory experiments and culture devices after scale-up. The literature search device 1303 collects publicly available information. The database 1304 collects various types of data. The data obtained from the data input device 1301, the fluid analysis device 1302, and the literature search device 1303 are collected in this database 1304. The correlation analysis device 1305 analyzes the correlation between various important process parameters 73. The laboratory experiment device 1306 conducts laboratory experiments and obtains culture conditions, analysis, and measurement instrument values. The design space creation device 1307 creates design spaces for laboratory experiments and for scale-up projects.

[0104] The management system 1300 will now be described starting from step S1. In step S1, the product characteristic target 70 for manufacturing the target product is input using an input device (e.g., a keyboard) not shown in Figure 13, and set as a setting item in the management system 1300. In step S2, the data input device 1301 and the literature search device 1303 are used to collect past data and publicly available literature information, and store them in the database 1304 as candidates for important quality characteristics 71 and / or important productivity characteristics 72.

[0105] In step S3, laboratory experiments are conducted using the laboratory experiment apparatus 1306 with the parameters of each candidate stored in the database 1304 to identify important quality characteristics 71 and / or important productivity characteristics 72. In step S4, candidate important process parameters 73 that may be the cause (potentially correlated) of the important quality characteristics 71 and / or important productivity characteristics 72 identified above are searched for using the data input device 1301, literature search device 1303, database 1304, and correlation analysis device 1305, and the results are stored in the database 1304. In step S5, the parameters of each candidate are evaluated using the laboratory experiment apparatus 1306, or fluid analysis (e.g., quantification of shear stress distribution) is performed using the fluid analysis device 1302 to identify important process parameters 73. Then, in step S6, the time to create the design space is determined by performing a time-series analysis of the process 1308. In step S7, the design space is created using the design space creation device 1307, and in step S8, the possible range of manufacturing conditions is determined.

[0106] (An Embodiment of a Manufacturing Conditions Management System) A specific embodiment of the management system 1300 described above is as follows. Figure 14 is a schematic diagram illustrating a specific embodiment of a manufacturing conditions management system 1300 according to one embodiment of the present invention. As shown in Figure 14, the management system 1300 comprises a first extraction means 1312, a first identification means 1313, a discrimination means 1314, a selection means 1315, a second extraction means 1316, a second identification means 1317, a creation means 1318, and a determination means 1319. Furthermore, as shown in Figure 14, the management system 1300 may also include a setting means 1311 before the first extraction means 1312. Figure 14 includes the setting means 1311 in its illustration. These means in the management system 1300 correspond to the setting step S11, first extraction step S12, first identification step S13, discrimination step S14, selection step S15, second extraction step S16, second identification step S17, creation step S18, and determination step S19 in the management method described above.

[0107] The setting means 1311 sets product characteristic targets 70 in order to manufacture the target product. This corresponds to step S1 and setting step S11 described above. The first extraction means 1312 extracts important quality characteristics 71 (quality characteristics) that may correlate with the product characteristic targets 70 described above, i.e., candidates for quality characteristics related to the quality in the manufacture of the target product and / or candidates for important productivity characteristics 72 (productivity characteristics) related to the productivity in the manufacture of the product. This corresponds to step S2 and first extraction step S12 described above. The first identification means 1313 identifies the target important quality characteristics 71 and / or important productivity characteristics 72 based on evidence from the extracted candidates for important quality characteristics 71 and / or important productivity characteristics 72. This corresponds to step S3 and first identification step S13 described above. The determination means 1314 determines whether the manufacturing process in which the product is manufactured is in a steady state or a transient state. This corresponds to part of step S6 and determination step S14 described above. The selection means 1315 selects any one time within the manufacturing process to create a design space if the state is steady, and selects one or more times within the manufacturing process to create a design space if the state is not steady. This corresponds to part of step S6 and the selection step S15 described above. The second extraction means 1316 extracts candidates for critical process parameters 73 that cause the critical quality characteristics 71 and / or critical productivity characteristics 72 identified above. This corresponds to step S4 and the second extraction step S16 described above. The second identification means 1317 identifies a critical process parameter 73 from the candidates for critical process parameters 73 extracted above, based on evidence. This corresponds to step S5 and the second identification step S17 described above. The creation means 1318 determines the range of the critical process parameters 73 identified above and creates a design space. This corresponds to step S7 and the creation step S18 described above. The determination means 1319 determines the manufacturing conditions based on the design space created above. This corresponds to step S8 and the determination step S19 described above.

[0108] The fluid analysis device 1302, literature search device 1303, correlation analysis device 1305, laboratory experiment device 1306, and design space creation device 1307 included in the management system 1300 shown in Figure 13 may each be individual devices. However, if a computer 1310, CPU 1320, GPU, etc. functions as the setting means 1311, first extraction means 1312, first identification means 1313, discrimination means 1314, selection means 1315, second extraction means 1316, second identification means 1317, creation means 1318, and determination means 1319 shown in Figure 14, then the computer 1310, etc. may be considered to correspond to the fluid analysis device 1302, literature search device 1303, correlation analysis device 1305, and design space creation device 1307. In other words, the management system 1300 may not only consist of the individual devices described above, but the computer 1310, CPU 1320, GPU, etc. may also perform the functions and roles of those devices.

[0109] The manufacturing condition management system 1300 described above comprises one or more computers 1310 as shown in Figure 14. In Figure 14, the computer 1310 comprises a CPU 1320, a storage unit 1321, a communication port 1322, an input / output port 1323, and a media port 1324. Here, the storage unit 1321 comprises a RAM 1321a, a ROM 1321b, and an SSD (Solid State Drive) 1321c.

[0110] The communication port 1322 is connected to the communication circuit 1325. The input / output port 1323 is connected to the input / output device 1326. The input / output device 1326 is a display device such as a flat panel display. The media port 1324 reads and writes data to and from the recording medium 1327. The ROM 1321b stores the IPL (Initial Program Loader) and other data executed by the CPU 1320. The SSD 1321c stores application programs and various data. The CPU 1320 executes application programs and other data read from the SSD 1321c into the RAM 1321a, thereby realizing various functions such as the setting means 1311, the first extraction means 1312, the first identification means 1313, the discrimination means 1314, the selection means 1315, the second extraction means 1316, the second identification means 1317, the creation means 1318, and the determination means 1319 shown in Figure 14. Furthermore, various functions may be distributed across multiple computers 1310. Incidentally, the management system 1300 shown earlier in Figures 13 and 14 primarily represents functions implemented by application programs and the like as blocks.

[0111] The management system 1300 is equipped with the setting means 1311 to determination means 1319, particularly the first extraction means 1312 to determination means 1319, and as described above, it can efficiently create a highly accurate design space while considering various factors and parameters. Furthermore, the management system 1300 can efficiently select important process parameters 73 that affect the product characteristic target 70 and grasp the appropriate range with fewer experiments, thereby reducing unexpected fluctuations in productivity and quality before and after scaling up or when changing manufacturing conditions. In addition, the management system 1300 classifies parameters into secondary important process parameters 82 and primary important process parameters 81, and reduces redundant parameters by performing sensitivity analysis on each parameter, thereby reducing the time and cost required to collect necessary experimental data by creating a design space in a part of the manufacturing process. It goes without saying that the preferred embodiments described in the management method described above can also be applied to the management system 1300. It also goes without saying that the same effects as the example use case described in the management method described above can be obtained with the management system 1300.

[0112] [Control of Manufacturing Conditions in Antibody Drug Production] As part of the production of antibody drugs, we investigated the control of manufacturing conditions in the manufacturing process, which involves culturing Chinese Hamster Ovary (CHO) cells and separating and purifying antibody proteins secreted from CHO cells.

[0113] Figure 15 is a flowchart illustrating the outline of the manufacturing process for producing antibody proteins. As shown in Figure 15, this manufacturing process is broadly composed of (1) seed culture, (2) production culture, (3) purification, and (4) filling. (1) The seed culture process increases the number and volume of cells. (2) The production culture process cultures cells in large quantities to produce antibodies. (3) The purification process removes cells and impurities and recovers the produced antibodies. (4) The filling process fills vials with the recovered antibodies.

[0114] While the control of manufacturing conditions is considered for all processes as described above, in this embodiment, it is difficult to identify the important process parameters 73 due to the involvement of complex biological mechanisms, and the important process parameters 73 have a significant impact on the product characteristic target 70. (2) An example of determining manufacturing conditions for the production culture process is described. Note that the manufacturing conditions for other processes can be determined using a similar procedure.

[0115] The production culture process described in (2) above will now be explained. The production culture apparatus used in this process mainly consists of a production culture tank, a stirrer, aeration mechanism, inline measuring instruments for pH, temperature, and dissolved oxygen (DO), and a control device (culture controller). During culture, oxygen necessary for cell respiration (consuming oxygen and releasing carbon dioxide) was supplied by aeration while stirring. In addition, the values ​​obtained from various measuring instruments were controlled to reach various set values. If the pH was higher than the set value, carbon dioxide was aerated to the surface of the culture medium, and if it was lower than the set value, an alkaline solution was added. If the DO was higher than the set value, no control was made (DO was higher than the set value), and if it was lower than the set value, air or oxygen was aerated through the aeration mechanism to raise the DO. In order to culture cells in the culture apparatus, the CHO cells that had been proliferated in the (1) seed culture process were aseptically added to the production culture tank which was pre-filled with fresh culture medium. This process facilitates the transition from the seed culture tank to production culture and dilution. Cells were grown while controlling temperature, DO, pH, and stirring speed to set values, and antibodies were produced simultaneously (see Figure 16). Figure 16 is a graph showing an example of the growth curve and the change in antibody quantity over time. In the example shown in Figure 16, additional culture medium was added from the 5th day after the start of culture. Also, in the example shown in Figure 16, the culture medium was collected on the 13th day after the start of culture, and the process proceeded to (3) purification step.

[0116] In the process described above, the manufacturing conditions were examined. First, in step S1, product characteristic targets 70 were set. In this embodiment, the product characteristic targets 70 related to quality were defined as therapeutic effect and side effects, and the product characteristic target 70 related to productivity was defined as the amount of antibody produced relative to the raw material costs in manufacturing.

[0117] Next, in step S2, candidate key quality characteristics 71 (quality characteristics) and key productivity characteristics 72 (productivity characteristics) were extracted based on experimental data and published literature conducted so far, in relation to the product characteristic target 70 defined in step S1 (see Table 1).

[0118]

[0119] However, since the candidates extracted in step S2 do not necessarily show a correlation in the manufacturing process in this embodiment, in step S3, the extracted candidates were prioritized based on the likelihood of correlation and the severity (magnitude of impact) if a correlation occurred, and experiments were conducted in order of priority. Then, as a result of examining the important quality characteristics 71 and important productivity characteristics 72 for the product characteristic target 70, it was determined that the important productivity characteristic 72 was the antibody production amount per culture tank. The important quality characteristic 71 was determined to be the glycan structure (for example, the proportion of glycans at a specific site of the antibody). Specifically, the proportion of glycans refers to the proportion of glycans on Asn297 of the antibody product. If the glycans of Asn297 are deficient, it may cause a decrease or loss of activity, a decrease in thermal stability, or an increase in polymers. Therefore, the manufacturing conditions are determined so as to maximize the proportion of glycans added to Asn297 or to be above a predetermined proportion. In this case, productivity is not considered, and experimental data and literature information regarding productivity are not used. On the other hand, when the manufacturing conditions change during culture, the metabolism of the cells (microorganisms) changes, which not only changes the rate of glycan addition but also changes the growth rate of the cells (microorganisms) and the secretion rate of the target substance. As a result, both the rate of glycan addition of Asn297 and the productivity of the antibody change. In other words, the important quality characteristic 71 and the important productivity characteristic 72 are correlated and, in some cases, have a trade-off relationship. In this case, if only the important quality characteristic 71 is considered, the productivity of the target substance will decrease significantly. In cases where there is a correlation between the important quality characteristic 71 and productivity as described above, it is necessary to balance the important quality characteristic 71 and the important productivity characteristic 72. In this example, both the important quality characteristic 71 and the important productivity characteristic 72 were considered. First, the minimum required quality (in the above example, the rate of glycan addition of Asn297) was determined, and productivity was maximized within the range that satisfies this. To accomplish this, multi-objective optimization was used with important quality characteristic 71 and important productivity characteristic 72 as the objective variables.

[0120] Next, in step S4, candidates for important process parameters 73 were extracted based on the results of experiments conducted to date and published literature for the important quality characteristics 71 and important productivity characteristics 72 determined in step S3 (see Table 1). However, since the candidates extracted in step S4 do not necessarily show a correlation in the manufacturing process in this embodiment, in step S5, the extracted candidates were prioritized based on the likelihood of correlation and the severity (magnitude of influence) if a correlation occurred, and experiments were conducted in order of priority. Then, the important process parameters 73 for the product characteristic target 70 were identified as growth rate, viable cell density, and cell viability (see Table 1). For some candidates, no new experiments were conducted, and already accumulated experimental results were used.

[0121] Figure 17 is a correlation diagram showing the correlation of each parameter in the example. Figure 18 is a correlation diagram showing the correlation of the secondary important process parameter 82 and the primary important process parameter 81 in the example. In the steps described above, candidate product characteristic targets 70, candidate important quality characteristics 71, candidate important productivity characteristics 72, and candidate important process parameters 73 were determined. These candidates were found to have the correlation potential shown in Figure 17. Furthermore, it was found that there is a correlation between the product characteristic targets 70, important quality characteristics 71, important productivity characteristics 72, and important process parameters 73 as shown in Figure 18. Based on these results, in this example, as shown in Figure 18, the secondary important process parameters 82 were determined to be growth rate, viable cell density, cell viability, dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, mixing time constant, shear stress, and bubbles, and the primary important process parameters 81 were determined to be stirring rotation speed and liquid aeration rate. More specifically, as shown in Figure 18, it was found that the stirring rotation speed, a primary important process parameter 81, correlates with the dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, mixing time constant, and shear stress, which are secondary important process parameters 82. Furthermore, it was found that the amount of air permeation in the liquid, a primary important process parameter 81, correlates with the dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, shear stress, and bubbles, which are secondary important process parameters 82. The dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, mixing time constant, shear stress, and bubbles, which are secondary important process parameters 82, were found to correlate with the growth rate. The mixing time constant, shear stress, and bubbles, which are secondary important process parameters 82, were found to correlate with the number density of viable cells. Furthermore, it was found that the mixing time constant, shear stress, and bubbles, which are secondary important process parameters 82, correlate with the cell viability. In addition, it was found that the dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, and mixing time constant correlate with the proportion of sugar chain addition, which is an important quality characteristic 71. It was found that proliferation rate, viable cell density, and cell viability correlate with antibody levels, which are important productivity characteristics 72.

[0122] Next, in step S6, the time periods within the manufacturing process when design space should be created were determined. Creating design space for every moment of the two-week production culture process is extremely labor-intensive. Therefore, based on the five secondary critical process parameters 82 of the five performance requirements determined above—oxygen supply capacity, carbon dioxide removal capacity, shear stress, mixing rate, and foaming—the times when harsh conditions would be anticipated were identified to determine when design space should be created.

[0123] Oxygen supply capacity: The objective is to supply oxygen to the amount of oxygen consumed by cells through respiration via aeration and agitation. Therefore, an oxygen supply capacity greater than the oxygen consumption rate by the total number of cells in the tank is required. The required oxygen supply capacity is highest from the late logarithmic growth phase, when the number of viable cells is large and their respiration activity is high, until the number of viable cells plateaus (see Figure 16).

[0124] Carbon dioxide removal capacity: Since the purpose is to remove carbon dioxide expelled by cells through respiration by aeration and agitation, a carbon dioxide removal capacity greater than the rate of increase of carbon dioxide accumulating in the tank is required. The required carbon dioxide removal capacity is highest from the late logarithmic growth phase, when the number of viable cells is large and the respiration activity of viable cells is high, until the number of viable cells plateaus (see Figure 16).

[0125] Shear stress: Cells are damaged by shear stress from the impeller during aeration and stirring. The effect varies depending on the rotation speed of the impeller, so it is sufficient if the maximum rotation speed during the culture period does not damage the cells. In this culture, the rotation speed of the impeller is constant throughout the culture period, so the design space can be chosen at any time.

[0126] Mixing Rate: When adding alkaline solutions or feed media for pH control during culture, slow diffusion of the solution after addition can lead to uneven solution concentration (high-concentration areas), potentially affecting cells. In this culture, the effect of adding alkaline solutions for pH control is greater than that of adding feed media, so the appropriate mixing rate for alkaline addition was investigated experimentally. Adding alkaline solutions for pH adjustment is performed throughout the culture period. Therefore, the design space can be created at any time when adding alkaline solutions, but considering the increased frequency of additions, it is more desirable to create the design space in the later stages of culture when cells are proliferating and secreting of lactic acid, an acidic substance, is high.

[0127] Foaming: When aeration is performed, a foam layer forms on the surface of the culture medium. Depending on the aeration rate, the foam and foam layer in the culture medium may damage the cells. Therefore, the appropriate aeration rate (aeration tower velocity) that does not damage the cells should be determined through experimentation. The aeration rate is highest at the same time as the time when oxygen supply is at its highest, which is from the late logarithmic growth phase to the time when the cell number plateaus (see Figure 16).

[0128] All of the secondary critical process parameters 82 can be expressed using two primary critical process parameters 81: wingtip velocity and ventilator velocity. While each secondary critical process parameter 82 may be evaluated at different times and superimposed to create a design space, in this embodiment, the design space was created in the next step S7 by unifying the time shown by ★ in Figure 16 (day 9).

[0129] In step S7, a design space was created. In this step S7, correlation equations were formulated between the critical quality characteristics 71 and critical productivity characteristics 72 and the critical process parameters 73, and the range of conditions under which manufacturing is possible was determined as the design space. Generally, the design space determines the range of critical process parameters 73 that satisfy the requirements of the critical quality characteristics 71 and critical productivity characteristics 72. However, in this embodiment, as described above, the design space was created as the range of primary critical process parameters 81 that satisfy the requirements of secondary critical process parameters 82. First, the conditions that the secondary critical process parameters 82 must satisfy were determined experimentally. The experiment is described below.

[0130] <Determination of the appropriate range of secondary critical process parameter 82 through laboratory experiments> The culture experiments conducted to determine the appropriate range of secondary critical process parameter 82 are described below. • Experimental apparatus Figure 19 is a schematic diagram showing the configuration of a 1L culture apparatus. Although the capacity and material of the culture tank differ, it mainly consists of a 1L culture tank, a stirrer (agitator blade), aeration mechanism, in-line measuring instruments for pH, temperature (TE), and dissolved oxygen (DO), and a control device (culture controller), similar to the production culture apparatus. During culture, oxygen necessary for cell respiration (consuming oxygen and releasing carbon dioxide) is supplied by aeration while stirring. A constant amount of aeration was performed on the sintered sparger and the culture medium surface at all times. In addition, the values ​​obtained from various measuring instruments were controlled to meet various set values.

[0131] • Culture Method: For the culture experiment, Chinese hamster ovary cells (CHO cells; CRL-12445 cells) that produce antibody proteins were purchased from the American Type Culture Collection (ATCC), and the adherent cells were used in suspension. The culture medium used was Dulbecco's Modified Eagle Medium supplemented with Fetal Bovine Serum (FBS) (final concentration 10%). Culture was performed in a 1L culture tank. The seeding density was 2 × 10⁶ ５During cell / mL culture, dissolved oxygen, pH, and temperature were controlled under various conditions. Sterile sampling was performed 1-2 times per day, and the culture medium components were analyzed. Alkaline solution was added as needed to adjust the pH. From the sampled culture medium, (1) the number of viable cells and (2) the culture medium components (glucose, glutamine, lactate, ammonia, antibody protein) were quantified. The analytical method is described below.

[0132] (1) Counting of live cells The number of live cells was determined using the Vi-CELL (Beckman Coulter) cell viability and death detection device. The cell culture medium was placed in the Vi-CELL, live and dead cells were distinguished by trypan blue staining, image data of the cells was acquired, and the number of live cells was obtained by automatic counting.

[0133] (2) Analysis of culture medium components (glucose, glutamine, lactate, ammonia, antibody protein) Glucose, glutamine, lactate, and ammonia in the culture medium were measured using Cedex (Roche).

[0134] Experiments on each secondary critical process parameter 82 The tolerance range of the secondary critical process parameter 82 (usually critical quality characteristics 71 or critical productivity characteristics 72) was determined using the experimental apparatus described above in the following experiment.

[0135] Oxygen supply capacity: The oxygen supply capacity must be higher than the rate at which cells consume oxygen through respiration. Therefore, the required oxygen supply capacity is determined by the rate at which cells consume oxygen. In this example, a dynamic method was used to calculate the oxygen supply capacity from the change in DO over time when oxygen supply was stopped and then restarted during culture. Figure 20 is a graph showing an overview of the change in DO over time when determining the oxygen consumption rate of cells. The change in DO over time is as shown in Figure 20, and the slope in the graph is given by the equation shown in the figure (dDO / dt = kLa(DO*-DO)-r ｏ２ Since it can be expressed as ), by obtaining the change in DO over time through culture experiments, the oxygen consumption rate (OUR = r ｏ２ ) can be calculated. Note that DO* indicates the saturated oxygen concentration.

[0136] The oxygen consumption rate by cells was determined by the following culture experiment. Cells were cultured in a 1L culture tank using the same culture method as production culture. Oxygen supply was stopped on the 7th day of culture, and oxygen supply was restarted when the DO value of the culture medium reached 10% of the saturated oxygen concentration. The maximum oxygen consumption rate OUR at the total number of cells was determined by the formula described in the principle above. ｍａｘ The following was calculated. Therefore, the oxygen supply capacity is the maximum oxygen consumption rate OUR ｍａｘ The above method was determined. Separately from the above method, if the oxygen consumption rate per cell is known from literature or experiment, the maximum oxygen consumption rate OUR can be calculated by multiplying it by the assumed maximum number of cells. ｍａｘ It is possible to find this.

[0137] Carbon dioxide degassing capacity: The carbon dioxide degassing capacity must be higher than the rate at which cells expel carbon dioxide through respiration. Therefore, the required carbon dioxide degassing capacity is determined by the rate at which cells expel carbon dioxide. The rate at which cells expel carbon dioxide can be determined from the mass balance. (Amount of carbon dioxide in the gas discharged from the culture vessel) = (Amount of carbon dioxide in the input gas) + (Amount of carbon dioxide expelled by cells)

[0138] The rate of carbon dioxide elimination by cells was determined through the following culture experiment. On the 7th day of culture, a flow meter and gas analyzer were installed on the gas discharged from the culture tank, and time-dependent data of the changes in each analytical value (flow rate, carbon dioxide content ratio) were obtained. In addition, the input gas (oxygen supply gas and CO2 for pH control) was also analyzed. ２ Similarly, flow meters and gas analyzers were installed for the gas to measure the flow rate and carbon dioxide content. The carbon dioxide emission rate released by the cells was calculated from the above values ​​and the mass balance formula. From the above, the carbon dioxide degassing capacity is the maximum carbon dioxide emission rate CER. ｍａｘ The above method was determined. Separately from the above method, if the carbon dioxide efflux rate per cell is known from literature or experiment, the maximum carbon dioxide efflux rate CER can be calculated by multiplying it by the assumed maximum number of cells. ｍａｘ It is possible to find this.

[0139] Shear stress: Shear stress is distributed within the culture vessel and is difficult to measure directly. Therefore, the shear stress distribution when the stirring speed is changed, based on the structure of the culture vessel used in the culture experiment, was quantified by fluid analysis (simulation). Conventionally, inclined paddle-type impellers are often used, but in this case, even at the maximum rotation speed, the magnitude of the shear stress after scaling up is not reached. Therefore, in this study, the impeller diameter was increased, and culture experiments were conducted at various rotation speeds using elongated flat paddles and disk turbine impellers to evaluate the growth rate, cell viability, and antibody production, and the appropriate shear stress range was determined.

[0140] Production culture tanks are 100L or 10m ３ Because the resulting culture tank is larger than a 1L tank, a 1L tank cannot generate sufficient shear stress. When considering manufacturing conditions for such a large culture tank in production, in addition to the 1L tank, a culture device capable of generating high shear stress with a small capacity is required. In this example, in addition to the 1L tank, the appropriate shear stress range is determined using the shear stress effect evaluation device shown in Figure 21. Figure 21 is a schematic diagram showing the configuration of a shear stress effect evaluation device equipped with a flow chamber type culture device.

[0141] The shear stress effect evaluation device shown in Figure 21 measures metabolic changes in cells while they are exposed to a uniform shear stress. To this end, the shear stress effect evaluation device is designed to detect changes in the concentrations of components (glucose, glutamine, lactate, ammonia) of the culture medium and metabolites, as well as the dissolved oxygen concentration. The shear stress effect evaluation device consists of (1) a flow chamber, (2) a dissolved oxygen (DO) measurement unit (DO electrode flow cell equipped with a DO electrode), (3) a culture medium preparation tank, (4) a liquid delivery pump (peristallic pump), and (5) a pulse damper, with each component connected using low-gas-permeability Pharmed tubing. In the culture medium preparation tank, the DO, pH, and temperature of the culture medium are controlled. The culture medium is delivered by the liquid delivery pump (peristallic pump), passes through a pulse damper to suppress pump pulsation, and then delivered to the flow chamber where the cells are fixed. The flow chamber outlet is equipped with a DO electrode on the flow cell, allowing for the measurement of the oxygen concentration consumed by the cells by taking the difference between the dissolved oxygen in the culture medium preparation tank and the DO electrode. Details of each component are described below. Based on the results of the culture experiment described above, the appropriate shear stress range is the maximum shear stress t ｍａｘ It was determined to be as follows:

[0142] Tolerable foam height: An experiment was conducted to evaluate the effect of bubbles on the secondary critical process parameter 82. To investigate the effect of bubbles, it is necessary to supply bubbles to the culture medium while maintaining a constant dissolved oxygen level. Therefore, in this culture experiment, oxygen in the culture medium was forcibly removed from the liquid surface by continuously blowing nitrogen onto the liquid surface, and bubbles were supplied by supplementing the removal of oxygen through submerged aeration. In a water test, it was confirmed that there was no difference in dissolved oxygen concentration near the liquid surface (1 cm below the liquid surface) and at the bottom of the culture tank. In liquid surface aeration, carbon dioxide was mixed in appropriate amounts in addition to nitrogen to maintain pH. In submerged aeration, air or pure oxygen was aerated to investigate the effect of different gas types. The extent to which cells and antibody proteins were damaged on day 7 of culture was investigated at various aeration rates (aeration tower velocity). The foam height for each aeration rate was also measured. The foam height at which no change occurred in cell number, viability, or culture medium components due to aeration was defined as the tolerable foam height hf. ｍａｘ That was the decision.

[0143] <Correlation formula between secondary critical process parameter 82 and primary critical process parameter 81> The correlation formula between the determined secondary critical process parameter 82 and primary critical process parameter 81 was formulated as follows, based on fluid dynamics calculations in chemical engineering.

[0144] - Oxygen supply capacity: If OUR is the rate of oxygen consumption by cells, then kLab is the oxygen transfer capacity coefficient for aeration in the liquid, kLas is the oxygen transfer capacity coefficient at the surface of the culture medium, and ΔpO is the difference between the saturated dissolved oxygen concentration and the dissolved oxygen concentration in the culture medium. ２ The difference between the oxygen concentration at the surface of the culture medium and the dissolved oxygen concentration in the culture medium is ΔpO ２ If we let it be s, it can be expressed by equation 8: OUR = kLab・ΔpO ２ +kLas・ΔpoO ２ s...(Formula 8)

[0145] If we let N be the tip velocity (rotational speed) and Us be the ventilator velocity, then from chemical engineering calculations, kLab and kLas can be expressed as functions f(N,Us) and g(N,Us), respectively, with N and Us as variables, and OUR = f(N,Us)・ΔpO ２ +g(N,Us)・ΔpO ２ s ... (Equation 9) ΔpO ２ and ΔpO ２ s can be determined experimentally, and OUR can be expressed using the primary critical process parameters 81, namely the blade tip velocity (rotational speed) N and the aeration tower velocity Us.

[0146] If CER is the rate of carbon dioxide secretion from cells capable of degassing carbon dioxide, then kLab' is the carbon dioxide transfer capacity coefficient for aeration in the liquid, kLas' is the carbon dioxide transfer capacity coefficient at the surface of the culture medium, and ΔpCO2 is the difference between the saturated dissolved carbon dioxide concentration and the dissolved carbon dioxide concentration in the culture medium. ２ The difference between the carbon dioxide concentration at the surface of the culture medium and the dissolved carbon dioxide concentration in the culture medium is ΔpCO2. ２ If we let it be s, it can be expressed by equation 10. CER = CO ２ kLab'・ΔpCO ２ +kLas'・ΔpCO ２s...(Formula 10)

[0147] If we let N be the tip velocity (rotational speed) and Us be the ventilator velocity, then from chemical engineering calculations, kLab' and kLas' can be expressed as functions h(N,Us) and j(N,Us), respectively, with N and Us as variables, and CER = h(N,Us)・ΔpCO ２ +j(N,Us)・ΔpCO ２ s ... (Equation 11) ΔpCO ２ and ΔpCO ２ s can be determined experimentally, and CER can be expressed in terms of the primary critical process parameters 81, namely the blade tip velocity (rotational speed) N and the ventilator velocity Us. However, the relationships kLab ≈ kLab' and kLas = kLas' hold true.

[0148] From chemical engineering calculations of shear stress, the shear stress t can be expressed by equation 12, and can be expressed in terms of the primary critical process parameters 81, namely the blade tip velocity (rotational speed) N and the ventilator velocity Us. t = c １ ・N・Di+c ２ ・N・Di・Us+c ３ • Us … (Equation 12) Here, c １ , c ２ , c ３ is a system-dependent constant, and Di is the diameter of the impeller.

[0149] Mixing rate Q ｉ / V can be expressed by equation 13 and can be expressed in terms of the primary critical process parameters 81, namely the blade tip velocity (rotational speed) N and the ventilator velocity Us. Q ｉ / V=c ４ ・N・Di ３ / V+c ５ Us / V １／３ …(Equation 13) Here, c ４ , c ５ V is a system-dependent constant, Di is the diameter of the stirring blade, and V is the volume of the culture medium.

[0150] From chemical engineering calculations for allowable foam height, the allowable foam height Hf can be expressed by equation 14, and can be expressed in terms of the primary critical process parameters 81, namely the blade tip velocity (rotational speed) N and the venting tower velocity Us. Hf = c ６ ・Us・Exp^(-c ７ ・N ３ ・Di ５ / V) ... (Equation 14) Here, c ６ , c ７ V is a system-dependent constant, Di is the diameter of the stirring blade, and V is the volume of the culture medium.

[0151] In this embodiment, the secondary critical process parameter 82 was formulated using the primary critical process parameter 81.

[0152] <Boundaries of the design space> Our determined through experimentation ｍａｘ、 CER ｍａｘ ,t ｍａｘ Q ｉ / V ｍａｘ , Hf ｍａｘ Using equations 8 through 14, which were formulated through chemical engineering calculations, we created a design space. ｍａｘ ≦f(N,Us)・ΔpO ２ +g(N,Us)・ΔpO ２ s...(Formula 15) CER ｍａｘ ≦h(N,Us)・ΔpCO ２ +j(N,Us)・ΔpCO ２ s...(Formula 16) t ｍａｘ ≥ c １ ・N・Di+c ２ ・N・Di・Us+c ３ • Us … (Equation 17) Q ｉ / V ｍａｘ ≥ c ４ ・N・Di ３ / V+c ５ Us / V １／３ …(Equation 18) Hf ｍａｘ ≥ c６ ・Us・Exp^(-c ７ ・N ３ ・Di ５ / V) …(Equation 19)

[0153] OUR ｍａｘ = f(N, Us)・ΔpO ２ + g(N, Us)・ΔpO ２ s …(Equation 20) CER ｍａｘ = h(N, Us)・ΔpCO ２ + j(N, Us)・ΔpCO ２ s …(Equation 21) t ｍａｘ = c １ ・N・Di + c ２ ・N・Di・Us + c ３ ・Us …(Equation 22) Q ｉ / V ｍａｘ = c ４ ・N・Di ３ / V + c ５ ・Us / V １／３ …(Equation 23) Hf ｍａｘ = c ６ ・Us・Exp^(-c ７ ・N ３ ・Di ５ / V) …(Equation 24)

[0154] Using equations 15 to 19, the ranges that the primary critical process parameters 81, namely the blade tip velocity and the ventilator velocity, must satisfy can be determined. To draw the boundary line in the design space, equations 20 to 24, which are made into equals, must be solved analytically. In this embodiment, however, a solver from calculation software was used. The results are shown in Figure 22. Figure 22 is an explanatory diagram showing an example of the design space created in this embodiment. Note that in equation 20, the change in blade tip velocity in response to the change in ventilator velocity was too large, so the solver did not function properly. Therefore, in this embodiment, the problem caused by the solver was solved by taking an inverse mapping, and the boundary line in equation 20 could be drawn. In this way, the design space (Figure 22) created in step S7 was created.

[0155] Next, in step S8, the conditions for the manufacturing apparatus were determined using the design space created in step S7. Figure 23 is an explanatory diagram showing an example of the design space before and after scale-up. As shown in the left diagram of Figure 23, by creating a design space for a small-scale culture tank (1000+L) slightly exceeding 1000L in laboratory experiments, it was possible to create a design space for a large-scale culture tank (10000+L) exceeding 10000L (right diagram of Figure 23) that was subsequently anticipated. For commercial manufacturing conditions, the manufacturing conditions within the cultureable area of ​​the design space in the right diagram of Figure 23 were determined as the actual operating conditions for the apparatus. Furthermore, since errors occur in operation control, the design and specifications of the apparatus were determined so that even including these errors, they would fit within the aforementioned design space.

[0156] On the other hand, even with the same scale of manufacturing equipment, the design space changes when the cell density increases. In this case, even with the same equipment, manufacturing could be carried out by setting the operating conditions to the design space shown in Figure 24, which represents the cultureable area. Figure 24 is an explanatory diagram showing an example of the design space when the cell number density is different. The left side of Figure 24 shows the design space when culturing at a standard cell number density in a 1000+ L culture tank. The right side of Figure 24 shows the design space when culturing at a high cell number density in a culture tank of the same scale (1000+ L). As shown in the right side of Figure 24, even with the same scale of manufacturing equipment, the cultureable area becomes smaller when the cell density increases.

[0157] [Examples of substance production by Corynebacterium] Figure 25 is an explanatory diagram showing the metabolic map of Corynebacterium. As shown in Figure 25, Corynebacterium can produce various useful substances through intracellular metabolism using glucose as a raw material. In this example, we investigated the conditions for producing arginine by culturing Corynebacterium.

[0158] The Corynebacterium strain used in this embodiment was genetically modified to metabolize arginine by altering the enzymes in each metabolic pathway. The strain was cultured in suspension under controlled conditions, and after the bacteria proliferated, arginine was recovered in a purification step. Since the culture step is the most important step in the manufacturing process, this embodiment describes the determination of the manufacturing conditions in the culture step.

[0159] Unlike antibodies, which have variants with different sugar chains, it is important that arginine is metabolized without stopping midway through metabolism or excessively branching into metabolic pathways, as can be seen from the metabolic pathway shown in Figure 25. Therefore, since maximizing arginine production reduces the impurities of intermediate metabolites, in this example, we focused only on the important productivity characteristics 72, rather than the important quality characteristics 71, when investigating manufacturing conditions.

[0160] The production culture apparatus used in the culture process of this embodiment mainly consists of a production culture tank, a stirrer, aeration mechanism, inline measuring instruments for pH, temperature, and dissolved oxygen (DO), and a control device (culture controller) (see Figure 19). During culture, oxygen necessary for the respiration of Corynebacterium (consuming oxygen and releasing carbon dioxide) was supplied by aeration while stirring. In addition, the values ​​obtained from various measuring instruments were controlled to reach various set values. If the pH was higher than the set value, no control was performed, and if it was lower than the set value, an alkaline solution was added. A constant amount of aeration was performed at all times on the sintered sparger and the culture medium surface. If the DO was higher than the set value, no control was performed (DO was higher than the set value), and if it was lower than the set value, the DO was increased by increasing the stirring speed. In order to culture cells in the culture apparatus, Corynebacterium grown in the seed culture process was aseptically inoculated into the production culture tank, which was pre-filled with fresh culture medium. This process facilitates the transfer from the seed culture tank to the production culture and dilution. Corynebacterium was grown while controlling the temperature, DO, pH, and stirring speed to set values. After 72 hours from the start of cultivation, the culture medium was collected and the process moved to the purification stage.

[0161] In the process described above, the manufacturing conditions were examined. First, as step S1, a target product characteristic 70 was set. In this example, as mentioned earlier, no important quality characteristic 71 was set, and the target product characteristic 70 related to productivity was set as the amount of arginine produced relative to the raw material glucose.

[0162] Next, in step S2, candidates for important productivity characteristics 72 were extracted based on experimental data and published literature from previous steps, in relation to the product characteristic target 70 defined in step S1 (see Table 1). However, since the candidates extracted in step S2 do not necessarily show a correlation in the manufacturing process in this embodiment, in step S3, the extracted candidates were prioritized based on the likelihood of correlation and the severity (magnitude of influence) if a correlation occurs, and experiments were conducted in order of priority. The important productivity characteristics 72 were then defined as the growth rate of Corynebacterium, absorbance (OD) reflecting the bacterial density, and metabolic rate in each metabolic reaction shown in Figure 25. For some candidates, no new experiments were conducted, and already accumulated experimental results were used.

[0163] Next, in step S4, candidates for important process parameters 73 were extracted based on the results of experiments conducted so far and published literature for the important productivity characteristics 72 determined in step S3 (see Table 1). However, the candidates extracted in step S4 do not necessarily show a correlation in the manufacturing process in this embodiment. Therefore, in step S5, the extracted candidates were prioritized based on the likelihood of correlation and the severity (magnitude of influence) if a correlation occurred, and experiments were conducted in order of priority. Then, the important process parameters 73 for the important productivity characteristics 72 were DO (dissolved oxygen concentration), DCO ２ The following factors were identified: (dissolved carbon dioxide concentration), temperature, pH, timing of inducer addition, amount of inducer added, aeration rate in the liquid, maximum stirring speed, and minimum stirring speed. Among these, DO and DCO were identified. ２ The following were identified as secondary critical process parameters 82. In addition, temperature, pH, timing of inducer addition, amount of inducer addition, aeration rate in the liquid, maximum stirring speed, and minimum stirring speed were identified as primary critical process parameters 81. For some candidates, no new experiments were conducted, and existing experimental results were used.

[0164] Next, in step S6, the time points within the manufacturing process for which design spaces should be created were determined. Creating design spaces for all time points in the 72-hour production culture process is extremely labor-intensive. Therefore, based on the secondary critical process parameter 82 determined above, the time points for which design spaces should be created were identified by considering which times would be the most severe conditions. As a result, one of the times was when the inducer was added, and the other was the time when the bacterial count (OD) was maximum during the growth phase of the bacteria at different times. For each of these two time points, small-scale (laboratory experiment) design spaces (with secondary critical process parameter 82 as the dependent variable and primary critical process parameter 81 as the independent variable) and large-scale (commercial) design spaces (with secondary critical process parameter 82 as the dependent variable and primary critical process parameter 81 as the independent variable) were created (step S7, figure omitted).

[0165] In step S7, one manufacturing condition was selected from the cultureable region of the large-scale design space obtained, and this was adopted as the manufacturing condition for commercial production. The specifications of the manufacturing equipment were then determined so that the accuracy range of each primary critical process parameter 81 fell within the cultureable region (step S8). As described above, in this embodiment, steps S1 to S8 allowed for the efficient creation of a highly accurate design space while considering various factors and parameters related to substance production by Corynebacterium.

[0166] [Control of Manufacturing Conditions in the Production of Therapeutic Cells] Human mesenchymal stem cells (hMSCs) are cultured to increase their cell count, then appropriately differentiated, and finally transplanted into the human body for therapeutic purposes. In this example, we investigated the manufacturing conditions for a production method in which undifferentiated hMSCs are proliferated and then differentiated into nerve cells.

[0167] Unlike the suspension culture example described above, hMSCs grow planarly, adhering to the bottom of the culture vessel. hMSCs were seeded in a culture flask, and when they grew and approached confluence, a cell detachment agent (trypsin was used in this example) was used to detach the cells from the bottom of the culture vessel. After collecting the cells using a centrifuge, they were diluted with fresh medium and inoculated into a larger culture flask. During culture, an incubator (5% CO2) was used. ２ The cells were left standing (under saturated steam). After the cells had grown to the desired number, fibronectin was coated on the bottom of the culture vessel, and the detached cells were seeded on top of it. To induce differentiation into nerve cells, the cells were cultured using mesenchymal stem cell neuronal differentiation medium (Promo Cell), and the medium was changed every three days. When the nerve cells reached the desired proportion, the nerve cells were harvested.

[0168] We investigated the manufacturing conditions for inducing nerve cells from the aforementioned hMSCs. The production culture apparatus used in the culture process mainly consists of a production culture vessel and an incubator. First, as step S1, we set the target product characteristic 70. In this example, the proportion of undifferentiated cells was set as the target product characteristic 70 before differentiation induction, and the proportion of nerve cells was set as the target product characteristic 70 after differentiation induction.

[0169] Next, in step S2, candidates for important quality characteristics 71 and important productivity characteristics 72 were extracted based on experimental data and published literature conducted to date, with respect to the product characteristic target 70 defined in step S1 (see Table 1). However, since the candidates extracted in step S2 do not necessarily show a correlation in the manufacturing process in this embodiment, in step S3, the extracted candidates were prioritized based on the likelihood of correlation and the severity (magnitude of impact) if a correlation occurs, and experiments were conducted in order of priority. The important productivity characteristic 72 was defined as the number of cells, and the important quality characteristics 71 were defined as the percentage of cells expressing membrane proteins in undifferentiated cells and the percentage of cells expressing membrane proteins in nerve cells. For some candidates, no new experiments were conducted, and the results of already accumulated experiments were used.

[0170] Next, in step S4, candidates for important process parameters 73 were extracted based on the results of experiments conducted so far and published literature for the important productivity characteristics 72 determined in step S3 (see Table 1). However, the candidates extracted in step S4 do not necessarily show a correlation in the manufacturing process in this embodiment. Therefore, in step S5, the extracted candidates were prioritized based on the likelihood of correlation and the severity (magnitude of influence) if a correlation occurred, and experiments were conducted in order of priority. Then, the important process parameters 73 for the important productivity characteristics 72 were DO (dissolved oxygen concentration), DCO ２ The following factors were identified: (dissolved carbon dioxide concentration), temperature, pH, timing of subculturing, fibronectin density (density of scaffolding material), seeding density, cell uniformity at seeding, shear stress, flow rate during pipetting, cell density at subculturing, and culture period. Among these, DO and DCO were identified. ２ pH and shear force were identified as secondary critical process parameters 82. Temperature, subculturing timing, fibronectin density (scaffolding density), seeding density, cell uniformity at seeding, flow rate during pipetting, cell number density at subculturing, and culture period were identified as primary critical process parameters 81. For some candidates, no new experiments were conducted, and existing experimental results were utilized.

[0171] Next, in step S6, we considered the time points within the manufacturing process when design spaces should be created. As a result, we found that one was immediately before differentiation induction (the time when the cell number density is highest in the undifferentiated state), another was the time of medium change after differentiation induction, another was the time during the subculturing process before differentiation induction, and the third was the time of seeding cells onto fibronectin. For each of these four time points, we created small-scale (laboratory experiment) design spaces (with secondary critical step parameter 82 as the dependent variable and primary critical step parameter 81 as the independent variable) and large-scale (commercial) design spaces (with secondary critical step parameter 82 as the dependent variable and primary critical step parameter 81 as the independent variable) (step S7, figure omitted).

[0172] In step S7, one manufacturing condition was selected from the cultureable region of the large-scale design space obtained, and this was designated as the manufacturing condition for commercial production. The specifications of the manufacturing equipment were then determined so that the accuracy range of each primary critical process parameter 81 fell within the cultureable region (step S8). As described above, in this embodiment, steps S1 to S8 enabled the efficient creation of a highly accurate design space while considering various factors and parameters for the manufacturing conditions in therapeutic cell production.

[0173] The following describes exemplary embodiments of this embodiment. [Embodiment 1] A method for managing manufacturing conditions, comprising: a first extraction step of extracting candidate quality characteristics related to the quality in the manufacture of a target product and / or candidate productivity characteristics related to the productivity in the manufacture of a product; a first identification step of identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or candidate productivity characteristics; a determination step of determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection step of selecting any one time within the manufacturing process to create a design space if it is in a steady state, and selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction step of extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification step of identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation step of determining the range of the identified important process parameters and creating a design space; and a determination step of determining manufacturing conditions based on the design space. [Aspect 2] The manufacturing condition management method according to aspect 1, characterized in that, in the first extraction step, candidates for the quality characteristics and candidates for the productivity characteristics are extracted; in the first identification step, the quality characteristics and productivity characteristics are identified based on evidence for the candidates for the quality characteristics and candidates for the productivity characteristics; and in the second extraction step, candidates for important process parameters that cause the quality characteristics and productivity characteristics are extracted. [Aspect 3] The manufacturing condition management method according to aspect 1, characterized in that the important process parameters are classified into primary important process parameters that have a distant correlation with the quality characteristics and / or the productivity characteristics, and secondary important process parameters that have a close correlation with the quality characteristics and / or the productivity characteristics. [Aspect 4] The manufacturing condition management method according to aspect 3, characterized in that the design space is created with the secondary important process parameters as the objective variable and the primary important process parameters as the explanatory variables.[Aspect 5] The manufacturing condition control method according to Aspect 1, characterized in that the product characteristic target, which is the quality and productivity that the product should satisfy in the manufacturing process, is one or more selected from the group of product function, product yield, and yield. [Aspect 6] The manufacturing condition control method according to Aspect 1, characterized in that the quality characteristic is one or more selected from the group of sugar chain structure, host-derived DNA, monomer, polymer, and degradation product, and the productivity characteristic is one or more selected from the group of antibody concentration, antibody production per glucose consumption, and antibody production per culture vessel. [Aspect 7] The manufacturing condition control method according to Aspect 1, characterized in that the important process parameter is one or more selected from the group of culture medium state parameter, temperature distribution difference, dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, fluid parameter, cell state parameter, setting parameter, and design parameter. [Aspect 8] The method for controlling manufacturing conditions according to Aspect 1, characterized in that the important process parameter is one or more selected from the group consisting of ammonia concentration, glucose concentration, lactate concentration, alanine concentration, glutamine concentration, mixing time constant, shear stress, Kolmogorov, oxygen transfer capacity coefficient, growth rate, cell shape, cell diameter distribution difference, cell viability, glutamine consumption rate, glucose consumption rate, antibody ratio production rate, ammonia secretion rate, lactate secretion rate, viable cell number density, gene expression level, pH, temperature, stirring rotation speed, dissolved oxygen concentration, cell separation method, perfusion rate, culture vessel capacity, agitator shape, culture vessel aspect ratio, liquid aeration rate, and bubble diameter. [Aspect 9] The method for controlling manufacturing conditions according to Aspect 3, characterized in that the secondary important process parameter is one or more selected from the group consisting of culture medium state parameter, temperature distribution difference, dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, fluid parameter, and cell state parameter.[Aspect 10] The method for controlling manufacturing conditions according to aspect 3, characterized in that the secondary important step parameter is one or more selected from the group consisting of ammonia concentration, glucose concentration, lactate concentration, alanine concentration, glutamine concentration, mixing time constant, shear stress, Kolmogorov scale, oxygen transfer capacity coefficient, growth rate, cell shape, cell diameter distribution difference, cell viability, glutamine consumption rate, glucose consumption rate, antibody ratio production rate, ammonia secretion rate, lactate secretion rate, viable cell number density, and gene expression level. [Aspect 11] The method for controlling manufacturing conditions according to aspect 3, characterized in that the primary important step parameter is one or more selected from the group consisting of setting parameters and design parameters. [Aspect 12] The method for controlling manufacturing conditions according to aspect 3, characterized in that the primary important step parameter is one or more selected from the group consisting of pH, temperature, stirring speed, dissolved oxygen concentration, cell separation method, perfusion rate, culture vessel capacity, stirring blade shape, culture vessel aspect ratio, liquid aeration rate, and bubble diameter. [Aspect 13] The manufacturing condition management method according to Aspect 1, characterized in that, in a manufacturing process in which multiple manufacturing conditions affect a product quality index, when determining the manufacturing conditions necessary to achieve a certain quality standard, the interrelationship between two or more of the multiple manufacturing conditions is made into a function and / or an inverse function, so that each of the two or more manufacturing conditions is uniquely determined. [Aspect 14] The manufacturing condition management method according to Aspect 13, characterized in that the relationship between the product quality index and the multiple manufacturing conditions is determined by experiment and / or theoretical analysis. [Aspect 15] The manufacturing condition management method according to Aspect 13, characterized in that when determining the manufacturing conditions, the domain and / or range of these conditions are changed so that each of the two or more manufacturing conditions is uniquely determined. [Aspect 16] The manufacturing condition management method according to Aspect 13, characterized in that when determining the manufacturing conditions, either of the correlations between the two or more manufacturing conditions is strictly monotonically increasing or strictly monotonically decreasing, and when the value of the larger rate of change of the two or more manufacturing conditions is uniquely determined, the value of the smaller rate of change is also uniquely determined. [Aspect 17] The method for controlling manufacturing conditions according to aspect 1, characterized in that the manufacturing process is carried out in a culture apparatus.[Aspect 18] The manufacturing condition management method according to Aspect 1, characterized in that when there are two or more product quality indicators for the product, an explanatory variable is selected for each product quality indicator and a different design space is created for each. [Aspect 19] The manufacturing condition management method according to Aspect 1, characterized in that a design space is created for each manufacturing apparatus of different scales or for each manufacturing apparatus of different cell number densities. [Aspect 20] The manufacturing condition management method according to Aspect 1, characterized in that in the second specific step, when the manufacturing process in which the product is manufactured includes culture in a culture tank, the important step parameters are identified based on a fluid simulation in the culture tank, or when the manufacturing process in which the product is manufactured includes culture in a culture tank, the important step parameters are identified based on experimental values ​​for the candidate important step parameters and the calculation results of a fluid simulation in the culture tank for the candidate important step parameters, or the important step parameters are identified based on the result of correcting experimental values ​​for the candidate important step parameters with the calculation results of a fluid simulation in the culture tank for the candidate important step parameters. [Aspect 21] The manufacturing condition management method according to aspect 1, wherein a display step is provided between the creation step and the determination step, and the display step displays the created design space on a display unit as a graph with a plurality of important process parameters as explanatory variables. [Aspect 22] The manufacturing condition management method according to aspect 21, wherein the display step creates the graph by using an important process parameter selected from the identified important process parameters as variables in the graph, and using important process parameters other than the selected important process parameter as fixed values. [Aspect 23] The manufacturing condition management method according to aspect 1, wherein a control step is provided after the determination step, and the control step creates a control value determined based on the design space and a control instruction based on the manufacturing conditions based on the control value, and transmits the control instruction to the equipment in the manufacturing process to be executed by the equipment in the manufacturing process.[Aspect 24] The manufacturing condition management method according to aspect 23, characterized in that the control step creates a control instruction that changes in time series based on the control value and the range information of the design space at different points in time. [Aspect 25] The manufacturing condition management method according to aspect 23, characterized in that the control step associates the information regarding the control value with the information regarding the created design space and displays it on the display unit. [Aspect 26] The manufacturing condition management method according to aspect 23, characterized in that the information of the design space displayed on the display unit in the control step is a graph in which the design space is a graph with a plurality of important process parameters as explanatory variables, and the graph is created with important process parameters selected from the identified important process parameters as variables in the graph, and important process parameters other than the selected important process parameters as fixed values. [Aspect 27] The manufacturing condition management method according to aspect 26, characterized in that the control step overlays and displays information indicating the position of the control value on the graph representing the design space. [Aspect 28] The manufacturing condition management method according to aspect 26, characterized in that the control step displays time-series changing control instructions and design space information, created based on information of the control values ​​and the range of the design space at different points in time, on the graph representing the design space.[Aspect 29] A manufacturing condition management program characterized by causing a computer to perform: a first extraction step of extracting candidate quality characteristics related to the quality in the manufacture of a target product and / or candidate productivity characteristics related to the productivity in the manufacture of a product; a first identification step of identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or productivity characteristics; a determination step of determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection step of selecting any one time within the manufacturing process to create a design space if it is in a steady state, and selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction step of extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification step of identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation step of determining the range of the identified important process parameters and creating a design space; and a determination step of determining manufacturing conditions based on the design space.[Aspect 30] A manufacturing condition management system comprising: a first extraction means for extracting candidate quality characteristics related to the quality in the manufacture of a target product and / or candidate productivity characteristics related to the productivity in the manufacture of a product; a first identification means for identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or candidate productivity characteristics; a determination means for determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection means for selecting any one time within the manufacturing process to create a design space if it is in a steady state, and for selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction means for extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification means for identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation means for determining the range of the identified important process parameters and creating a design space; and a determination means for determining manufacturing conditions based on the design space.[Aspect 31] A computer-readable recording medium recording a program for causing a computer to perform the following steps: a first extraction step of extracting candidate quality characteristics related to the quality in the manufacture of a target product and / or candidate productivity characteristics related to the productivity in the manufacture of a product; a first identification step of identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or productivity characteristics; a determination step of determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection step of selecting any one time within the manufacturing process to create a design space if it is in a steady state, and selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction step of extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification step of identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation step of determining the range of the identified important process parameters and creating a design space; and a determination step of determining manufacturing conditions based on the design space.

[0174] The manufacturing condition management method, manufacturing condition management program 120, and manufacturing condition management system 1300 according to the present invention have been described in detail above with reference to embodiments and examples. However, the present invention is not limited to the embodiments and examples described above, and various modifications are included. For example, the embodiments described above are described in detail in order to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add a configuration from another embodiment to the configuration of one embodiment. In addition, it is possible to add, delete, or replace a part of the configuration of each embodiment with other configurations.

[0175] 70 Product characteristics target 71 Important quality characteristics (quality characteristics) 72 Important productivity characteristics (productivity characteristics) 73, 73a-73e Important process parameters 81 Primary important process parameters 82 Secondary important process parameters 91 Primary design space 92 Secondary design space 120 Manufacturing condition management program 121 Recording medium 1300 Manufacturing condition management system 1310 Computer 1311 Setting means 1312 First extraction means 1313 First identification means 1314 Discrimination means 1315 Selection means 1316 Second extraction means 1317 Second identification means 1318 Creation means 1319 Decision means S11, S21 Setting process S12, S22 First extraction process S13, S23 First identification process S14, S24 Discrimination process S15, S25 Selection process S16, S26 Second extraction step S17, S27 Second identification step S18, S28 Production step S19, S29 Determination step

Claims

1. A method for managing manufacturing conditions, comprising: a first extraction step of extracting candidate quality characteristics related to the quality in the manufacturing of a target product and / or candidate productivity characteristics related to the productivity in the manufacturing of a product; a first identification step of identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or candidate productivity characteristics; a determination step of determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection step of selecting any one time within the manufacturing process to create a design space if it is in a steady state, and selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction step of extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification step of identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation step of determining the range of the identified important process parameters and creating a design space; and a determination step of determining manufacturing conditions based on the design space.

2. A method for managing manufacturing conditions according to claim 1, characterized in that, in the first extraction step, candidates for quality characteristics and candidates for productivity characteristics are extracted; in the first identification step, the quality characteristics and productivity characteristics are identified based on evidence for the candidates for quality characteristics and candidates for productivity characteristics; and in the second extraction step, candidates for important process parameters that cause the quality characteristics and productivity characteristics are extracted.

3. The method for controlling manufacturing conditions according to claim 1, characterized in that the critical process parameters are classified into primary critical process parameters that have a distant correlation with the quality characteristics and / or productivity characteristics, and secondary critical process parameters that have a close correlation with the quality characteristics and / or productivity characteristics.

4. The method for managing manufacturing conditions according to claim 3, characterized in that the design space is created using the secondary critical process parameters as the objective variables and the primary critical process parameters as the explanatory variables.

5. The method for controlling manufacturing conditions according to claim 1, characterized in that the product characteristic targets, which are the quality and productivity that the product should satisfy in the manufacturing process, are one or more selected from the group consisting of product function, product yield, and yield.

6. The method for controlling manufacturing conditions according to claim 1, characterized in that the quality characteristic is one or more selected from the group consisting of sugar chain structure, host-derived DNA, monomer, polymer, and degradation product, and the productivity characteristic is one or more selected from the group consisting of antibody concentration, antibody production per glucose consumption, and antibody production per culture vessel.

7. The method for controlling manufacturing conditions according to claim 1, characterized in that the important process parameter is one or more selected from the group consisting of culture medium state parameters, temperature distribution difference, dissolved carbon dioxide distribution difference, dissolved oxygen concentration distribution difference, fluid parameters, cell state parameters, setting parameters, and design parameters.

8. The method for controlling manufacturing conditions according to claim 3, characterized in that the secondary critical process parameter is one or more selected from the group consisting of culture medium state parameters, temperature distribution differences, dissolved carbon dioxide distribution differences, dissolved oxygen concentration distribution differences, fluid parameters, and cell state parameters.

9. The method for controlling manufacturing conditions according to claim 3, characterized in that the primary critical process parameter is one or more selected from the group of setting parameters and design parameters.

10. The manufacturing condition management method according to claim 1, characterized in that, in a manufacturing process in which multiple manufacturing conditions affect product quality indicators, when determining the manufacturing conditions necessary to achieve a certain quality standard, the interrelationship between two or more of the multiple manufacturing conditions is made into a function and / or inverse function, so that each of the two or more manufacturing conditions is uniquely determined.

11. The method for controlling manufacturing conditions according to claim 10, characterized in that the relationship between the product quality index and the plurality of manufacturing conditions is determined by experiment and / or theoretical analysis.

12. The method for managing manufacturing conditions according to claim 10, characterized in that when determining the manufacturing conditions, one of the correlations between the two or more manufacturing conditions is either strictly monotonically increasing or strictly monotonically decreasing, and when the larger of the two or more manufacturing conditions' rates of change is determined to a single value, the smaller of the two or more manufacturing conditions' rates of change is also determined to a single value.

13. The method for managing manufacturing conditions according to claim 1, characterized in that, if there are two or more product quality indicators for the product, an explanatory variable is selected for each product quality indicator and a different design space is created for each.

14. The method for controlling manufacturing conditions according to claim 1, characterized in that the design space is created for each manufacturing apparatus of different scales, or for each manufacturing apparatus of different cell number densities.

15. A manufacturing condition management program characterized by causing a computer to execute: a first extraction step of extracting candidate quality characteristics related to the quality in the manufacture of a target product and / or candidate productivity characteristics related to the productivity in the manufacture of the product; a first identification step of identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or productivity characteristics; a determination step of determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection step of selecting any one time within the manufacturing process to create a design space if it is in a steady state, and selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction step of extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification step of identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation step of determining the range of the identified important process parameters and creating a design space; and a determination step of determining manufacturing conditions based on the design space.

16. A manufacturing condition management system comprising: a first extraction means for extracting candidate quality characteristics related to the quality in the manufacture of a target product and / or candidate productivity characteristics related to the productivity in the manufacture of a product; a first identification means for identifying the target quality characteristics and / or productivity characteristics based on evidence for the extracted candidate quality characteristics and / or candidate productivity characteristics; a determination means for determining whether the manufacturing process in which the product is manufactured is in a steady state or a transient state; a selection means for selecting any one time within the manufacturing process to create a design space if it is in a steady state, and for selecting one or more times within the manufacturing process to create a design space if it is in a transient state; a second extraction means for extracting candidate important process parameters that cause the identified quality characteristics and / or productivity characteristics; a second identification means for identifying important process parameters from the extracted candidate important process parameters based on evidence; a creation means for determining the range of the identified important process parameters and creating a design space; and a determination means for determining manufacturing conditions based on the design space.

Images