Calibration curve model creating device and calibration curve model creating method

Abstract

To facilitate the creation of a calibration curve model that estimates living cell density or the like.SOLUTION: A calibration curve model creation device includes: a control unit 14 that creates a calibration curve model 152 for estimating living cell density in a culture medium based on impedance measurements taken by a sensor 11, and an operation unit 12 that accepts creation conditions for creating a parameter set, including parameters to create the calibration curve model 152. The control unit 14 creates parameter sets in accordance with the creation conditions; for each of the created parameter sets, creates a plurality of calibration curve models 152 using data sets for model creation; and stores one selected from among the plurality of calibration curve models 152.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a calibration curve model creating device and a calibration curve model creating method for creating a calibration curve model for estimating viable cell density (VCD). [Background technology]

[0002] BACKGROUND ART Conventionally, there is a prediction technique for predicting the viable cell density in cell culture including, for example, animal cells, microorganisms, plant cells, etc., using a cell culture device (bioreactor) (see, for example, Patent Document 1).

[0003] Patent Document 1 discloses a cell testing device that includes an impedance sensor that measures the impedance of a culture medium, a memory unit that classifies a predetermined period within the culture period from the start of cell culture to cell death into multiple periods and stores a coefficient for estimating the number of viable cells surviving in the culture medium during the predetermined period using the impedance for each of the multiple classified periods, and a viable cell number estimation unit that acquires the impedance and estimates the number of viable cells using at least one of the coefficients for each period stored in the memory unit against the impedance. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2019-124594 Summary of the Invention [Problem to be solved by the invention]

[0005] Here, a conventional method for creating a calibration curve model for estimating viable cell density will be described. Fig. 10 is a flowchart showing the conventional method for creating a calibration curve model. In this specification, a calibration curve model is a set of coefficients for estimating the viable cell density in a culture solution from the impedance of the culture solution measured by a bioreactor or the like or the capacitance calculated from the impedance.

[0006] As shown in FIG. 10, in step S1001, a control device such as a computer reads a data set for creating a calibration curve model (hereinafter referred to as a calibration curve model creation data set).

[0007] Next, in step S1002, the user specifies parameters for creating a calibration curve model. For example, the user specifies the boundary between the first phase (first half of the culture) and the second phase (second half of the culture) of the cell culture, i.e., the elapsed time (seconds) from the start of the culture, as a parameter.

[0008] Next, in step S1003, the control device creates a calibration curve model using the read calibration curve model creation data set and parameters designated by the user.

[0009] Next, the created calibration curve model is validated. In step S1004, the control device loads a data set for validation designated by the user (hereinafter referred to as a validation data set) into the created calibration curve model.

[0010] Then, in step S1005, the control device verifies (validates) the created calibration curve model by comparing the viable cell density estimated by the calibration curve model using the validation dataset with the known viable cell density of the validation dataset.

[0011] Finally, in step S1006, the user checks the validation results, and if there are no problems, ends this flow, but if there are problems, starts the process over again from step S1002.

[0012] As described above, in the conventional flow for creating a calibration curve model, the user is required to specify parameters, specify a data set for validation, and check the validation results, and if there is a problem, the user is required to specify the parameters again, specify a data set for validation, and check the validation results again. Therefore, in order to create an optimal calibration curve model, the user is required to repeatedly perform tasks such as specifying parameters, which is time-consuming for the user, which has been an issue.

[0013] Therefore, the present invention proposes a calibration curve model creation device and a calibration curve model creation method that can easily create a calibration curve model for estimating the viable cell density and the like. [Means for solving the problem]

[0014] The calibration curve model creation device of the present invention includes a control unit that creates a calibration curve model that estimates the viable cell density in a culture solution from the impedance measured by a sensor that measures the impedance of the culture solution, and a reception unit that receives generation conditions for generating one or more parameter sets including one or more parameters for creating the calibration curve model. The control unit generates one or more parameter sets according to the generation conditions received by the reception unit, and for each of the generated parameter sets, creates multiple calibration curve models using one or more parameters included in the parameter set, and a model creation dataset that includes the elapsed time since the start of culture of the culture solution for model creation, the impedance of the culture solution for model creation or a capacitance calculated from the impedance, and the viable cell density in the culture solution for model creation measured by a cell counter that measures the viable cell density in the culture solution, and stores one selected from the multiple calibration curve models.

[0015] Furthermore, the calibration curve model creation method of the present invention includes receiving creation conditions for generating one or more parameter sets including one or more parameters for creating a calibration curve model that estimates the viable cell density in a culture medium from the impedance measured by a sensor that measures the impedance of the culture medium; generating one or more parameter sets in accordance with the received creation conditions; creating, for each of the generated one or more parameter sets, a plurality of calibration curve models using one or more parameters included in the parameter set, and a model creation dataset including the elapsed time from the start of culture of the culture medium for model creation, the impedance of the culture medium for model creation or a capacitance calculated from the impedance, and the viable cell density in the culture medium for model creation measured by a cell counter that measures the viable cell density in the culture medium; and storing one selected from the plurality of calibration curve models. [Effects of the Invention]

[0016] According to the present invention, it is possible to easily create a calibration curve model for estimating the viable cell density and the like. [Brief explanation of the drawings]

[0017] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a calibration curve model creation system according to an embodiment. [Figure 2] FIG. 2 is a hardware block diagram of a control unit according to the embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of a model creation dataset according to an embodiment. [Figure 4] FIG. 10 is a diagram showing a method for calculating an average error that serves as an index for verifying a calibration curve model according to an embodiment. [Figure 5] 10 is a flowchart showing a method for creating a calibration curve model according to an embodiment. [Figure 6] FIG. 10 is a diagram showing a list of parameter sets according to the embodiment. [Figure 7] 10 is a flowchart showing a method for verifying a calibration curve model according to an embodiment. [Figure 8] FIG. 10 is a diagram (table) showing the verification results of the calibration curve model according to the embodiment. [Figure 9] FIG. 10 is a diagram (graph) showing the verification results of a calibration curve model according to the embodiment. [Figure 10] 1 is a flowchart showing a conventional method for creating a calibration curve model. DETAILED DESCRIPTION OF THE INVENTION

[0018] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0019] (Calibration curve model creation device 100) 1 is a diagram showing an example of the configuration of a calibration curve model creating device 100 according to this embodiment. As shown in FIG. 1, the calibration curve model creating device 100 includes a bioreactor 1 and a cell counter 2.

[0020] (Bioreactor 1) The bioreactor 1 includes a culture tank 10, a sensor 11, an operation unit 12, a display unit 13, a control unit 14, and a storage device 15. The sensor 11 has a probe 111 and a sensor unit 112. The storage device 15 stores a calibration curve model creation tool 151, a calibration curve model 152, a model creation dataset 153, and one or more validation datasets 154. The culture tank 10, the sensor 11, the operation unit 12, the display unit 13, and the storage device 15 are communicably connected to the control unit 14. The cell counter 2 is also communicably connected to the control unit 14. In this embodiment, the control unit 14 and the cell counter 2 are communicatively connected, but the measurement results of the cell counter 2 may be input to the control unit 14 using a portable memory such as a USB (Universal Serial Bus) memory. Communication between the control unit 14 and its peripheral devices (culture vessel 10, sensor 11, operation unit 12, display unit 13, storage device 15, or cell counter 2) may be wired or wireless.

[0021] (Culture tank 10) A culture solution (culture medium) containing cells is placed in the culture tank 10. The culture tank 10 is equipped with a sensor 11, an agitator (not shown), a heater, etc. Agitation by the agitator and heating by the heater are controlled by a control unit 14.

[0022] (Sensor 11) The sensor 11 measures the impedance of the culture solution in real time by an electrical impedance measurement method.

[0023] The probe 111 includes a measurement probe and an impedance sensor. The impedance sensor is a sensor that measures the impedance of a culture solution containing cells. The impedance sensor measures the impedance at each of a plurality of different frequencies (for example, 287 kHz, 473 kHz, ..., 20000 kHz). The probe 111 measures the impedance of the culture solution and outputs the measured value to the sensor unit 112. The probe 111 may also have an amplifier unit that amplifies the measured value.

[0024] The sensor unit 112 separates the measurement value output by the probe 111 into capacitance (dielectric constant) and dielectric conductance (electrical conductivity) by a well-known method, and outputs the separated capacitance and dielectric conductance to the control unit 14. The sensor unit 112 may also have an amplifier that amplifies the measurement value measured by the probe 111.

[0025] (Operation unit 12) The operation unit 12 is, for example, a touch panel sensor and operation buttons provided on the display unit 13. The operation unit 12 receives operation contents from the user and outputs the received operation contents to the control unit 14. The operation contents include the cell name, set values ​​of the culture conditions, a culture start instruction, a culture end instruction, generation conditions described below, etc. The operation unit 12 (reception unit) receives generation conditions for generating one or more parameter sets including one or more parameters for creating the calibration curve model 152.

[0026] (parameter, parameter set) Here, parameters and parameter sets will be described. Parameters are values ​​used to create the calibration curve model 152, and examples of parameters include parameter A indicating the boundary (the elapsed time from the start of culture) between the first phase (the first half of culture) and the second phase (the second half of culture) of cell culture, and parameter B indicating the order for estimating the viable cell density in the second phase. A parameter set is a collection of parameters including one or more parameters. Note that a parameter set may consist of only one parameter, or may be a collection of multiple parameters.

[0027] (Display section 13) The display unit 13 is, for example, a liquid crystal display device, an organic EL (Electro Luminescence) display device, etc. The display unit 13 displays a display image output by the control unit 14. The display image is an image such as a culture condition setting screen, a condition specification screen for specifying the generation conditions described below, and a verification result screen (table in FIG. 8, graph in FIG. 9) of the multiple calibration curve models that have been created.

[0028] (Control unit 14) The control unit 14 is a computer system that controls the operation of peripheral devices that are communicatively connected to the control unit 14. For example, in response to a culture start instruction, the control unit 14 starts culture by controlling the agitator, heater, etc. of the culture tank 10 according to the set values ​​of the culture conditions. The control unit 14 starts timing at the start of culture and measures the elapsed time (seconds) from the start of culture.

[0029] The control unit 14 executes the calibration curve model creation tool 151 stored in the storage device 15, and creates a plurality of calibration curve models 152 using a model creation dataset 153. Specifically, the control unit 14 generates one or more parameter sets in accordance with the creation conditions accepted by the operation unit 12. Then, for each of the one or more generated parameter sets, the control unit 14 creates a plurality of calibration curve models 152 using one or more parameters included in the parameter set and the model creation dataset 153.

[0030] The control unit 14 also verifies the multiple calibration curve models 152. Specifically, the control unit 14 causes each of the multiple created calibration curve models 152 to read the elapsed time and capacitance of the validation dataset and estimate the viable cell density in the culture solution for validation. The control unit 14 then verifies the calibration curve model 152 from the error between the viable cell density estimated by the calibration curve model 152 and the viable cell density of the validation dataset.

[0031] Furthermore, when there are multiple validation data sets, the control unit 14 causes each of the multiple created calibration curve models 152 to read the elapsed time and capacitance of the multiple validation data sets and estimate the viable cell density in the validation culture solution for each of the multiple validation data sets. Then, the control unit 14 verifies the calibration curve model 152 from the average error between the viable cell density for each of the multiple validation data sets estimated by the calibration curve model 152 and the viable cell density of the multiple validation data sets.

[0032] Then, the control unit 14 stores in the storage device 15 the calibration curve model 152 selected by the user from among the plurality of calibration curve models 152 or the calibration curve model 152 selected automatically.

[0033] Then, the control unit 14 reads the capacitance of the culture medium received from the sensor unit 112 into the calibration curve model 152 stored in the storage device 15, and estimates the density of living cells in the culture medium.

[0034] (Hardware configuration of control unit 14) FIG. 2 is a hardware block diagram of a control unit according to an embodiment. The hardware configuration of the control unit 14 will now be described. The control unit 14 is a computer including a processor 141, a memory 142, an I / F (interface) 143, an auxiliary storage unit 144, and the like. The processor 141 is a central processing unit (CPU), a digital signal processor (DSP), or the like, and loads a program (e.g., a calibration curve model creation tool 151) stored in the storage device 15 into the memory 142 and executes it. The processor 141 also has a clocking means such as a real-time clock (RTC), and measures the elapsed time (seconds) from the start of culture, and the like. The memory 142 is a random access memory (RAM), and the like. The processor 141 communicates with peripheral devices and external devices via the I / F 143. The auxiliary storage unit 144 is a read-only memory (ROM), and the like, and stores the boot program of the control unit 14, and the like.

[0035] (Storage device 15) The storage device 15 stores various programs, data used in the programs, etc. The storage device 15 is an HDD (Hard Disc Drive), an SSD (Solid State Drive), etc. The storage device 15 of this embodiment stores a calibration curve model creation tool 151 for creating a calibration curve model 152, the calibration curve model 152 created by the calibration curve model creation tool 151, a model creation dataset 153 used to create the calibration curve model 152, and one or more validation datasets 154 used to validate the calibration curve model 152.

[0036] (Calibration curve model creation tool 151) The calibration curve model creation tool 151 is a program for creating the calibration curve model 152. The calibration curve model creation tool 151 is executed by the processor 141 of the control unit 14, and performs the following processes. A process of displaying a condition specification screen on the display unit 13 for specifying generation conditions for generating a parameter set - Automatically generate multiple parameter sets according to the generation conditions received via the condition specification screen A process of creating a calibration curve model 152 for each of the generated parameter sets. · Verification process for multiple calibration curve models 152 created ·Process to display the verification results A process of storing the calibration curve model 152 selected by the user or automatically selected in the storage device 15

[0037] (Calibration curve model 152) The calibration curve model 152 is a set of coefficients for estimating the viable cell density from the capacitance of the culture medium received from the sensor unit 112. The following are examples of the coefficients. Coefficient for estimating viable cell density in the first phase of cell culture (first half of culture) Coefficient for estimating viable cell density in the second phase of cell culture (later stage of culture) Coefficient for estimating whether the cell culture is in the first or second phase

[0038] (Model creation dataset 153) As shown in FIG. 3 , the model creation dataset 153 includes the elapsed time (seconds) from the start of culturing the culture solution for model creation, the capacitance of the culture solution for model creation, and the viable cell density in the culture solution for model creation measured by the cell counter 2. Since the impedance sensor of the probe 111 measures impedance at a plurality of different frequencies, the example of FIG. 3 includes the capacitance received from the sensor unit 112 for each frequency (287 kHz, 473 kHz, ... 2000 kHz). As described above, the model creation dataset 153 is a dataset obtained by measuring the culture solution for model creation. Furthermore, the model creation dataset 153 may be measured in the bioreactor 1 or may be measured in another bioreactor 1.

[0039] (Validation dataset 154) The validation dataset 154 is used to validate the multiple calibration curve models 152 created by the calibration curve model creation tool 151. Like the model creation dataset 153, the validation dataset 154 includes the elapsed time (seconds) from the start of culture of the validation culture solution, the capacitance of the validation culture solution, and the viable cell density in the validation culture solution measured by the cell counter 2. The validation dataset 154 may be a past dataset measured in the bioreactor 1 or may be a dataset measured in another bioreactor. It may also include a viable cell density measured by a cell counter different from the cell counter 2. Note that the measurement conditions for the measurement values ​​of the validation dataset 154 and the model creation dataset 153 are preferably the same.

[0040] 4, a plurality of validation datasets 154 (validation datasets 1 to 3 in the figure) are prepared. The average of the errors between the viable cell density estimated by a calibration curve model 152 (model 1 in the figure) created by a calibration curve model creation tool 151 and each of the plurality of validation datasets 154 is taken as the average error, and this average error is used as an index for validating the calibration curve model 152.

[0041] (Cell Counter 2) The cell counter 2 samples the culture medium, stains the viable cells, and measures the viable cell density (VCD) [cell / mL] in the culture medium. The timing at which the sensor 11 measures the capacitance and the timing at which the viable cell density is obtained may or may not coincide. The measurement timing may be, for example, such that the capacitance is measured continuously and the viable cell density is obtained intermittently. The cell counter 2 outputs information indicating the measured viable cell density to the bioreactor 1. The measurement by the cell counter 2 is performed offline.

[0042] (Calibration curve model creation flow) Next, a method for creating the calibration curve model 152 (calibration curve model creating method) will be described with reference to Fig. 5. Fig. 5 is a flowchart showing the method for creating the calibration curve model. The processor 141 of the control unit 14 executes the calibration curve model creating tool 151 in the storage device 15, thereby executing each step of the flowchart in Fig. 5.

[0043] In step S500, when the calibration curve model creation tool 151 is executed, the control unit 14 displays on the display unit 13 a condition designation screen for designating the creation conditions for creating a parameter set.

[0044] In step S501, the user specifies the generation conditions via the condition specification screen. Specifically, the user specifies, for example, as the generation condition for parameter A, the range of the boundary between the first phase (first half of culture) and the second phase (second half of culture) of cell culture (for example, 500,000 to 700,000 (seconds) from the start of culture), and the number of boundaries to be picked from that range (for example, 5). In addition, the user specifies, for example, as the generation condition for parameter B, the range of the order (for example, 3 to 5) for estimating the viable cell density in the second phase.

[0045] In step S502, the control unit 14 automatically generates a parameter set according to the generation conditions specified by the user. When the above-mentioned generation conditions for parameter A (range of boundary between the first phase (first half of culture) and the second phase (second half of culture) of cell culture: 500,000 to 700,000 (seconds), number of boundaries to be picked: 5) are input, the control unit 14 picks 500,000, 550,000, 600,000, 650,000, and 700,000 (seconds) as candidates for parameter A. When the above-mentioned generation conditions for parameter B (candidate range of order for estimating the viable cell density in the second phase: 3 to 5) are input, the control unit 14 picks 3, 4, and 5 as candidates for parameter B.

[0046] Fig. 6 is a diagram showing a list of parameter sets according to an embodiment. The list of parameter sets generated according to generation conditions specified by the user includes 15 patterns of parameter sets (parameter sets 1 to 15 in the figure) that combine five candidates for parameter A (500000, 550000, 600000, 650000, 700000 (seconds)) with three candidates for parameter B (3, 4, 5), as shown in Fig. 6.

[0047] In step S503, the control unit 14 repeats the following steps S504 and S505 the number of times equal to the number of generated parameter sets (15 patterns).

[0048] In step S504, the control unit 14 selects one of the multiple automatically generated parameter sets, and creates the calibration curve model 152 using the parameters A and B of the selected parameter set and the model creation data set.

[0049] In step S505, the control unit 14 verifies the created calibration curve model 152. Details of the verification of the calibration curve model 152 will be described later.

[0050] After creating the calibration curve models 152 (step S504) and verifying the calibration curve models 152 (step S505) for the number of parameter sets, in step S506, the control unit 14 displays the verification results (see FIGS. 8 and 9) of the plurality of calibration curve models 152 on the display unit 13. The details of the verification results of the plurality of calibration curve models 152 will be described later.

[0051] In step S507, the control unit 14 stores the calibration curve model 152 selected by the user in a file in the storage device 15.

[0052] (Calibration curve model verification flow) Next, a method for verifying the calibration curve model 152 will be described with reference to Fig. 7. Fig. 7 is a flowchart showing a method for verifying the calibration curve model according to the embodiment.

[0053] In step S701, the control unit 14 reads one of the calibration curve models 152 created in step S504.

[0054] In step S702, the control unit 14 initializes the total error to 0 in advance in order to calculate the average of the errors between the estimated values ​​of the read calibration curve model 152 and the validation data set.

[0055] In step S703, the control unit 14 repeats the following processing of steps S704 to S707 the number of times equal to the number of selected validation data sets.

[0056] In step S704, the control unit 14 reads one validation data set from among the multiple validation data sets.

[0057] In step S705, the control unit 14 uses the read calibration curve model 152 to calculate the viable cell density from the capacitance of the validation data set.

[0058] In step S706, the control unit 14 calculates the error between the calculated viable cell density and the viable cell density measured by the cell counter in the validation data set.

[0059] In step S707, the control unit 14 adds the calculated error to the total error.

[0060] After repeating the processes of steps S704 to S707 for the number of selected validation data sets, in step S708, the control unit 14 divides the sum of the errors by the number of validation data sets to calculate the average error.

[0061] (Verification results of the calibration curve model) Fig. 8 is a diagram (table) showing the verification results of the calibration curve model according to the embodiment, and Fig. 9 is a diagram (graph) showing the verification results of the calibration curve model according to the embodiment. The table in Fig. 8 and the graph in Fig. 9 are displayed on display unit 13.

[0062] 8, the average error calculated in step S708 is displayed in association with the parameter set of the calibration curve model 152 (information indicating the calibration curve model 152). The table in FIG. 8 is an example in which a parameter set is generated that includes a parameter A that indicates the boundary between the first phase (first half of the culture) and the second phase (second half of the culture) of the cell culture, and does not include any parameters other than the parameter A (for example, parameter B). In the example in FIG. 8, 500,000 to 620,000 seconds is specified as the range of the boundary between the first phase (first half of the culture) and the second phase (second half of the culture) of the cell culture, and 13 is specified as the number of boundaries to be picked up.

[0063] As shown in FIG. 8, the parameter set and average error of the calibration curve model 152 with the smallest average error are highlighted and displayed.

[0064] 8 can be graphed to graphically display the relationship between the parameter set and the average error of the calibration curve model 152. In the graph of Fig. 9, the horizontal axis represents the parameter set of the calibration curve model, and the vertical axis represents the average error calculated in step S708.

[0065] (Effects of this embodiment) In this embodiment, the user simply inputs the generation conditions for generating a parameter set, and the control unit 14 automatically generates multiple parameter sets and automatically creates multiple calibration curve models for the generated multiple parameter sets. In this way, in this embodiment, an appropriate calibration curve model can be saved in a short time and with minimal user operations, and the viable cell density can be estimated using this calibration curve model.

[0066] Furthermore, in this embodiment, the created calibration curve model can be verified using multiple validation data sets, thereby improving the accuracy of estimating the viable cell density using the calibration curve model.

[0067] Although the present invention has been described above with reference to the embodiments, the above embodiments are merely illustrative of specific examples of how the present invention can be implemented, and the technical scope of the present invention should not be construed as being limited by these embodiments. In other words, the present invention can be implemented in various forms without departing from its technical concept or main features.

[0068] For example, in the above-described embodiment, the weighting of the error in the first phase (first half of culture) of cell culture and the error in the second phase (second half of culture) was not described, but it may be possible to specify the weighting of the error in the first phase (first half of culture) of cell culture and the error in the second phase (second half of culture). For example, the error in the first phase (first half of culture) of cell culture and the error in the second phase (second half of culture) may be weighted one-to-one, or in order to emphasize the error in a specific phase, the weight of the error in a specific phase may be made greater than the weight of the error in other phases.

[0069] Furthermore, it may be possible to display only the error in a specific phase (for example, the error in the second phase (later stage of culture)), or it may be possible to display both the error in all phases and the error in a specific phase.

[0070] In the above embodiment, an example was described in which the number to be picked from the specified range was specified, but instead of the number to be picked, a picking interval (for example, 50,000 seconds) may be specified. Also, instead of the specified range, a data index range (from the 500th data to the 509th data) may be specified.

[0071] In the above-described embodiment, the calibration curve model is verified by calculating the error or average error between the estimated value by the calibration curve model and the validation dataset. However, an index such as a correlation coefficient may be added instead of or in addition to the error.

[0072] Also, there may be three or more aspects of cell culture.

[0073] In the above-described embodiment, the control unit 14 installed in the bioreactor 1 creates the calibration curve model 152 and verifies the calibration curve model 152, but the control unit 14 may also be an information-processing computer such as a cloud server, an on-premise server, a personal computer, a smartphone, or a tablet. [Explanation of symbols]

[0074] 1: Bioreactor, 2: Cell counter, 10: Culture tank, 11: Sensor, 12: Operation unit, 13: Display unit, 14: Control unit, 15: Storage device, 100: Calibration curve model creation device, 111: Probe, 112: Sensor unit, 141: Processor, 142: Memory, 143: I / F (interface), 144: Auxiliary storage unit, 151: Calibration curve model creation tool, 152: Calibration curve model, 153: Data set for model creation, 154: Data set for validation

Claims

1. a control unit that creates a calibration curve model for estimating the viable cell density in the culture solution from the impedance measured by a sensor that measures the impedance of the culture solution; a reception unit that receives a generation condition for generating one or more parameter sets including one or more parameters for creating the calibration curve model, The control unit generating one or more parameter sets in accordance with the generation conditions received by the reception unit; For each of the one or more parameter sets generated, a plurality of calibration curve models are created using one or more parameters included in the parameter set, and a model creation data set including the elapsed time from the start of culture of the culture solution for model creation, the impedance of the culture solution for model creation or a capacitance calculated from the impedance, and the viable cell density in the culture solution for model creation measured by a cell counter that measures the viable cell density in the culture solution; One selected from the plurality of calibration curve models is stored. A calibration curve model creating device characterized by:

2. The control unit a validation data set including the elapsed time from the start of culture of the validation culture solution, the impedance of the validation culture solution or the capacitance calculated from the impedance, and the viable cell density in the validation culture solution measured by the cell counter, is read into each of the created plurality of calibration curve models, thereby estimating the viable cell density in the validation culture solution; The calibration curve model is validated based on an error between the viable cell density estimated by the calibration curve model and the viable cell density of the validation dataset.

2. The calibration curve model creating device according to claim 1.

3. The control unit reading the elapsed time and the impedance or the capacitance of the plurality of validation data sets into each of the plurality of created calibration curve models, and estimating a viable cell density in the validation culture solution for each of the plurality of validation data sets; The calibration curve model is verified based on an average error between the viable cell density for each of the plurality of validation data sets estimated by the calibration curve model and the viable cell density of the plurality of validation data sets.

3. The calibration curve model creating device according to claim 2.

4. The control unit The information indicating the plurality of calibration curve models and the average error are displayed on a display unit in association with each other, and the information indicating the calibration curve model with the smallest average error is displayed in an emphasized manner.

4. The calibration curve model creating device according to claim 3.

5. The production conditions include a parameter range indicating a boundary between a first phase, which is the first half of the cell culture, and a second phase, which is the second half of the cell culture, and the number of boundaries selected from the range.

2. The calibration curve model creating device according to claim 1.

6. The generation conditions include a range of orders for estimating the viable cell density of the second phase.

6. The calibration curve model creating device according to claim 5.

7. receiving generation conditions for generating one or more parameter sets including one or more parameters for creating a calibration curve model for estimating the viable cell density in the culture solution from the impedance measured by a sensor that measures the impedance of the culture solution; generating one or more parameter sets in accordance with the received generation conditions; For each of the one or more parameter sets generated, creating a plurality of calibration curve models using one or more parameters included in the parameter set, and a model creation data set including the elapsed time from the start of culture of the culture solution for model creation, the impedance of the culture solution for model creation or a capacitance calculated from the impedance, and the viable cell density in the culture solution for model creation measured by a cell counter that measures the viable cell density in the culture solution; and and storing a selected one of the plurality of calibration curve models. A method for creating a calibration curve model.

8. a validation data set including the elapsed time from the start of culture of the validation culture solution, the impedance of the validation culture solution or the capacitance calculated from the impedance, and the viable cell density in the validation culture solution measured by the cell counter, is read into each of the created plurality of calibration curve models, thereby estimating the viable cell density in the validation culture solution; and and validating the calibration curve model based on an error between the viable cell density estimated by the calibration curve model and the viable cell density of the validation dataset.

8. The method for creating a calibration curve model according to claim 7.

9. Estimating the viable cell density in the culture medium for validation includes: and reading the elapsed time and the impedance or the capacitance of the plurality of validation data sets into each of the plurality of created calibration curve models, and estimating the viable cell density in the validation culture solution for each of the plurality of validation data sets; Validating the calibration curve model includes: and validating the calibration curve model from an average error between the viable cell density for each of the plurality of validation data sets estimated by the calibration curve model and the viable cell density of the plurality of validation data sets.

9. The method for creating a calibration curve model according to claim 8.

10. The method further comprises displaying information indicating the plurality of calibration curve models and the average error in association with each other on a display unit, and highlighting information indicating the calibration curve model with the smallest average error. The calibration curve model creating method according to claim 9 .

11. The production conditions include a parameter range indicating a boundary between a first phase, which is the first half of the cell culture, and a second phase, which is the second half of the cell culture, and the number of boundaries selected from the range.

8. The method for creating a calibration curve model according to claim 7.

12. The generation conditions include a range of orders for estimating the viable cell density of the second phase. The method for creating a calibration curve model according to claim 11.

Images