WO2022168774A1 - 推定装置、学習装置、最適化装置、推定方法、学習方法、および最適化方法 - Google Patents
推定装置、学習装置、最適化装置、推定方法、学習方法、および最適化方法 Download PDFInfo
- Publication number
- WO2022168774A1 WO2022168774A1 PCT/JP2022/003522 JP2022003522W WO2022168774A1 WO 2022168774 A1 WO2022168774 A1 WO 2022168774A1 JP 2022003522 W JP2022003522 W JP 2022003522W WO 2022168774 A1 WO2022168774 A1 WO 2022168774A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- cells
- data
- quality
- prediction
- learning
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims description 66
- 238000005457 optimization Methods 0.000 title claims description 59
- 238000005259 measurement Methods 0.000 claims abstract description 187
- 229940088679 drug related substance Drugs 0.000 claims abstract description 145
- 239000008186 active pharmaceutical agent Substances 0.000 claims abstract description 138
- 239000000126 substance Substances 0.000 claims abstract description 73
- 238000012258 culturing Methods 0.000 claims abstract description 59
- 229960000074 biopharmaceutical Drugs 0.000 claims abstract description 39
- 238000010899 nucleation Methods 0.000 claims abstract description 22
- 230000008569 process Effects 0.000 claims description 35
- 239000002207 metabolite Substances 0.000 claims description 18
- 235000015097 nutrients Nutrition 0.000 claims description 13
- 230000005526 G1 to G0 transition Effects 0.000 claims description 12
- 230000003698 anagen phase Effects 0.000 claims description 11
- 230000034994 death Effects 0.000 claims description 7
- 230000004060 metabolic process Effects 0.000 claims description 5
- 238000012545 processing Methods 0.000 abstract description 17
- 239000001963 growth medium Substances 0.000 abstract description 9
- 239000002609 medium Substances 0.000 description 59
- 238000003860 storage Methods 0.000 description 34
- 238000004458 analytical method Methods 0.000 description 25
- 238000010586 diagram Methods 0.000 description 24
- 238000004113 cell culture Methods 0.000 description 20
- 238000012360 testing method Methods 0.000 description 19
- 238000000746 purification Methods 0.000 description 18
- 230000015654 memory Effects 0.000 description 17
- 238000012986 modification Methods 0.000 description 15
- 230000004048 modification Effects 0.000 description 15
- 230000008859 change Effects 0.000 description 14
- 238000004891 communication Methods 0.000 description 13
- 239000003153 chemical reaction reagent Substances 0.000 description 11
- 239000000306 component Substances 0.000 description 9
- 230000006870 function Effects 0.000 description 7
- 238000004519 manufacturing process Methods 0.000 description 7
- 239000003814 drug Substances 0.000 description 5
- 238000011534 incubation Methods 0.000 description 5
- 230000036962 time dependent Effects 0.000 description 5
- 239000002028 Biomass Substances 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 239000012533 medium component Substances 0.000 description 4
- 101000786631 Homo sapiens Protein SYS1 homolog Proteins 0.000 description 3
- 102100025575 Protein SYS1 homolog Human genes 0.000 description 3
- 150000001413 amino acids Chemical class 0.000 description 3
- 230000004071 biological effect Effects 0.000 description 3
- 230000030833 cell death Effects 0.000 description 3
- 230000003833 cell viability Effects 0.000 description 3
- 238000012512 characterization method Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 230000010365 information processing Effects 0.000 description 3
- 239000004973 liquid crystal related substance Substances 0.000 description 3
- 230000000704 physical effect Effects 0.000 description 3
- 230000002062 proliferating effect Effects 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 102000004190 Enzymes Human genes 0.000 description 2
- 108090000790 Enzymes Proteins 0.000 description 2
- 230000019522 cellular metabolic process Effects 0.000 description 2
- 238000011109 contamination Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 229940079593 drug Drugs 0.000 description 2
- 230000000984 immunochemical effect Effects 0.000 description 2
- 238000009616 inductively coupled plasma Methods 0.000 description 2
- 230000003834 intracellular effect Effects 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000003277 amino acid sequence analysis Methods 0.000 description 1
- 229940125644 antibody drug Drugs 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000010261 cell growth Effects 0.000 description 1
- 238000004132 cross linking Methods 0.000 description 1
- 230000012010 growth Effects 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 238000004811 liquid chromatography Methods 0.000 description 1
- 238000004895 liquid chromatography mass spectrometry Methods 0.000 description 1
- 238000001000 micrograph Methods 0.000 description 1
- 238000012510 peptide mapping method Methods 0.000 description 1
- 230000035755 proliferation Effects 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 230000001172 regenerating effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000013076 target substance Substances 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 229960005486 vaccine Drugs 0.000 description 1
- 230000035899 viability Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/30—Prediction of properties of chemical compounds, compositions or mixtures
-
- C—CHEMISTRY; METALLURGY
- C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
- C12P—FERMENTATION OR ENZYME-USING PROCESSES TO SYNTHESISE A DESIRED CHEMICAL COMPOUND OR COMPOSITION OR TO SEPARATE OPTICAL ISOMERS FROM A RACEMIC MIXTURE
- C12P21/00—Preparation of peptides or proteins
- C12P21/02—Preparation of peptides or proteins having a known sequence of two or more amino acids, e.g. glutathione
-
- C—CHEMISTRY; METALLURGY
- C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
- C12M—APPARATUS FOR ENZYMOLOGY OR MICROBIOLOGY; APPARATUS FOR CULTURING MICROORGANISMS FOR PRODUCING BIOMASS, FOR GROWING CELLS OR FOR OBTAINING FERMENTATION OR METABOLIC PRODUCTS, i.e. BIOREACTORS OR FERMENTERS
- C12M41/00—Means for regulation, monitoring, measurement or control, e.g. flow regulation
- C12M41/48—Automatic or computerized control
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/5005—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving human or animal cells
- G01N33/5008—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving human or animal cells for testing or evaluating the effect of chemical or biological compounds, e.g. drugs, cosmetics
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/80—Data visualisation
Definitions
- the present disclosure relates to an estimation device, a learning device, an optimization device, an estimation method, a learning method, and an optimization method.
- Patent Document 1 discloses a cell culture method for culturing host cells to produce a target protein.
- Biopharmaceuticals are produced using cell metabolism that is difficult to artificially control, and the drugs used may contain ingredients derived from living cells. Therefore, it is difficult to make biopharmaceuticals of uniform quality even if they are manufactured according to a prescribed protocol. Therefore, typically, a biopharmaceutical drug substance obtained by cell culture is subjected to characterization to confirm the quality of the drug substance.
- biopharmaceuticals requires a very long time, from several weeks to several months, including the cultivation, purification, and processing processes.
- the drug substance to be characterized is obtained in a purification step subsequent to the completion of the culturing step. In other words, since the results of characterization cannot be obtained at the stage of the culturing process, it is only after the purification process is completed that it is determined whether or not the target drug substance has been obtained by culturing.
- the estimation device of the present disclosure includes a reception unit that receives input of measurement data, a prediction unit that generates quality prediction data indicating the quality of a drug substance by inputting the measurement data received by the reception unit into a prediction model, and a prediction and an output unit for outputting the quality prediction data generated by the unit.
- the measurement data includes measurement results obtained by measuring at least one substance in a culture vessel containing the cells and the medium at least one timing after a predetermined period of time has passed since the cells were seeded in the medium.
- a prediction model is a model for predicting the quality of biopharmaceutical drug substances manufactured by culturing cells.
- the learning device of the present disclosure includes a reception unit that receives learning data, and a model generation unit that generates a prediction model by executing learning processing using the learning data received by the reception unit.
- the learning data includes measurement data including measurement results obtained by measuring at least one substance in the culture vessel containing the cells and the medium at multiple timings after the cells are seeded in the medium, and the data produced from the cells. and quality data obtained by analyzing the biopharmaceutical drug substance.
- the prediction model is based on the measurement results obtained by measuring at least one substance in the culture vessel during cell culture, and the biomass produced from the cells contained in the culture vessel during cell culture. This is a model for generating quality prediction data that indicates the quality of active pharmaceutical ingredients.
- the estimation method of the present disclosure includes the steps of placing cells and a medium in a culture vessel and culturing the cells, and at least one timing after a predetermined period of time after seeding the cells in the medium. a step of measuring at least one substance; a step of inputting measurement data including measurement results obtained in the measuring step into a prediction model to generate quality prediction data indicating the quality of the drug substance; and outputting.
- a prediction model is a model for predicting the quality of biopharmaceutical drug substances manufactured by culturing cells.
- the learning method of the present disclosure includes the steps of placing cells and a medium in a culture vessel and culturing the cells, and measuring at least one substance in the culture vessel at a plurality of timings after seeding the cells in the medium. analyzing the quality of a biopharmaceutical drug substance manufactured from cells; and generating a predictive model by performing a learning process using the learning data.
- the learning data includes measurement data including measurement results obtained by the measuring step and quality data obtained by the analyzing step.
- the prediction model is based on the measurement results obtained by measuring at least one substance in the culture vessel during cell culture, and the biomass produced from the cells contained in the culture vessel during cell culture. This is a model for generating quality prediction data that indicates the quality of active pharmaceutical ingredients.
- the optimization device of the present disclosure analyzes the values of a plurality of parameters that define the culture conditions when cells are seeded in a medium and the quality obtained by analyzing the drug substance of the biopharmaceutical manufactured by culturing the cells an estimating unit for estimating an optimal combination of multiple parameter values based on the values of the multiple parameters received by the receiving unit and the quality data; and the multiple parameters estimated by the estimating unit. and an output unit for outputting a combination of values of .
- the optimization method of the present disclosure includes the steps of placing cells and medium in a culture vessel and culturing the cells under culture conditions defined by the values of a plurality of parameters; and a step of estimating the optimal combination of the values of the parameters, using the values of the parameters and the quality data obtained by the analyzing step as inputs. .
- the step of culturing the cells the cells are cultured under the optimum combination of the values of the plurality of parameters estimated by the estimating step as new culturing conditions.
- the user can optimize the conditions for improving the quality of the drug substance obtained by proceeding with the culture at a stage in the middle of the culture. Therefore, based on the predicted quality of the drug substance, the user can make various changes such as reviewing the protocol, extending the incubation time, or stopping the culture at the stage of the culture process before the purification process to obtain the drug substance. As a result, the manufacturing cost of biopharmaceuticals can be reduced.
- FIG. It is a block diagram which shows the hardware constitutions of an estimation apparatus.
- 2 is a block diagram showing the hardware configuration of a learning device;
- FIG. FIG. 4 is a diagram for explaining measurement timing; It is a figure which shows an example of measurement data.
- It is a block diagram which shows an example of a structure of a prediction part.
- It is a figure which shows an example of the output result which an output part outputs.
- 4 is a flow chart showing the flow of an estimation method according to the first embodiment; 4 is a flow chart showing the flow of the learning method according to the first embodiment;
- FIG. 11 is a block diagram showing the configuration of a learning device according to a modification; It is a figure which shows the modification of a model generation part. It is a block diagram which shows the structure of the estimation apparatus concerning a modification. It is a figure which shows the modification of a prediction model.
- FIG. 12 is a diagram schematically showing the overall configuration of a prediction system according to Embodiment 2; FIG. It is a block diagram which shows the hardware constitutions of an optimization apparatus. It is a block diagram which shows an example of a structure of an optimization apparatus. 4 is a flow chart showing the flow of processing of the optimization device; It is a figure which shows the modification of the prediction system which concerns on this Embodiment 2.
- FIG. 12 is a diagram schematically showing the overall configuration of a prediction system according to Embodiment 2; FIG. It is a block diagram which shows the hardware constitutions of an optimization apparatus. It is a block diagram which shows an example of a structure of an optimization apparatus. 4 is a flow chart showing the flow of processing of
- FIG. 1 is a diagram schematically showing the overall configuration of the prediction system according to the first embodiment.
- the prediction system SYS is a system for predicting the quality of biopharmaceutical drug substances manufactured by cell culture.
- biopharmaceuticals are pharmaceuticals manufactured using cells, such as antibody pharmaceuticals and vaccines.
- biopharmaceuticals may also include cells themselves used in regenerative medicine.
- the "drug substance” is a target substance obtained by cell culture, for example, after a culture process for culturing cells, it is obtained through a purification process for extracting the target component.
- Prediction system SYS includes measurement device 100 , estimation device 200 , learning device 300 , and server 400 .
- the measurement device 100 measures the substance in the culture container 10 containing the cells 1 and the medium 2 as the measurement target.
- the substance to be measured is preferably a substance that changes during the process of culturing the cell 1, such as the cell 1 itself, metabolites of the cell 1, nutrients of the cell 1, and the like.
- the user can select any measuring device 100 according to the object to be measured. For example, if the object to be measured is cell 1, the user can determine the number of cells in the culture solution composed of cell 1 and medium 2 in culture vessel 10, the viability, the shape of the cell (circularity, length A microscope and software for analyzing an image obtained by the microscope can be selected as the measurement device 100 in order to obtain measurement results such as the length of the axis, the length of the minor axis, and the like.
- the measurement target is a metabolite or nutrient of the cell 1
- the user uses a liquid chromatography mass spectrometer as the measurement device 100 to obtain measurement results such as the concentration of metabolites and nutrients in the culture medium.
- LC-MS liquid chromatography mass spectrometer
- ICP Inductively Coupled Plasma
- the estimating device 200 predicts the quality of the drug substance using the measurement data 510 including the measurement results for the measurement target after a predetermined period of time has passed since the cells 1 were seeded in the medium 2.
- the measurement data 510 is obtained by collecting the culture medium in the culture container 10 at least one timing after a predetermined period of time has passed since the cells 1 were seeded in the medium 2, and using the measurement device 100. obtained by measuring the substance of
- estimation apparatus 200 includes reception unit 210, prediction unit 220, and output unit 230. Prepare.
- the reception unit 210 receives input of the measurement data 510 .
- the prediction unit 220 inputs the measurement data 510 received by the reception unit 210 into the prediction model 420 to generate quality prediction data 540 indicating the quality of the drug substance.
- the prediction model 420 is a model for predicting the quality of the drug substance of the biopharmaceutical manufactured by culturing the cells 1, receives the input of the measurement data 510, and produces a prediction result indicating the quality of the drug substance. Output.
- the predictive model 420 is generated by the model generation unit 320 of the learning device 300 executing supervised learning processing.
- the quality prediction data 540 may be the prediction result output from the prediction model 420 itself, or may be generated based on the prediction result output from the prediction model 420.
- a prediction result output from the prediction model 420 is a result corresponding to the quality data 520 used as correct data when the model generation unit 320 generates. For example, if the quality data 520 is the data itself indicating the analysis result obtained by performing any analysis on the drug substance 3, the prediction model 420 outputs the predicted analysis result as the prediction result. When the quality data 520 is data indicating the degree of similarity with the target drug substance obtained based on the analysis result, the prediction model 420 outputs the degree of similarity predicted as the prediction result.
- the output unit 230 outputs the quality prediction data 540 generated by the prediction unit 220 to any output destination such as a display, printer, or server.
- the learning device 300 uses the learning data 530 to generate the prediction model 420 .
- the learning device 300 includes a reception unit 310 and a model generation unit 320 as an example of a software configuration implemented by the processor 31 (see FIG. 3) executing the learning program 580 (see FIG. 3).
- the reception unit 310 receives input of the learning data 530 .
- Learning data 530 includes time-series data 512 and quality data 520 .
- the time-series data 512 is one type of the measurement data 510, and measures at least one substance in the culture vessel 10 at a plurality of timings (timings t1, . obtained by The time-series data 512 is data indicating the change over time of the substance in the culture container 10 containing the cells 1 and the medium 2 as the object of measurement. The time-series data 512 is obtained by collecting the culture medium in the culture container 10 at a plurality of timings (timing t1, .
- the time-series data 512 indicates time-dependent changes in cell count, time-dependent changes in cell viability, time-dependent changes in cell shape, time-dependent changes in nutrient concentrations, time-dependent changes in metabolite concentrations, and the like.
- the quality data 520 is obtained by analyzing the drug substance 3 of the biopharmaceutical manufactured from the cell 1.
- the quality data 520 is data obtained through the purification process and the analysis process after culturing the cells 1, and is obtained by performing arbitrary analysis in the analysis process on the drug substance 3 obtained through the purification process.
- 2 is data showing the quality of drug substance 3 obtained.
- the analysis performed in the analysis step is arbitrarily selected according to the type of drug substance 3 .
- the quality of drug substance 3 is assessed in terms of one or more aspects, such as physical properties, chemical properties, biological activity, immunochemical properties, and the like.
- the quality data 520 may be data indicating an analysis result obtained by performing an arbitrary analysis on the drug substance 3, and also indicates the degree of similarity with the target drug substance obtained based on the analysis result. It may be data.
- the quality data 520 may include data indicating how similar the amino acid composition is to the target drug substance as a result of the amino acid composition analysis.
- the model generation unit 320 generates the prediction model 420 by executing the learning process using the quality data 520 included in the learning data 530 as correct data. For example, the model generation unit 320 inputs the time series data 512 included in the learning data 530 to the prediction model 420, obtains the error between the output quality prediction result and the quality data 520, which is the correct data, Optimize predictive model 420 so that the error is small.
- the server 400 is an example of a storage device for storing the prediction model 420.
- the prediction model 420 may be stored in the storage 23 (see FIG. 2) of the estimation device 200, or may be stored in the storage 33 (see FIG. 3) of the learning device 300.
- the estimation device 200 and the learning device 300 may be realized by one device.
- reception unit 210 and reception unit 310 may be implemented as one reception unit.
- the culture conditions are the same when generating the prediction model 420 and when generating the quality prediction data 540 using the generated prediction model 420 .
- the same “culturing conditions” means culturing according to a common protocol.
- the drug substance obtained by cell culture may not be the desired drug substance because the cells themselves are different and the drug used may contain components derived from living cells. may exhibit different characteristics. Therefore, in order to ensure the quality of biopharmaceuticals, it is common to perform characterization of drug substances obtained by cell culture.
- the drug substance 3 is obtained by the activity of the cell 1. Therefore, it is expected that there is a relationship between the properties of the drug substance 3 and the activity of the cells 1 .
- the substances in the culture vessel 10 containing the cells 1 show different changes over time depending on whether the activities such as proliferation and metabolism of the cells 1 are active and when they are not active. Therefore, it is expected that there is a relationship between the state of substances in the culture vessel 10 and the activity of the cells 1 . In other words, it is expected that there is an indirect relationship between the state of substances in the culture vessel 10 and the properties of the drug substance 3 .
- the learning device 300 performs learning processing using learning data 530 including time-series data 512 indicating changes over time of substances in the culture container 10 and quality data 520 indicating the quality of the drug substance 3. Generate predictive model 420 . In other words, the learning device 300 learns the relationship between the change over time of the substances in the culture container 10 and the quality of the drug substance 3 to generate the prediction model 420 .
- the user repeats the culture test including the culture process and the purification process while changing the culture conditions.
- Quality data 520 is obtained for each culture test.
- the measuring device 100 outputs time-series data 512 for each culture test.
- a large number of learning data 530 corresponding to the number of culture tests is collected by repeating the culture test.
- Learning device 300 iteratively optimizes predictive model 420 based on a large amount of learning data.
- the measurement data 510 is obtained by sampling the culture solution in the culture vessel 10 at least one timing after a predetermined period of time has passed since the cells 1 were seeded in the medium 2, and using the measurement device 100. It is obtained by measuring substances in the collected culture medium. That is, the measurement data 510 indicates the state of substances in the culture medium at least one timing after a predetermined period of time has elapsed since the cells 1 were seeded in the medium 2 .
- the measurement data 510 indicates the state of the substance after a predetermined period of time has passed since the cells 1 were seeded in the medium 2, it was generated by learning the relationship between the change over time of the substance and the quality of the drug substance 3.
- the predictive model 420 By inputting the measurement data 510 into the predictive model 420, a predictive result indicating the quality of the drug substance 3 obtained after proceeding with the culture is obtained.
- the estimating apparatus 200 can culture the cells 1 in the culture container 10 based on the measurement data 510 including the measurement results obtained by measuring the measurement target in the culture container 10. to generate quality prediction data 540 indicating the quality of the drug substance obtained by advancing Therefore, it is possible to predict the quality of the drug substance obtained by proceeding with the culture in the middle of the culture.
- the learning device 300 can generate a prediction model for predicting the quality of the drug substance obtained by advancing the culture during the culture. Based on the quality prediction data 540, the user can take various measures such as reviewing the protocol, extending the culture time, or stopping the culture at the stage of the culture process before the purification process for obtaining the drug substance. As a result, the manufacturing cost of biopharmaceuticals can be reduced.
- the learning device 300 uses learning data 530 including time-series data 512 indicating temporal changes in substances in the culture vessel 10 and quality data 520 indicating the quality of the drug substance 3.
- learning data 530 including time-series data 512 indicating temporal changes in substances in the culture vessel 10
- quality data 520 indicating the quality of the drug substance 3.
- FIG. 2 is a block diagram showing the hardware configuration of the estimation device 200.
- the estimating device 200 includes a processor 21, a main memory 22, a storage 23, a communication interface (I/F) 24, a USB (Universal Serial Bus) interface (I/F) 25, an input unit 26, and a display unit. 27. These components are connected via a processor bus 28 .
- I/F communication interface
- I/F Universal Serial Bus
- I/F Universal Serial Bus
- the processor 21 is composed of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), etc., reads out a program stored in the storage 23 (for example, an estimation program 560), expands it in the main memory 22, and executes it. .
- the processor 21 executes a program to perform a series of processes for predicting the quality of drug substances.
- the main memory 22 is composed of a volatile storage device such as RAM (Random Access Memory) or DRAM (Dynamic RAM), for example.
- the storage 23 is configured by, for example, a non-volatile storage device such as a HDD (Hard Disk Drive) or an SSD (Solid State Drive).
- the storage 23 contains measurement data 510 sent via the USB I/F 25, quality prediction data 540 generated based on the measurement data 510, and an estimation program for performing a series of processes for predicting the quality of the drug substance. 560 and the like are stored.
- the communication I/F 24 uses wired communication or wireless communication to exchange signals with the server 400 storing the prediction model 420 .
- a USB memory (not shown) is detachably attached to the USB I/F 25, and the measurement data 510 stored in the USB memory is read. Note that the measurement data 510 may be sent from the measuring device 100 to the estimating device 200 by connecting the measuring device 100 and the estimating device 200 in a communicable manner.
- the input unit 26 accepts user operations and is typically composed of a touch panel, keyboard, mouse, and the like.
- the display unit 27 is an example of an output destination of the quality prediction data 540, and is configured by a liquid crystal panel or the like capable of displaying an image.
- FIG. 3 is a block diagram showing the hardware configuration of the learning device 300.
- the learning device 300 includes a processor 31 , a main memory 32 , a storage 33 , a communication I/F 34 , a USB I/F 35 , an input section 36 and a display section 37 . These components are connected via a processor bus 38 .
- the processor 31 is composed of a CPU, GPU, etc., reads a program (for example, a learning program 580) stored in the storage 33, expands it in the main memory 32, and executes it. Processor 31 performs a series of processes for generating prediction model 420 by executing a program.
- a program for example, a learning program 580
- the main memory 32 is composed of a volatile storage device such as RAM and DRAM, for example.
- the storage 33 is configured by, for example, a non-volatile storage device such as an HDD or SSD.
- the storage 33 stores learning data 530 sent via the USB I/F 35 and a learning program 580 for performing a series of processes for generating the prediction model 420 .
- the communication I/F 34 uses wired or wireless communication to exchange signals with the server 400 storing the prediction model 420 .
- a USB memory (not shown) is detachably attached to the USB I/F 35, and the learning data 530 stored in the USB memory is read.
- the learning data 530 may be sent from the information processing device to the estimation device 200 .
- the input unit 36 accepts user operations and is typically composed of a touch panel, keyboard, mouse, and the like.
- the display unit 37 is composed of a liquid crystal panel or the like capable of displaying an image.
- the measurement data 510 received by the estimation device 200 for predicting the quality of the drug substance 3 includes measurement results obtained by measuring the measurement target at least one timing after a predetermined period of time has elapsed since the cells 1 were seeded in the medium 2. I made it.
- the measurement data 510 used to predict the quality of the drug substance 3 is the measurement result obtained by measuring the measurement target at a plurality of timings after the cells 1 are seeded in the medium 2. It may be time-series data 512 that indicates changes over time.
- FIG. 4 is a diagram for explaining the measurement timing.
- a typical cell growth curve is shown in FIG.
- FIG. 4 in general, immediately after cells 1 are seeded in medium 2, they do not grow immediately, but enter a logarithmic growth phase in which they grow through a lag phase (or lag phase). After that, the logarithmic growth phase ends and the stationary phase (also called the stationary phase) begins due to a decrease in nutrients or accumulation of waste products (metabolites). In the stationary phase, there is a balance between the number of cells 1 proliferating and the number of cells 1 dying. After the stationary phase continues for a while, when the number of dying cells 1 increases, the death phase (or decline phase) begins.
- the measurement data 510 may include measurement results obtained by measuring the measurement target at the timing t0 when culturing the cell 1 is started. In addition, when the culture of the cells 1 is started, the timing before seeding the cells 1 in the medium 2 and the timing immediately after seeding the cells 1 in the medium 2 can be included.
- the measurement result at timing t0 indicates the initial state inside the culture container 10 . If the measurement data 510 includes a measurement result indicating the initial state inside the culture vessel 10, the measurement data 510 indicates a change from the initial state.
- the measurement data 510 may include measurement results obtained by measuring the measurement target between timing t1 and timing t2, which are logarithmic growth phases. That is, at least one timing after a predetermined period of time has elapsed since seeding the cells 1 in the medium 2 is timing between timing t1 and timing t2, which are logarithmic growth phases. In the logarithmic growth phase, the number of cells 1 fluctuates greatly, so the state of substances in the culture vessel 10 fluctuates greatly.
- the estimation device 200 can use the data more accurately showing the state change of the cell 1 to obtain the drug substance 3. can predict the quality of
- the measurement data 510 may include measurement results obtained by measuring the object to be measured during the stationary period from timing t2 to timing t3. That is, at least one timing after the lapse of a predetermined period from the seeding of the cells 1 in the medium 2 is the timing between timing t2 and timing t3, which is the stationary phase. In the stationary phase, there is a balance between the number of cells 1 proliferating and the number of cells 1 dying, and the number of cells 1 is maximal. In this way, by including in the time-series data 512 the result of measuring the object to be measured at the timing when the number of cells 1 reaches its maximum, the estimating device 200 calculates the culture container 10 when the number of cells 1 reaches its maximum.
- the data describing the state of the substances in the drug substance 3 can be used to predict the quality of the drug substance 3.
- the measurement data 510 includes the measurement results measured during the death period after timing t3.
- Cell death exposes intracellular enzymes to the culture medium, which may affect the quality of the drug substance. In this case, measurements taken during cell death could serve as an example of poor quality data for machine learning.
- FIG. 5 is a diagram showing an example of measurement data 510.
- FIG. 5 the time-series data 512a to 512f measured at a plurality of timings after the cells 1 are seeded in the medium 2 are shown as an example. is seeded in the medium 2, and at least one measurement result obtained by measuring the measurement object at a timing after a predetermined period has elapsed may be included.
- the measurement data 510 are measurement results obtained by measuring the cell 1, such as time-series data 512a of cell count, time-series data 512b of cell viability, time-series data 512c of cell shape, or time-series data of microscopic images. data 512d and so on.
- the measurement data 510 includes, for example, time-series data 512a of metabolites as measurement results obtained by measuring metabolites produced by the metabolism of the cell 1.
- the metabolite includes at least one of a metabolite secreted by the cell 1 and a metabolite contained in the cell 1 itself.
- the measurement data 510 includes time-series data 512f of medium components, for example, as measurement results obtained by measuring nutrients taken by the cells 1.
- the time-series data 512f of medium components can be the time-series data of metabolites.
- the substance to be measured is at least one of the nutrients ingested by the cell 1, the metabolites produced by the metabolism of the cell 1, and the cell 1.
- the measurement data 510 may include information such as the pH and temperature of the culture solution in the culture vessel 10 .
- Time series data 512 included in learning data 530 received by learning device 300 may also include multiple types of time series data, similar to measurement data 510 received by estimating device 200 .
- the time-series data 512 included in the learning data 530 includes at least one substance among the nutrients taken in by the cell 1, the metabolites produced by the metabolism of the cell 1, and the substance of the cell 1 as measurement targets. It may also indicate changes over time of the object.
- the time-series data 512 included in the learning data 530 is time-series data of cell count, time-series data of cell viability, time-series data of cell shape, or time-series data of cell shape obtained by measuring cells. Includes time-series data of microscope images.
- the time-series data 512 included in the learning data 530 includes time-series data of metabolites obtained by measuring metabolites.
- the time-series data 512 included in the learning data 530 includes time-series data of medium components obtained by measuring nutrients. Note that the time-series data 512 included in the learning data 530 may include information such as pH and temperature of the culture solution in the culture container 10 .
- time-series data 512 used when generating the prediction model 420 and the measurement data 510 input to the prediction model 420 when generating the quality prediction data 540 using the prediction model 420 are common to each other. It includes at least the measurement results obtained with the substance as the measurement target. That is, when the prediction model 420 is generated using time-series data indicating changes in the number of cells over time, the measurement data 510 input to the prediction model 420 includes at least measurement results indicating the number of cells.
- FIG. 6 is a block diagram showing an example of the configuration of the prediction section 220.
- the prediction unit 220 includes an estimation processing unit 222 and a determination unit 224 .
- the estimation processing unit 222 inputs the measurement data 510 received by the receiving unit 210 to the prediction model 420 to obtain a prediction result.
- the prediction result is the prediction value 542, for example.
- the prediction model 420 may be configured to output a prediction value 542 when evaluated from a specific point of view as a prediction result.
- the predicted value 542 is, for example, an estimated value of the result obtained when the drug substance 3 obtained from cells in culture is analyzed from a specific point of view, or a difference between the obtained estimated value and the result of the target drug substance. It is a degree of similarity indicating whether the degree of similarity is similar.
- Viewpoints for evaluating the quality of drug substance 3 include, for example, physical properties, chemical properties, biological activity, and immunochemical properties.
- Prediction model 420 may be configured to output a plurality of prediction values obtained when drug substance 3 is evaluated from each of a plurality of perspectives.
- the prediction model 420 has a first predicted value obtained when evaluated from a first viewpoint (eg, physical properties) and a second predicted value obtained when evaluated from a second viewpoint (eg, biological activity) and may be output.
- the prediction model 420 includes a first prediction model for outputting a first prediction value when evaluating the quality of the drug substance from the first viewpoint, and a second prediction model for outputting a first prediction value when evaluating the quality of the drug substance from the second viewpoint. and a second prediction model for outputting a prediction value.
- the determination unit 224 determines the quality of the drug substance 3 based on the predicted value 542 and generates a determination result 544 .
- the judgment result 544 may be a two-level evaluation result (eg, good/bad), or may be a multiple-level evaluation result (eg, A, B, C, etc.), or may be a score. It may be a form of evaluation result (eg, 50 points, 95 points, etc.).
- the determination unit 224 determines the quality of the drug substance 3 from the plurality of predicted values. is determined to generate a determination result 544 .
- the determination unit 224 may weight each of the plurality of predicted values according to the degree of importance of the viewpoint.
- the quality prediction data 540 includes a prediction value 542 generated by the estimation processing unit 222 and a determination result 544 generated by the determination unit 224.
- the quality prediction data 540 includes a plurality of predicted values (first 1 predicted value, 2nd predicted value, etc.).
- the quality prediction data 540 includes a plurality of predicted values obtained when evaluating the drug substance 3 from each of a plurality of viewpoints, so that the quality of the drug substance can be evaluated from various viewpoints.
- the quality prediction data 540 includes the determination result 544, the prediction result for the quality of the drug substance can be presented in a more comprehensible manner.
- the output unit 230 outputs the quality prediction data 540.
- FIG. 7 is a diagram showing an example of the output result output by the output unit 230. As shown in FIG.
- the output unit 230 may output determination results 544 together with predicted values 542 obtained for each viewpoint.
- the prediction unit 220 does not have to include the determination unit 224 that comprehensively determines the quality of the drug substance.
- FIG. 8 is a flow chart showing the flow of the estimation method according to the first embodiment.
- the estimation method comprises a culture step (S100) of culturing cells, a measurement step (S200) of measuring substances in the culture vessel, and a prediction step of predicting the quality of drug substance 3 based on measurement data 510 including measurement results. (S300) and an output step (S400) of outputting the quality prediction data 540 obtained in the prediction step.
- the culturing step (S100) includes a step of seeding cells 1 in medium 2 (S110).
- the culturing step (S100) may include a step of replacing the medium 2 (S112), a step of adding a reagent (S114), and the like, depending on the culturing conditions (predetermined protocol).
- the medium 2 may be either liquid or solid.
- the measuring step (S200) may be performed once or more during the culturing step (S100).
- the measurement step (S200) includes, for example, the timing of the step of seeding the cells 1 in the medium 2 (S110), the timing of the step of replacing the medium 2 (S112), and the step of adding the reagent (S114). It can be done at any time, such as when
- the measurement timing includes the timing at which the step of seeding the cells 1 in the medium 2 (S110) is performed, the timing at which the step of replacing the medium 2 (S112) is performed, and the timing at which the reagent addition step (S114) is performed.
- the measurement timing may be during the logarithmic growth phase, during the stationary phase, or during the death phase.
- the substance in the culture vessel 10 is measured at least at one timing after a predetermined period of time has passed since the cells 1 were seeded in the medium 2. measurement results are obtained.
- the culture step (S100) and the measurement step (S200) may be performed manually, or may be automated or semi-automated by a machine.
- the prediction step (S300) includes a step of receiving input of measured data 510 (S310), a step of inputting measured data 510 to prediction model 420 (S312), and a step of obtaining prediction results from prediction model 420 (S314). , and generating quality prediction data 540 (S316). That is, in the prediction step (S300), the quality prediction data 540 is generated by inputting the measurement data 510 including the measurement results obtained in the measurement step (S200) into the prediction model 420.
- S310 to S316 and S400 are specifically performed by the estimating device 200.
- FIG. 9 is a flow chart showing the flow of the learning method according to the first embodiment.
- the learning method includes a culture step (S10) for culturing cells, a measurement step (S20) for measuring substances in the culture vessel, a purification step (S30) for obtaining a drug substance, and an analysis for analyzing the quality of the drug substance. It includes a step (S40) and a learning step (S550) of executing a learning process using learning data 530 to generate prediction model 420.
- FIG. 10 includes a culture step (S10) for culturing cells, a measurement step (S20) for measuring substances in the culture vessel, a purification step (S30) for obtaining a drug substance, and an analysis for analyzing the quality of the drug substance. It includes a step (S40) and a learning step (S550) of executing a learning process using learning data 530 to generate prediction model 420.
- the culture step (S10) is common to the culture step (S100) included in the estimation method.
- the culturing step (S10) includes a step of seeding cells 1 in medium 2 (S11).
- the culture step (S10) may include a step of replacing the medium 2 (S12), a step of adding a reagent (S13), and the like, depending on the culture conditions (predetermined protocol).
- the medium 2 may be either liquid or solid.
- a plurality of measurement steps (S20) are performed during the culture step (S10).
- time-series data 512 indicating changes over time of the object to be measured is obtained.
- the measurement step (S200) includes, for example, the timing of the step of seeding the cells 1 in the medium 2 (S11), the timing of the step of replacing the medium 2 (S12), and the step of adding the reagent (S13). It can be done at any time, such as when
- the measurement timing includes the timing at which the step (S11) of seeding the cells 1 in the medium 2 is performed, the timing at which the step (S112) of replacing the medium 2 is performed, and the timing at which the step (S114) of adding a reagent is performed.
- the measurement timing may be during the logarithmic growth phase, during the stationary phase, or during the death phase.
- the substances in the culture vessel 10 are measured at multiple timings after the cells 1 are seeded in the medium 2, and the measurement results are obtained.
- the culture step (S10) and the measurement step (S20) may be performed manually, or may be automated or semi-automated by a machine.
- the purification step (S30) is a task of extracting only the target component from the culture solution in the culture vessel 10. Through the purification step (S30), a drug substance containing the target component with high purity can be obtained.
- the analysis step (S40) one or more types of analysis are performed to obtain quality data 520 in order to evaluate the drug substance from one or more perspectives.
- the quality data 520 may include multiple types of analysis results (analysis results).
- the learning step (S50) includes a step (S51) of receiving an input of learning data 530, a step (S52) of performing a learning process using the learning data 530, and outputting a prediction model 420 generated by the learning process. and a step (S53).
- the output destination of prediction model 420 is server 400 .
- the prediction model 420 may be stored in the storage 33 of the learning device 300 without being output. S51 to S53 are specifically performed by the learning device 300. FIG.
- learning device 300 acquires learning data 530 from the outside.
- the learning device may have a function of generating the learning data 530 .
- the measurement timing for obtaining measurement data 510 received by estimation apparatus 200 has been described with reference to FIG.
- the cell 1 growth curve differs depending on the cell 1 type. Therefore, the learning device may have a function for specifying measurement timing at the time of prediction for generating quality prediction data 540 .
- FIG. 10 is a block diagram showing the configuration of a learning device according to a modification.
- FIG. 10 shows a configuration that the learning apparatus 300 according to the first embodiment does not have.
- Learning device 300 a differs from learning device 300 in that it further includes a learning data generation unit 312 and a timing identification unit 330 .
- the learning data generation unit 312 generates learning data 530 based on the time series data 512 and the quality data 520 .
- the learning data generation unit 312 sends the generated learning data 530 to the reception unit 310 (not shown).
- the receiving unit 310 receives input of the sent learning data 530 .
- the timing identification unit 330 identifies the measurement timing at the time of prediction for generating the quality prediction data 540 based on the cell count time series data 512a. More specifically, the timing identifying unit 330 identifies the lag phase, the logarithmic growth phase, the stationary phase, and the death phase based on the time-series data 512a (see FIG. 4), and can identify timings t1 to t3. generate useful information.
- the information that can specify the timings t1 to t3 includes, for example, the time after seeding the cells 1 in the medium 2. FIG. The user can determine the measurement timing during prediction based on the information that can specify the timings t1 to t3.
- the culture conditions are the same when generating the prediction model 420 and when generating the quality prediction data 540 using the generated prediction model 420 . It should be noted that the culture conditions may not be common between when the prediction model 420 is generated and when the quality prediction data 540 is generated using the generated prediction model 420 . A learning method and an estimation method in this case will be described with reference to FIGS. 11 to 13. FIG.
- FIG. 11 is a diagram showing a modified example of the model generation unit.
- a model generation unit 320a according to a modification generates a prediction model 420a based on multiple types of learning data 530a and 530b.
- Each of the learning data 530 a and 530 b includes condition data 550 .
- the condition data 550 is information indicating culture conditions. For example, by changing the culture conditions and repeating the culture process, measurement process, purification process, and analysis process (see FIG. 9), learning data 530a and 530b generated under mutually different culture conditions can be obtained.
- “Different culture conditions” means that the components of the medium 2 are different, the types of reagents to be added are different, and the types of chemicals used are different, and in addition, the temperature conditions are different, and the reagent addition timings are different. etc. can also be included.
- the condition data 550 may include information on chemicals used, temperature conditions, reagent addition timing, and the like. Note that the condition data 550 is not information on a specific culture procedure, but information merely indicating that the learning data 530a and 530b were obtained under mutually different culture conditions (for example, condition A and condition B). etc.).
- the model generation unit 320a determines the relationship between the change in the substance in the culture vessel 10 over time and the quality of the drug substance 3, as well as the relationship between the difference in culture conditions and the change in the substance in the culture vessel 10 over time. Alternatively, the relationship between the difference in culture conditions and the quality of the drug substance 3 is learned to generate the prediction model 420a.
- FIG. 12 is a block diagram showing the configuration of an estimation device according to a modification.
- FIG. 12 omits the description of the configuration of the output unit according to the first embodiment.
- Estimating apparatus 200a differs from estimating apparatus 200 according to the first embodiment in that it includes reception section 210a instead of reception section 210 and prediction section 220a instead of prediction section 220 .
- the reception unit 210a receives input of condition data 550 indicating culture conditions.
- the prediction unit 220a inputs the condition data 550 in addition to the measurement data 510 to the prediction model 420a generated by the model generation unit 320a shown in FIG. 13 to obtain a prediction result indicating the quality of the drug substance.
- the prediction model 420a is a model that takes into account the relationship between the difference in culture conditions and the change over time of the substance in the culture vessel 10, or the relationship between the difference in culture conditions and the quality of the drug substance 3. It can more accurately predict the quality of drug substance 3 compared to predictive model 420 . Even if the culture conditions differ between when the prediction model 420a is generated and when the quality prediction data 540 is generated using the generated prediction model 420a, the prediction model 420a takes into account the difference in the culture conditions. Therefore, by inputting the condition data 550 in addition to the measurement data 510 to the prediction model 420a, the prediction unit 220a can obtain a prediction result indicating the quality of the drug substance. That is, the prediction unit 220a can obtain a prediction result that considers the relationship between the difference in culture conditions and the quality.
- FIG. 13 is a diagram showing a modification of the prediction model.
- the model generation unit may generate a prediction model for each culture condition, such as a prediction model 420-1 dedicated to culture condition A and a prediction model 420-2 dedicated to culture condition B. good.
- the estimation device may use a prediction model that matches the culture conditions based on the condition data 550.
- FIG. 14 is a diagram schematically showing the overall configuration of the prediction system according to the second embodiment.
- the prediction system SYS2 like the prediction system SYS1, has a function of predicting the quality of biopharmaceutical drug substances manufactured by cell culture.
- the prediction system SYS2 has a function of optimizing culture conditions in addition to the functions of the prediction system SYS1.
- the prediction system SYS2 includes an optimization device 600 that optimizes culture conditions.
- the optimization device 600 estimates optimal culture conditions for obtaining a high-quality drug substance 3 based on the time-series data 512 and quality data 520 included in the learning data 530.
- the optimization device 600 outputs the estimated optimal culture conditions.
- the user performs a new culture test based on the culture conditions output from the optimization device 600 .
- the culture test includes a culture step (step S10), a purification step (step S30), and an analysis step (step S40). Each of these steps is the same as each step described as the first embodiment. The details of each step are shown in FIG. 9, so the description thereof will not be repeated here.
- new learning data 530 is input to the optimization device 600 .
- the optimization device 600 re-estimates the optimal culture conditions using the previously input learning data 530 and the newly input learning data 530 .
- the optimization device 600 outputs newly estimated culture conditions. Based on the culture conditions output from the optimization device 600, the user carries out the next culture test.
- the culture conditions are optimized as the number of culture tests increases.
- the properties of the learning data 530 change in the direction toward the generation of the drug substance 3 with higher quality.
- the quality of the learning data 530 increases as the culture conditions are optimized.
- the optimization device 600 outputs optimized culture conditions according to user instructions.
- the user can culture cells based on the culture conditions output from the optimization device 600 when predicting the quality of the drug substance using the estimation device 200 . Therefore, a user can predict drug substance quality in an environment where high quality drug substances can be produced.
- the optimization device 600 functions extremely effectively in both the process of generating the prediction model 420 and the process of using the prediction model 420 to predict the quality of the drug substance. According to the prediction system SYS2 according to the second embodiment, the possibility of efficiently obtaining higher-quality drug substances increases.
- FIG. 15 is a block diagram showing the hardware configuration of the optimization device 600.
- the optimization device 600 includes a processor 61 , a main memory 62 , a storage 63 , a communication I/F 64 , a USB I/F 65 , an input section 66 and a display section 67 . These components are connected via a processor bus 68 .
- the processor 61 is composed of a CPU, GPU, etc., reads a program (for example, an optimization program 630) stored in the storage 63, develops it in the main memory 62, and executes it.
- the processor 61 executes a program to perform a series of processes for searching for optimum values of parameters of culture conditions.
- the main memory 62 is composed of a volatile storage device such as RAM and DRAM, for example.
- the storage 63 is composed of, for example, a non-volatile storage device such as an HDD or SSD.
- Storage 63 stores learning (estimation) data 530 sent via USB I/F 65, optimization program 630 for searching optimal values 620 of parameters of culture conditions, and culture data newly determined by processor 61. Optimal values 620 for the parameters of the condition are stored.
- the communication I/F 64 uses wired communication or wireless communication to exchange signals with other communication devices.
- a USB memory (not shown) is detachably attached to the USB I/F 65, and the learning data 530 stored in the USB memory is read.
- the input unit 66 accepts user operations and is typically composed of a touch panel, keyboard, mouse, and the like.
- the display unit 67 is composed of a liquid crystal panel or the like capable of displaying an image.
- FIG. 16 is a block diagram showing an example of the configuration of the optimization device 600.
- Optimizing device 600 includes a receiving unit 601, an estimating unit 602, and an output unit 603 as an example of a software configuration realized by processor 61 (see FIG. 15) executing optimization program 630 (see FIG. 15). and a storage unit 604 .
- the optimization device 600 searches for culture conditions (hereinafter also referred to as optimal solutions) that increase the quality level of the drug substance 3 by Bayesian optimization.
- the culture condition is a combination of multiple parameter values.
- Bayesian optimization is a method of estimating an optimal event from observed events by a statistical approach based on Bayesian probability. In Bayesian optimization, a trial based on the set optimal solution is executed, and another optimal solution is searched for based on the results of the trial. The searched optimal solution is set as the optimal solution to be used for the next trial.
- the reception unit 601 receives estimation data.
- the estimation data is included in learning data 530 shown in FIG.
- the estimation data includes at least part of the time-series data 512 and quality data 520 .
- the estimation unit 602 searches for the optimal solution of the culture conditions by Bayesian optimization by executing the estimation program.
- the storage unit 604 includes a plurality of memories that store an estimation program, input estimation data, and searched optimal solutions.
- the output unit 603 is, for example, the display unit 67 (FIG. 15) that displays the optimal solution of the culture conditions.
- the output unit 603 may be an interface for outputting the optimum solution of culture conditions to a display or printer.
- one piece of estimation data consists of culture condition data and quality data used in one culture test.
- the culture condition data consists of a plurality of parameters p1, p2, . . . pn that define the culture conditions.
- Various parameters are, for example, nutrient concentration, pH, temperature, humidity, medium components, types of reagents to be added, reagent addition timings, and the like.
- Various parameters are, for example, medium conditions collected at the timing of seeding cells in the medium. This timing corresponds to the measurement timing t0 shown in FIG. That is, the values of various parameters are the values of the parameters measured at the measurement timing t0 in the time-series data 512 shown in FIG. However, values of various parameters measured at various measurement timings shown in FIG. For example, values representing changes in concentration of nutrients over time, changes in concentration of metabolites over time, changes in pH of the culture solution in the culture container, changes in temperature, changes in humidity, etc. are input to the reception unit 601 as data for estimation. may The estimation unit 602 may search for optimal culture conditions based on these estimation data.
- FIG. 17 is a flowchart showing the processing flow of the optimization device 600. As shown in FIG. The flow of processing for estimating optimal culture conditions by the optimization device 600 will be described below with reference to FIGS. 16 and 17. FIG.
- the estimation unit 602 determines whether estimation data has been input to the reception unit 601 (step S600).
- the estimation unit 602 stores the input estimation data in the storage unit 604 (step S601).
- the estimation data input to the reception unit 601 is accumulated in the storage unit 604 .
- the estimating unit 602 uses all the estimation data stored in the storage unit 604 to determine values of parameters p1, p2, .
- the estimating unit 602 searches for culture conditions that give the optimum value to the quality level of the drug substance 3 by Bayesian optimization.
- step S602 it is conceivable to perform the processing of step S602 every time 4 to 10 estimation data corresponding to 4 to 10 culture tests are input.
- the estimation unit 602 stores the newly determined optimum values of the parameters in the storage unit 604 . If the optimal parameter values determined in the previous search are stored in the storage unit 604, the estimating unit 602 updates these values with the newly determined optimal parameter values (step S603). Next, the estimation unit 602 outputs the newly determined parameter values from the output unit 603 (step S604).
- the parameter values output from the output unit 603 are the latest parameter values indicating the optimum culture conditions.
- the user performs a new culture test once or multiple times based on the output parameter values.
- the estimation unit 602 again searches for the optimal culture condition parameter values.
- the search result is stored in the storage unit 604 as the current optimal parameter value. Therefore, by repeating the culture test and the optimization of the culture conditions by the optimization device 600, the optimization of the parameter values of the culture conditions stored in the optimization device 600 progresses.
- the estimation unit 602 determines whether or not there is a request to output the parameter value (step S605). For example, when the user wants to predict the quality of a drug substance obtained by a certain culture test using the estimation device 200, it is necessary to define the culture conditions for the culture test. It is useful to refer to the parameter values stored in the optimization device 600 when determining the culture conditions for the culture test. If it is determined in step S600 that there is a request to output the parameter value, the estimation unit 602 outputs the parameter value stored in the storage unit 604 from the output unit 603 (step S604), and ends the processing based on this flowchart. .
- FIG. 18 is a diagram showing a modification of the prediction system according to the second embodiment.
- the learning device 300 includes an optimization section 6000 having the functions of the optimization device 600 shown in FIG.
- the optimization unit 6000 includes a reception unit 601 , an estimation unit 602 , an output unit 603 , and a storage unit 604 as in the optimization device 600 .
- the learning device 300 optimizes the predictive model 420 based on the learning data 530 and searches for the optimum values of the culture condition parameters. Therefore, according to the modified example, the added value of the learning device 300 can be increased. Furthermore, according to the modified example, since it is not necessary to provide the optimization device 600 separately from the learning device 300, the cost can be reduced.
- optimization device 600 or the optimization unit 6000 may optimize the values of the culture condition parameters using a grid search technique instead of Bayesian optimization.
- An estimation device includes a reception unit that receives input of measurement data, and inputs the measurement data received by the reception unit into a prediction model to generate quality prediction data that indicates the quality of the drug substance. and an output unit for outputting quality prediction data generated by the prediction unit.
- the measurement data includes measurement results obtained by measuring at least one substance in a culture vessel containing the cells and the medium at least one timing after a predetermined period of time has passed since the cells were seeded in the medium.
- a prediction model is a model for predicting the quality of biopharmaceutical drug substances manufactured by culturing cells.
- the estimation device described in paragraph 1 it is possible to predict the quality of the drug substance obtained by proceeding with the culture in the middle of the culture. Therefore, based on the predicted quality of the drug substance, the user can make various changes such as reviewing the protocol, extending the incubation time, or stopping the culture at the stage of the culture process before the purification process to obtain the drug substance. As a result, the manufacturing cost of biopharmaceuticals can be reduced.
- the measurement data includes measurement results obtained by measuring at least one substance at a plurality of timings after seeding the cells in the medium.
- the measurement data includes measurement results obtained by measuring the at least one substance when culturing the cells is started.
- the measurement data includes measurement results obtained by measuring at least one substance during the logarithmic growth phase of the cell.
- the estimation device described in paragraph 4 uses the data that more accurately indicates the state change of the cell to determine the quality of the drug substance. can be predicted.
- the measurement data includes measurement results obtained by measuring at least one substance during the stationary phase of the cell.
- the estimating device estimates the amount of substance in the culture container when the number of cells reaches its maximum.
- Conditional data can be used to predict drug substance quality.
- the measurement data includes measurement results obtained by measuring the at least one substance during the death period of the cell. (Fig. 5).
- the cells die, exposing the intracellular enzymes to the culture medium, which may affect the quality of the drug substance.
- measurements taken during cell death could serve as an example of poor quality data for machine learning.
- the at least one substance includes nutrients ingested by cells, metabolites produced by metabolism of cells, and at least one of them.
- the estimation device described in paragraph 8 can evaluate the quality of drug substances from various viewpoints.
- the predictor includes a determiner that determines the quality of the drug substance based on the predicted value.
- the quality prediction data includes judgment results for judging the quality of the drug substance.
- the reception unit further receives input of condition data indicating cell culture conditions.
- the prediction unit generates quality prediction data by inputting the measurement data and condition data received by the reception unit into the prediction model.
- prediction results are obtained that take into account the relationship between differences in culture conditions and quality, and more accurate prediction results are obtained.
- a learning device includes a reception unit that receives learning data, and a model generation unit that generates a prediction model by executing learning processing using the learning data received by the reception unit.
- the learning data includes measurement data including measurement results obtained by measuring at least one substance in the culture vessel containing the cells and the medium at multiple timings after the cells are seeded in the medium, and the data produced from the cells. This includes quality data obtained by analyzing the drug substance of biopharmaceuticals.
- the prediction model is based on the measurement results obtained by measuring at least one substance in the culture vessel during cell culture, and the biomass produced from the cells contained in the culture vessel during cell culture. This is a model for generating quality prediction data that indicates the quality of active pharmaceutical ingredients.
- the learning device described in paragraph 11 can generate a prediction model for predicting the quality of the drug substance obtained by proceeding with the culture in the middle of the culture. Therefore, based on the predicted quality of the drug substance, the user can make various changes such as reviewing the protocol, extending the incubation time, or stopping the culture at the stage of the culture process before the purification process to obtain the drug substance. As a result, the manufacturing cost of biopharmaceuticals can be reduced.
- the learning data further includes condition data indicating cell culture conditions, the condition data includes values of a plurality of parameters defining the culture conditions, and learning
- the device further comprises an optimization unit that optimizes a combination of the values of the multiple parameters, the optimization unit determines the values of the multiple parameters and the quality obtained when the cells are cultured based on the values of the multiple parameters. Estimates the optimal combination of values for multiple parameters using data as input.
- the learning device described in paragraph 12 can estimate a combination of multiple parameter values that define cell culture conditions in the middle of culturing. Therefore, the user can optimize the conditions for improving the quality of the drug substance obtained by advancing the culture in the middle of the culture.
- the learning data further includes condition data indicating conditions for culturing the cells contained in the culture vessel.
- the learning data received by the reception unit includes first learning data obtained when cells are cultured under a first condition, and second learning data obtained when cells are cultured under a second condition different from the first condition.
- data for The model generation unit generates a prediction model by executing a learning process using the first learning data and the second learning data.
- the learning device generates a prediction model that takes into account the relationship between the difference in culture conditions and the change over time of the substance in the culture vessel, or the relationship between the difference in culture conditions and the quality of the drug substance. can. Therefore, the learning device can generate a prediction model for obtaining more accurate prediction results.
- an estimation method includes the step of placing cells and a medium in a culture vessel and culturing the cells, and at least one A step of measuring at least one substance in the culture vessel at a timing, and a step of generating quality prediction data indicating the quality of the drug substance by inputting measurement data including measurement results obtained in the measuring step into the prediction model. and outputting quality prediction data.
- a prediction model is a model for predicting the quality of biopharmaceutical drug substances manufactured by culturing cells.
- the learning method includes the steps of putting cells and a medium in a culture container and culturing the cells, and at a plurality of timings after seeding the cells in the medium, comprising the steps of measuring at least one substance; analyzing the quality of a biopharmaceutical drug substance manufactured from cells; and generating a prediction model by performing a learning process using the learning data.
- the learning data includes measurement data including measurement results obtained by the measuring step and quality data obtained by the analyzing step.
- the prediction model is based on the measurement result obtained by measuring at least one substance in the culture vessel during cell culture, and the biomass manufactured from the cells contained in the culture vessel during cell culture. It is a model for generating quality prediction data that indicates the quality of active pharmaceutical ingredients.
- a prediction model for predicting the quality of the drug substance obtained by proceeding with the culture is generated in the middle of the culture. Therefore, based on the predicted quality of the drug substance, the user can make various changes such as reviewing the protocol, extending the incubation time, or stopping the culture at the stage of the culture process before the purification process to obtain the drug substance. As a result, the manufacturing cost of biopharmaceuticals can be reduced.
- the optimization device is a biopharmaceutical drug substance manufactured by culturing the values of a plurality of parameters that define culture conditions when cells are seeded in a medium and the cells are cultured.
- a receiving unit that receives quality data obtained by analysis; an estimating unit that estimates an optimum combination of the plurality of parameter values using the values of the plurality of parameters received by the receiving unit and the quality data as inputs; an output unit that outputs a combination of the values of the parameters estimated by the estimation unit.
- a combination of multiple parameter values that define cell culture conditions can be estimated in the middle of culture. Therefore, the user can optimize the conditions for improving the quality of the drug substance obtained by advancing the culture in the middle of the culture.
- the estimating unit estimates the optimum combination of the values of the parameters by performing Bayesian optimization with the values of the parameters and the quality data obtained by the analyzing step as inputs. good.
- an optimization method includes the steps of placing cells and a medium in a culture vessel and culturing the cells under culture conditions defined by the values of a plurality of parameters; The step of analyzing the quality of the drug substance of the biopharmaceutical manufactured by culturing, and the quality data obtained by the multiple parameter values and the analyzing step are input, and the optimal combination of multiple parameter values is determined. and estimating. In the step of culturing the cells, the cells are cultured under the optimum combination of the values of the plurality of parameters estimated by the estimating step as new culturing conditions.
- the user can optimize the conditions for improving the quality of the drug substance obtained by proceeding with the culture in the middle of the culture.
- a prediction system includes the estimation device according to any one of items 1 to 10, and the estimation device according to any one of items 11 to 13.
- a learning device as described and a measuring device for measuring at least one substance in a culture vessel.
- the estimation device and the measurement device may be realized by one information processing device.
- the reception unit of the estimation device and the reception unit of the measurement device may be implemented as one reception unit. That is, the reception unit receives measurement data obtained by measuring the substance with the measurement device.
- the measurement device includes a first measurement device that measures the first substance and a second measurement device that measures the second substance.
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Organic Chemistry (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Wood Science & Technology (AREA)
- Zoology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Immunology (AREA)
- Computing Systems (AREA)
- Theoretical Computer Science (AREA)
- Crystallography & Structural Chemistry (AREA)
- Molecular Biology (AREA)
- Biochemistry (AREA)
- Biotechnology (AREA)
- Analytical Chemistry (AREA)
- Microbiology (AREA)
- Data Mining & Analysis (AREA)
- Urology & Nephrology (AREA)
- Hematology (AREA)
- Genetics & Genomics (AREA)
- General Engineering & Computer Science (AREA)
- Cell Biology (AREA)
- Pathology (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Toxicology (AREA)
- General Physics & Mathematics (AREA)
- Tropical Medicine & Parasitology (AREA)
- Databases & Information Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Physics & Mathematics (AREA)
- Proteomics, Peptides & Aminoacids (AREA)
Abstract
Description
[予測システムの全体構成]
図1は、本実施の形態1にかかる予測システムの全体構成を模式的に示す図である。予測システムSYSは、細胞培養によって製造されるバイオ医薬品の原薬の品質を予測するためのシステムである。本実施の形態1において、「バイオ医薬品」とは、細胞を利用して製造される医薬品であって、たとえば、抗体医薬品、ワクチンなどである。また、本実施の形態1において、バイオ医薬品には、再生医療に用いられる細胞そのものも含み得る。本実施の形態1において、「原薬」とは、細胞培養によって得られる目的物質であって、たとえば、細胞を培養するための培養工程の後、目的成分を取り出すための精製工程を経て得られる。予測システムSYSは、測定装置100と、推定装置200と、学習装置300と、サーバ400とを含む。
図2は、推定装置200のハードウェア構成を示すブロック図である。推定装置200は、プロセッサ21と、メインメモリ22と、ストレージ23と、通信インターフェイス(I/F)24と、USB(Universal Serial Bus)インターフェイス(I/F)25と、入力部26と、表示部27とを備える。これらのコンポーネントは、プロセッサバス28を介して接続されている。
図3は、学習装置300のハードウェア構成を示すブロック図である。学習装置300は、プロセッサ31と、メインメモリ32と、ストレージ33と、通信I/F34と、USBI/F35と、入力部36と、表示部37とを備える。これらのコンポーネントは、プロセッサバス38を介して接続されている。
原薬3の品質を予測するために推定装置200が受け付ける測定データ510は、細胞1を培地2に播種してから所定期間経過した後の少なくとも1のタイミングにおいて測定対象を測定した測定結果を含むものとした。なお、原薬3の品質を予測する際に利用する測定データ510は、細胞1を培地2に播種してからの複数のタイミングで測定対象を測定することで得られる測定結果として、測定対象の経時変化を示す時系列データ512であってもよい。
図6は、予測部220の構成の一例を示すブロック図である。予測部220は、推定処理部222と判定部224とを備える。推定処理部222は、受付部210が受け付けた測定データ510を予測モデル420に入力し、予測結果を得る。予測結果は、たとえば、予測値542である。
図8は、本実施の形態1にかかる推定方法の流れを示すフローチャートである。推定方法は、細胞を培養する培養工程(S100)と、培養容器内の物質を測定する測定工程(S200)と、測定結果を含む測定データ510に基づいて原薬3の品質を予測する予測工程(S300)と、予測工程で得られた品質予測データ540を出力する出力工程(S400)とを含む。
図9は、本実施の形態1にかかる学習方法の流れを示すフローチャートである。学習方法は、細胞を培養する培養工程(S10)と、培養容器内の物質を測定する測定工程(S20)と、原薬を取得する精製工程(S30)と、原薬の品質を解析する解析工程(S40)と、学習用データ530を用いて学習処理を実行して予測モデル420を生成する学習工程(S550)とを含む。
[学習装置の変形例]
上記実施の形態1において、学習装置300は、学習用データ530を外部から取得するものとした。なお、学習装置は、学習用データ530を生成する機能を有していても良い。また、上記実施の形態1において、図4を参照して、推定装置200が受け付ける測定データ510を得るための測定タイミングについて説明した。細胞1の増殖曲線は、細胞1の種類に応じて異なる。そのため、学習装置は、品質予測データ540を生成する予測時における測定タイミングを特定するための機能を有していても良い。
上記実施の形態1において、培養条件は、予測モデル420を生成する時と、生成された予測モデル420を用いて品質予測データ540を生成する時とで同じであるものとした。なお、予測モデル420を生成する時と、生成された予測モデル420を用いて品質予測データ540を生成する時とで、培養条件は、共通でなくともよい。この場合の学習方法と推定方法について、図11~図13を参照して説明する。
[予測システムの全体構成]
図14は、本実施の形態2にかかる予測システムの全体構成を模式的に示す図である。予測システムSYS2は、予測システムSYS1と同様に、細胞培養によって製造されるバイオ医薬品の原薬の品質を予測する機能を備える。予測システムSYS2は、予測システムSYS1が有する機能に加えて、培養条件を最適化する機能を備える。予測システムSYS2は、培養条件を最適化する最適化装置600を備える。
図15は、最適化装置600のハードウェア構成を示すブロック図である。最適化装置600は、プロセッサ61と、メインメモリ62と、ストレージ63と、通信I/F64と、USBI/F65と、入力部66と、表示部67とを備える。これらのコンポーネントは、プロセッサバス68を介して接続されている。
図16は、最適化装置600の構成の一例を示すブロック図である。最適化装置600は、プロセッサ61(図15参照)が最適化プログラム630(図15参照)を実行することで実現されるソフトウェア構成の一例として、受付部601と、推定部602と、出力部603と、記憶部604とを備える。
図17は、最適化装置600の処理の流れを示すフローチャートである。以下、図16および図17を参照して、最適化装置600が最適な培養条件を推定する処理の流れを説明する。
図18は、本実施の形態2に係る予測システムの変形例を示す図である。変形例において、学習装置300は、図16に示した最適化装置600の機能を備える最適化部6000を備えている。最適化部6000は、最適化装置600と同様に、受付部601と、推定部602と、出力部603と、記憶部604とを備える。学習装置300は、学習用データ530に基づいて予測モデル420を最適化するとともに培養条件のパラメータの最適値を探索する。したがって、変形例によれば、学習装置300の付加価値を高めることができる。さらに、変形例によれば、学習装置300と別に最適化装置600を設ける必要がないため、コストの低減を図ることができる。
上述した各実施の形態は、以下の態様の具体例であることが当業者により理解される。
Claims (17)
- 細胞を培地に播種してから所定期間経過した後の少なくとも一のタイミングで当該細胞および当該培地を含む培養容器内の少なくとも一の物質を測定して得られる測定結果を含む測定データの入力を受け付ける受付部と、
前記受付部が受け付けた前記測定データを、前記細胞を培養することで製造されるバイオ医薬品の原薬の品質を予測するための予測モデルに入力することで、前記原薬の品質を示す品質予測データを生成する予測部と、
前記予測部が生成した前記品質予測データを出力する出力部とを備える、推定装置。 - 前記測定データは、前記細胞を前記培地に播種してからの複数のタイミングで前記少なくとも一の物質を測定することで得られる測定結果を含む、請求項1に記載の推定装置。
- 前記測定データは、前記細胞の培養を開始するときに前記少なくとも一の物質を測定して得られる測定結果を含む、請求項1または請求項2に記載の推定装置。
- 前記測定データは、前記細胞の対数増殖期に前記少なくとも一の物質を測定して得られる測定結果を含む、請求項1~請求項3のうちいずれか1項に記載の推定装置。
- 前記測定データは、前記細胞の定常期に前記少なくとも一の物質を測定して得られる測定結果を含む、請求項1~請求項4のうちいずれか1項に記載の推定装置。
- 前記測定データは、前記細胞の死滅期に前記少なくとも一の物質を測定して得られる測定結果を含む、請求項1~請求項4のうちいずれか1項に記載の推定装置。
- 前記少なくとも一の物質は、前記細胞が摂取する栄養素、前記細胞の代謝により生成される代謝物、および前記細胞のうちの少なくとも一つである、請求項1~請求項6のうちいずれか1項に記載の推定装置。
- 前記予測モデルからは、前記受付部が受け付けた前記測定データを入力することで、第1観点で前記原薬の品質を評価した場合の第1予測値と、第2観点で前記原薬の品質を評価した場合の第2予測値とが出力され、
前記品質予測データは、前記第1予測値と、前記第2予測値とを含む、請求項1~請求項7のうちいずれか1項に記載の推定装置。 - 前記予測モデルからは、前記受付部が受け付けた前記測定データを入力することで、所定の観点で前記原薬の品質を評価した場合の予測値が出力され、
前記予測部は、前記予測値に基づいて前記原薬の品質を判定する判定部を含み、
前記品質予測データは、前記原薬の品質を判定した判定結果を含む、請求項1~請求項7のうちいずれか1項に記載の推定装置。 - 前記受付部は、前記細胞の培養条件を示す条件データの入力をさらに受け付け、
前記予測部は、前記受付部が受け付けた前記測定データおよび前記条件データを前記予測モデルに入力することで前記品質予測データを生成する、請求項1~請求項9のうちいずれか1項に記載の推定装置。 - 細胞を培地に播種してからの複数のタイミングで当該細胞および当該培地を含む培養容器内の少なくとも一の物質を測定して得られる測定結果を含む測定データと、前記細胞から製造されるバイオ医薬品の原薬を解析して得られる品質データと含む学習用データを受け付ける受付部と、
前記受付部が受け付けた前記学習用データを用いて学習処理を実行することにより、前記細胞を培養している途中における培養容器内の前記少なくとも一の物質を測定して得られる測定結果に基づいて、当該培養している途中における培養容器に含まれる前記細胞から製造される前記バイオ医薬品の前記原薬の品質を示す品質予測データを生成するための予測モデルを生成するモデル生成部とを備える、学習装置。 - 前記学習用データは、前記細胞の培養条件を示す条件データをさらに含み、
前記条件データは、前記培養条件を定義する複数のパラメータの値を含み、
前記学習装置は、前記複数のパラメータの値の組み合わせを最適化する最適化部をさらに備え、
前記最適化部は、前記複数のパラメータの値と前記複数のパラメータの値に基づいて前記細胞を培養したときに得られた前記品質データとを入力として、前記複数のパラメータの値の最適な組み合せを推定する、請求項11に記載の学習装置。 - 前記学習用データは、前記細胞の培養条件を示す条件データをさらに含み、
前記受付部が受け付ける前記学習用データには、第1培養条件で前記細胞を培養したときに得られる第1学習用データと、第2培養条件で前記細胞を培養したときに得られる第2学習用データとが含まれ、
前記モデル生成部は、前記第1学習用データおよび前記第2学習用データを用いて前記学習処理を実行することで前記予測モデルを生成する、請求項11に記載の学習装置。 - 培養容器内に細胞および培地を入れて当該細胞を培養するステップと、
前記細胞を前記培地に播種してから所定期間経過した後の少なくとも一のタイミングで前記培養容器内の少なくとも一の物質を測定するステップと、
前記測定するステップにおいて得られた測定結果を含む測定データを、前記細胞を培養することで製造されるバイオ医薬品の原薬の品質を予測するための予測モデルに入力することで、前記原薬の品質を示す品質予測データを生成するステップと、
前記品質予測データを出力するステップとを含む、推定方法。 - 培養容器内に細胞および培地を入れて当該細胞を培養するステップと、
前記細胞を前記培地に播種してからの複数のタイミングで前記培養容器内の少なくとも一の物質を測定するステップと、
前記細胞を培養することで製造されるバイオ医薬品の原薬の品質を解析するステップと、
前記測定するステップにより得られた測定結果を含む測定データと、前記解析するステップにより得られた品質データとを含む学習用データを用いて学習処理を実行することにより、前記細胞を培養している途中における培養容器内の前記少なくとも一の物質を測定して得られる測定結果に基づいて、当該培養している途中における培養容器に含まれる前記細胞から製造される前記バイオ医薬品の前記原薬の品質を示す品質予測データを生成するための予測モデルを生成するステップとを含む、学習方法。 - 細胞を培地に播種するときの培養条件を定義する複数のパラメータの値と前記細胞を培養することで製造されるバイオ医薬品の原薬を解析して得られる品質データとを受け付ける受付部と、
前記受付部が受け付けた前記複数のパラメータの値と前記品質データとを入力として、最適な前記複数のパラメータの値の組み合せを推定する推定部と、
前記推定部が推定した前記複数のパラメータの値の組み合せを出力する出力部とを備える、最適化装置。 - 培養容器内に細胞および培地を入れて複数のパラメータの値により定義される培養条件の下で当該細胞を培養するステップと、
前記細胞を培養することで製造されるバイオ医薬品の原薬の品質を解析するステップと、
前記複数のパラメータの値と前記解析するステップにより得られた品質データとを入力として、前記複数のパラメータの値の最適な組み合せを推定するステップとを含み、
前記細胞を培養するステップは、前記推定するステップにより推定された前記複数のパラメータの値の最適な組み合せを新たな前記培養条件として、前記細胞を培養する、最適化方法。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202280013193.7A CN116830204A (zh) | 2021-02-08 | 2022-01-31 | 估计装置、学习装置、优化装置、估计方法、学习方法以及优化方法 |
US18/275,969 US20240120033A1 (en) | 2021-02-08 | 2022-01-31 | Estimation Device, Learning Device, Optimization Device, Estimation Method, Learning Method, and Optimization Method |
JP2022579518A JPWO2022168774A5 (ja) | 2022-01-31 | 推定装置 | |
EP22749646.0A EP4290523A1 (en) | 2021-02-08 | 2022-01-31 | Estimation device, learning device, optimization device, estimation method, learning method, and optimization method |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021-017968 | 2021-02-08 | ||
JP2021017968 | 2021-02-08 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022168774A1 true WO2022168774A1 (ja) | 2022-08-11 |
Family
ID=82741392
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/003522 WO2022168774A1 (ja) | 2021-02-08 | 2022-01-31 | 推定装置、学習装置、最適化装置、推定方法、学習方法、および最適化方法 |
Country Status (4)
Country | Link |
---|---|
US (1) | US20240120033A1 (ja) |
EP (1) | EP4290523A1 (ja) |
CN (1) | CN116830204A (ja) |
WO (1) | WO2022168774A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024203378A1 (ja) * | 2023-03-31 | 2024-10-03 | ソニーグループ株式会社 | 細胞処理システム、細胞処理方法および学習装置 |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009044974A (ja) * | 2007-08-16 | 2009-03-05 | Univ Nagoya | 細胞の品質を予測する予測モデルの構築法、予測モデルの構築用ブログラム、該プログラムを記録した記録媒体、予測モデルの構築用装置 |
JP4496086B2 (ja) | 2002-12-23 | 2010-07-07 | ブリストル−マイヤーズ スクイブ カンパニー | タンパク質製造のための哺乳類細胞培養方法を用いる生成物品質の増大 |
JP2019110767A (ja) * | 2017-12-20 | 2019-07-11 | 株式会社日立製作所 | 細胞培養装置の制御方法及び制御システム |
US20190272894A1 (en) * | 2018-03-02 | 2019-09-05 | Genzyme Corporation | Multivariate Spectral Analysis and Monitoring for Biomanufacturing |
WO2020021860A1 (ja) * | 2018-07-24 | 2020-01-30 | 富士フイルム株式会社 | 細胞培養支援装置の作動プログラム、細胞培養支援装置、細胞培養支援装置の作動方法 |
WO2020039683A1 (ja) * | 2018-08-22 | 2020-02-27 | 富士フイルム株式会社 | 細胞培養支援装置の作動プログラム、細胞培養支援装置、細胞培養支援装置の作動方法 |
-
2022
- 2022-01-31 WO PCT/JP2022/003522 patent/WO2022168774A1/ja active Application Filing
- 2022-01-31 EP EP22749646.0A patent/EP4290523A1/en active Pending
- 2022-01-31 US US18/275,969 patent/US20240120033A1/en active Pending
- 2022-01-31 CN CN202280013193.7A patent/CN116830204A/zh active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP4496086B2 (ja) | 2002-12-23 | 2010-07-07 | ブリストル−マイヤーズ スクイブ カンパニー | タンパク質製造のための哺乳類細胞培養方法を用いる生成物品質の増大 |
JP2009044974A (ja) * | 2007-08-16 | 2009-03-05 | Univ Nagoya | 細胞の品質を予測する予測モデルの構築法、予測モデルの構築用ブログラム、該プログラムを記録した記録媒体、予測モデルの構築用装置 |
JP2019110767A (ja) * | 2017-12-20 | 2019-07-11 | 株式会社日立製作所 | 細胞培養装置の制御方法及び制御システム |
US20190272894A1 (en) * | 2018-03-02 | 2019-09-05 | Genzyme Corporation | Multivariate Spectral Analysis and Monitoring for Biomanufacturing |
WO2020021860A1 (ja) * | 2018-07-24 | 2020-01-30 | 富士フイルム株式会社 | 細胞培養支援装置の作動プログラム、細胞培養支援装置、細胞培養支援装置の作動方法 |
WO2020039683A1 (ja) * | 2018-08-22 | 2020-02-27 | 富士フイルム株式会社 | 細胞培養支援装置の作動プログラム、細胞培養支援装置、細胞培養支援装置の作動方法 |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2024203378A1 (ja) * | 2023-03-31 | 2024-10-03 | ソニーグループ株式会社 | 細胞処理システム、細胞処理方法および学習装置 |
Also Published As
Publication number | Publication date |
---|---|
US20240120033A1 (en) | 2024-04-11 |
EP4290523A1 (en) | 2023-12-13 |
CN116830204A (zh) | 2023-09-29 |
JPWO2022168774A1 (ja) | 2022-08-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP7092879B2 (ja) | 細胞培養物の代謝状態の予測 | |
JP2022533003A (ja) | バイオ医薬品生産における細胞株選択のためのデータ駆動予測モデリング | |
US20200202051A1 (en) | Method for Predicting Outcome of an Modelling of a Process in a Bioreactor | |
Hernández Rodríguez et al. | Predicting industrial‐scale cell culture seed trains–A Bayesian framework for model fitting and parameter estimation, dealing with uncertainty in measurements and model parameters, applied to a nonlinear kinetic cell culture model, using an MCMC method | |
Wang et al. | Coupled metabolic‐hydrodynamic modeling enabling rational scale‐up of industrial bioprocesses | |
WO2022168774A1 (ja) | 推定装置、学習装置、最適化装置、推定方法、学習方法、および最適化方法 | |
EP3014268B1 (en) | A method of determining or predicting a characteristic of a cell | |
Rantanen et al. | An analytic and systematic framework for estimating metabolic flux ratios from 13 C tracer experiments | |
US20220146987A1 (en) | Multivariate process chart to control a process to produce a chemical, pharmaceutical, biopharmaceutical and/or biological product | |
Leduc et al. | Massively parallel sample preparation for multiplexed single-cell proteomics using nPOP | |
CN113838530A (zh) | 利用细胞代谢网络监测生物制造过程的方法 | |
JP7207643B2 (ja) | 連続培養条件のスクリーニング方法 | |
Goudar et al. | Towards industrial application of quasi real-time metabolic flux analysis for mammalian cell culture | |
US10591463B2 (en) | Method of predicting phenotypic instability in a cell | |
Bickle | High-content screening: a new primary screening tool? | |
Blöbaum et al. | Quantifying microbial robustness in dynamic environments using microfluidic single-cell cultivation | |
US20210257045A1 (en) | Method for verifying cultivation device performance | |
EP3514797B1 (en) | Method and system for analyzing metabolic state of a cell by measuring concentrations of metabolites | |
East et al. | A Comprehensive Examination of Metabolic Flux Analysis Techniques and Applications in Systems Biology | |
Gulati et al. | Profiling cell identity and tissue architecture with single-cell and spatial transcriptomics | |
WO2022239535A1 (ja) | 抗体生成支援方法、および抗体生成支援用プログラム | |
WO2024095564A1 (ja) | 代謝フラックス分布の推定方法、解析装置、プログラム | |
US20240136010A1 (en) | Method for creating cell mathematical model, cell mathematical model creation program, cell mathematical model creation apparatus, method for determining cell mathematical model, cell mathematical model determination program, and cell mathematical model determination apparatus | |
US20240086596A1 (en) | Method and system for predicting the performance of biopharmaceutical manufacturing processes | |
US20230002708A1 (en) | Adapting control of a cell culture in a production scale vessel with regard to a starting medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22749646 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2022579518 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280013193.7 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18275969 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2022749646 Country of ref document: EP Effective date: 20230908 |