Culture support device, observation device, and program
The culture support device addresses the challenge of evaluating and predicting cell culture states by using information acquisition and prediction units to enhance the accuracy and efficiency of cell culture processes.
Patent Information
- Application Number
- JP2025166248
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2018-02-22
- Filing Date
- 2025-10-02
- Publication Date
- 2026-01-14
AI Technical Summary
Existing technologies lack an efficient and accurate method for evaluating and predicting the culture state of cells, particularly in regenerative medicine, to determine successful differentiation and potential cancerous or infected states.
A culture support device with a first and second information acquisition unit, a state prediction unit, and a culture protocol selection unit that utilizes image processing and learning algorithms to predict cell culture states and select optimal protocols based on acquired information.
Enables accurate evaluation and prediction of cell culture states, allowing for precise control of cell culture processes and improving the efficiency and quality of cell culture outcomes.
Smart Images

Figure 2026004469000001 
Figure 2026004469000002 
Figure 2026004469000003
Abstract
Description
[Technical Field]
[0001] The present invention relates to a culture support device. This application claims priority based on Japanese Patent Application No. 2018-029336, filed on February 22, 2018, the contents of which are incorporated herein by reference. [Background technology]
[0002] In general, technology for evaluating the culture state of cells has become a fundamental technology in a wide range of fields, including advanced medical fields such as regenerative medicine and pharmaceutical screening. For example, in the field of regenerative medicine, there is a process of growing and differentiating cells in vitro. Furthermore, in the above process, it is necessary to accurately evaluate the culture state of cells, such as whether or not the cells have differentiated successfully, and whether or not the cells have become cancerous or infected. As an example, a method for determining the culture state of cells by image processing of captured images of the cells has been disclosed (see Patent Document 1). [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2004-229619 Summary of the Invention
[0004] One aspect of the present invention is a culture support device comprising: a first information acquisition unit that acquires first information regarding a first culture state of the cells at a first time point in a cell culture process; a second information acquisition unit that acquires second information regarding a second culture state of the cells at a second time point after the first time point; a state prediction unit that repeatedly performs a prediction process of the culture state after a predetermined period based on a culture protocol indicating the culture conditions of the cells during the culture process and a change in the culture state of the cells due to the culture protocol, information about the culture state, and the culture protocol until the first culture state is reached to the second culture state; and a culture protocol selection unit that selects the culture protocol used in the series of prediction processes performed by the state prediction unit as the culture protocol that will bring the cells from the first culture state to the second culture state, wherein the state prediction unit performs the prediction process based on the information about the culture state output by the immediately preceding prediction process, the culture protocol, and the learning result if the culture state output by the prediction process has not reached the second culture state. [Brief explanation of the drawings]
[0005] [Figure 1] 1 is a block diagram showing an overview of an incubator including a culture support device according to an embodiment of the present invention. [Figure 2] FIG. 1 is a diagram showing an example of a front view of an incubator according to an embodiment of the present invention. [Figure 3] FIG. 1 is a diagram showing an example of a plan view of an incubator according to an embodiment of the present invention. [Figure 4] FIG. 10 is a diagram illustrating an example of an observation operation in the incubator according to the embodiment of the present invention. [Figure 5] FIG. 2 is a diagram illustrating an example of a functional configuration of a control device according to an embodiment of the present invention. [Figure 6] FIG. 10 is a diagram illustrating an example of a learning stage process according to an embodiment of the present invention. [Figure 7] 10 is an example of information stored in a storage unit according to an embodiment of the present invention. [Figure 8] FIG. 10 is a diagram illustrating an example of event information to be implemented according to the embodiment of the present invention. [Figure 9] FIG. 10 is a diagram illustrating an example of processing in an application stage according to an embodiment of the present invention. [Figure 10] FIG. 4 is a diagram illustrating an example of an image stored in an image storage unit according to the embodiment of the present invention. [Figure 11] FIG. 10 is a diagram illustrating an example of an initial information input screen according to the embodiment of the present invention. [Figure 12] FIG. 10 is a diagram illustrating an example of a termination information input screen according to the embodiment of the present invention. [Figure 13] FIG. 10 is a diagram illustrating an example of a process for culturing process analysis according to an embodiment of the present invention. [Figure 14] FIG. 10 is a diagram illustrating an example of a process for predicting image feature amounts of a cell image in a final state according to an embodiment of the present invention. [Figure 15] FIG. 10 is a diagram showing an example of a culture mode selection screen according to an embodiment of the present invention. [Figure 16] FIG. 10 is a diagram illustrating an example of an analysis result screen according to the embodiment of the present invention. [Figure 17] FIG. 10 is a diagram illustrating an example of an event selection screen according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION
[0006] [Embodiment] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing an outline of an incubator 11 including a culture support device of the embodiment. Fig. 2 and Fig. 3 are diagrams showing an example of a front view and a plan view of the incubator 11 of the embodiment. The incubator 11 is an example of an observation device.
[0007] The incubator 11 of this embodiment has an upper casing 12 and a lower casing 13. When the incubator 11 is assembled, the upper casing 12 is placed on the lower casing 13. The internal spaces of the upper casing 12 and the lower casing 13 are divided into upper and lower spaces by a base plate 14.
[0008] First, an outline of the configuration of the upper casing 12 will be described. Inside the upper casing 12, a temperature-controlled chamber 15 in which cells are cultured is formed. This temperature-controlled chamber 15 has a temperature regulator 15a and a humidity regulator 15b, and the inside of the temperature-controlled chamber 15 is maintained in an environment suitable for cell culture (for example, an atmosphere with a temperature of 37°C and a humidity of 90%) (note that the temperature regulator 15a and humidity regulator 15b are not shown in Figures 2 and 3). In other words, the temperature-controlled chamber 15 can maintain its interior at predetermined environmental conditions.
[0009] A large door 16, a middle door 17, and a small door 18 are arranged on the front of the temperature-controlled chamber 15. The large door 16 covers the front of the upper casing 12 and the lower casing 13. The middle door 17 covers the front of the upper casing 12 and isolates the temperature-controlled chamber 15 from the outside environment when the large door 16 is open. The small door 18 is a door for loading and unloading a culture vessel 19 for culturing cells, and is attached to the middle door 17. Loading and unloading the culture vessel 19 through this small door 18 makes it possible to suppress environmental changes in the temperature-controlled chamber 15. The large door 16, middle door 17, and small door 18 are kept airtight by packings SL1, SL2, and SL3, respectively.
[0010] For example, these doors are opened when performing tasks such as "medium change," "passaging," "cleaning," and "transferring to another incubator." The type of these tasks is selected depending on the state of cell culture. In the following explanation, these tasks are also collectively referred to as "events." Note that "events" also include tasks that are performed without opening the doors, doing nothing, and stopping cell culture. "Culture exchange" refers to replacing all or part of the culture medium in a culture vessel with new culture medium. "Subculture" refers to harvesting proliferated cells and reseeding them in another culture vessel. "Cleaning" refers to cleaning the inside of an incubator. Cleaning includes disinfecting the incubator surface and sterilizing the water necessary to maintain humidity inside the incubator. "Transferring an incubator" refers to transferring a culture vessel to a different incubator when there are multiple incubators.
[0011] Also arranged in the temperature-controlled room 15 are a stocker 21, an observation unit 22, a container transport device 23, and a transport table 24. The transport table 24 is arranged in front of the small door 18, and the culture containers 19 are transported in and out through the small door 18.
[0012] The stocker 21 is disposed on the left side of the temperature-controlled room 15 when viewed from the front of the upper casing 12 (the lower side in FIG. 3). The stocker 21 has multiple shelves, and each shelf of the stocker 21 can store multiple culture vessels 19. Each culture vessel 19 contains cells to be cultured together with a medium. In this way, the temperature-controlled room 15 stores culture vessels for culturing cells.
[0013] The observation unit 22 is disposed on the right side of the temperature-controlled chamber 15 when viewed from the front of the upper casing 12. This observation unit 22 is capable of performing time-lapse observation of cells in the culture vessel 19. Here, time-lapse observation refers to a technique for observing changes in a sample over time by capturing images of the sample at predetermined time intervals based on a preset imaging schedule. Images of the sample may be captured at regular time intervals or at different time intervals.
[0014] Here, the observation unit 22 is fitted into the opening of the base plate 14 of the upper casing 12. The observation unit 22 has a sample stage 31, a stand arm 32 extending above the sample stage 31, and a main body 33 incorporating a microscopic optical system for phase contrast observation and an imaging device 34. The sample stage 31 and the stand arm 32 are placed in the temperature-controlled chamber 15, while the main body 33 is housed in the lower casing 13.
[0015] The sample stage 31 is made of a light-transmitting material, and the culture vessel 19 can be placed on it. The sample stage 31 is configured to be movable horizontally, allowing the position of the culture vessel 19 placed on its top surface to be adjusted. An LED light source 35 is built into the stand arm 32. The imaging device 34 can capture microscopic images of the cells in the culture vessel 19, which are transmitted and illuminated from above the sample stage 31 by the stand arm 32, via a micro-optical system. The imaging device 34 images the cells accommodated in the culture vessel in the temperature-controlled chamber 15 at predetermined time intervals.
[0016] The container transport device 23 is disposed in the center of the temperature-controlled room 15 when viewed from the front of the upper casing 12. The container transport device 23 transfers the culture container 19 between the stocker 21, the sample stage 31 of the observation unit 22, and the transport stage 24.
[0017] 3, the container transfer device 23 has a vertical robot 38 with an articulated arm, a rotation stage 39, a mini-stage 36, and an arm unit 37. The rotation stage 39 is attached to the tip of the vertical robot 38 via a rotation shaft 35a so as to be rotatable 180 degrees in the horizontal direction. Therefore, the rotation stage 39 can position the arm unit 37 facing the stocker 21, the sample stage 31, and the transfer table 24, respectively.
[0018] The mini-stage 36 is attached so as to be slidable in the horizontal direction relative to the rotary stage 39. An arm part 37 for gripping the culture vessel 19 is attached to the mini-stage 36.
[0019] Next, we will explain the outline of the configuration of the lower casing 13. Inside the lower casing 13, a main body 33 of the observation unit 22 and a control device 41 of the incubator 11 are housed.
[0020] The control device 41 is connected to the temperature adjustment device 15a, the humidity adjustment device 15b, the observation unit 22, and the container transport device 23. The control device 41 includes a calculation unit 42 and a storage unit 43, and controls all the components of the incubator 11 in accordance with a predetermined program. The control device 41 is an example of a culture support device.
[0021] As an example, the control device 41 controls the temperature adjustment device 15a and the humidity adjustment device 15b to maintain predetermined environmental conditions inside the temperature-controlled room 15. Furthermore, the control device 41 controls the observation unit 22 and the container transport device 23 based on a predetermined observation schedule to automatically execute an observation sequence for the culture container 19. Furthermore, the control device 41 executes a culture state evaluation process to evaluate the culture state of the cells based on images acquired in the observation sequence.
[0022] [Example of observation behavior] Next, an example of the observation operation in the incubator 11 will be described with reference to the flowchart of FIG. 4 is a diagram showing an example of an observation operation in the incubator 11 of this embodiment. The figure shows an example of an operation in which a culture vessel 19 carried into a temperature-controlled room 15 is subjected to time-lapse observation according to a registered observation schedule.
[0023] Step S101: The calculation unit 42 compares the observation schedule in the management data of the storage unit 43 with the current date and time to determine whether the observation start time of the incubation container 19 has arrived. If the observation start time has arrived (step S101: YES), the calculation unit 42 shifts the process to S102. On the other hand, if it is not time to observe the incubation container 19 (step S101: NO), the calculation unit 42 waits until the next scheduled observation time.
[0024] Step S102: The calculation unit 42 instructs the container transport device 23 to transport the culture container 19 corresponding to the observation schedule. Then, the container transport device 23 takes the instructed culture container 19 out of the stocker 21 and places it on the specimen stage 31 of the observation unit 22. When the culture container 19 is placed on the specimen stage 31, a bird's-eye view camera (not shown) built into the stand arm 32 captures an overall observation image of the culture container 19.
[0025] Step S103: The calculation unit 42 instructs the observation unit 22 to capture a microscopic image of the cells. The observation unit 22 turns on the LED light source 35 to illuminate the culture vessel 19, and also drives the imaging device 34 to capture a microscopic image of the cells in the culture vessel 19.
[0026] At this time, the imaging device 34 captures a microscope image based on the imaging conditions (magnification of the objective lens, observation point in the container) specified by the user, based on the management data stored in the memory unit 43. For example, when observing multiple points in the culture container 19, the observation unit 22 sequentially adjusts the position of the culture container 19 by driving the sample stage 31, and captures a microscope image at each point. The microscope image data acquired in S103 is read into the control device 41 and recorded in the memory unit 43 under the control of the calculation unit 42.
[0027] Step S104: After the observation schedule ends, the calculation unit 42 instructs the container transfer device 23 to transfer the incubation container 19. Then, the container transfer device 23 transfers the instructed incubation container 19 from the specimen stage 31 of the observation unit 22 to a predetermined storage position in the stocker 21. Thereafter, the calculation unit 42 ends the observation sequence and returns the process to S101.
[0028] Through the above-described procedure, the time-series image data observed by the incubator 11 is stored in the storage unit 43. In the following description, obtaining time-series image data by the incubator 11 is also referred to as time-lapse imaging.
[0029] [Presentation of culture protocols and final state predictions according to culture mode] Next, we will explain the function of presenting the culture protocol and final state prediction according to the culture mode with reference to Figures 5 to 16. Here, the culture protocol is information indicating the sequence of events to be performed and environmental conditions to be set. The final state is the state of the cells obtained at the end of culture.
[0030] 5 is a diagram showing an example of the functional configuration of the control device 41 of this embodiment. As described above, the control device 41 includes the calculation unit 42 and the storage unit 43. The calculation unit 42 has, as its functional units, an image acquisition unit 421, a display control unit 422, an operation detection unit 423, a memory control unit 424, a learning unit 425, an analysis information generation unit 426, a culture mode selection unit 427, a parameter acquisition unit 4285, and an analysis result presentation unit 428. The storage unit 43 includes an image storage unit 431 , an implemented event storage unit 432 , an environment log storage unit 433 , a learning result storage unit 434 , a culture mode storage unit 435 , and a prediction result storage unit 436 .
[0031] The image acquisition unit 421 acquires an image P from the imaging device 34. This image P is an image of cells being cultured captured, for example, at predetermined time intervals, based on an imaging schedule. When the image acquisition unit 421 acquires the image P from the imaging device 34, the image acquisition unit 421 adds imaging date and time information DT indicating the date and time of imaging, and stores the image in the image storage unit 431.
[0032] The display control unit 422 controls the screen display of the display unit 44. Specifically, the display control unit 422 displays an image P stored in the image storage unit 431 and candidate events related to this image P on the display unit 44. The display control unit 422 displays an initial information input screen DI for inputting initial information IS, a termination information input screen DF for inputting termination information FS, a culture mode selection screen DM for selecting a culture mode, an analysis result screen DO showing the analysis results of the culture process, and an event selection screen DE for selecting an event to be implemented.
[0033] The operation unit 45 includes a touch panel, a mouse, a keyboard, or the like. When the operation unit 45 is a touch panel, the operation unit 45 and the display unit 44 may be configured as an integrated unit. Alternatively, the operation unit 45 and the display unit 44 may be configured as a touch panel provided on the upper casing 12 or the lower casing 13. The observer selects an event to be performed by operating this operation unit 45. In the example described above, when a culture medium replacement operation is performed, the observer operates the operation unit 45 to select "culture medium replacement" as the event to be performed. The observer also operates the operation unit 45 to set the environmental conditions of the temperature-controlled room 15 . In the above example, the temperature and humidity of the temperature-controlled room 15 are set as the environmental conditions of the temperature-controlled room 15, but this is not limiting. The temperature-controlled room 15 may have a carbon dioxide concentration adjusting device that adjusts the carbon dioxide concentration, and this carbon dioxide concentration adjusting device may be controlled by the control device 41.
[0034] The operation detection unit 423 detects an operation on the operation unit 45. When the operation detection unit 423 detects an operation, it generates performed event information EV according to this operation. When the observer performs an operation on the operation unit 45 to select "culture medium change", the operation detection unit 423 detects that "culture medium change" has been selected as the performed event. Furthermore, the operation detection unit 423 generates performed event information EV indicating "culture medium change." Furthermore, when the operation detection unit 423 detects an operation to set the environmental conditions of the temperature-controlled room 15, it generates environmental log information EL according to this operation. When the observer operates the operation unit 45 to set the environmental conditions of the temperature-controlled room 15 to an atmosphere with a temperature of 37°C and a humidity of 90%, the operation detection unit 423 detects that an atmosphere with a temperature of 37°C and a humidity of 90% has been set as the environmental conditions.
[0035] The memory control unit 424 controls writing of information to the memory unit 43. Specifically, the memory control unit 424 associates the image acquired by the image acquisition unit 421 with executed event information EV indicating an event related to cell culturing at the time the image was captured, and stores the associated information in the executed event memory unit 432. The memory control unit 424 associates the image acquired by the image acquisition unit 421 with environmental log information EL indicating the environmental conditions of the temperature-controlled chamber 15 at the time the image was captured, and stores the associated information in the environmental log memory unit 433.
[0036] Furthermore, the storage control unit 424 may store information about the date and time when the event was held in addition to the held event information EV. In this case, the held event information EV includes information about the date and time when the event was held. Furthermore, the storage control unit 424 may add information about the date and time when the environmental log was acquired to the environmental log information EL and store the added information. In this case, the environmental log information EL includes information about the date and time when the environmental log was acquired.
[0037] The learning unit 425 learns the association between an image stored in the image storage unit 431, the implemented event information EV associated with this image, and the environment log information EL associated with this image. The learning unit 425 learns the association between these pieces of information using various known methods. The learning unit 425 stores the learning result in the learning result storage unit 434 as a learning result LR.
[0038] The analysis information generation unit 426 generates each piece of information used for analysis by the analysis result presentation unit 428. The analysis information generation unit 426 includes an initial information generation unit 4261, a current information generation unit 4262, an end information generation unit 4264, and an intermediate information generation unit 4263.
[0039] The initial information generation unit 4261 acquires images P that are acquired from the start of the culture process to the first culture medium change from among the images P stored in the image storage unit 431. The initial information generation unit 4261 acquires performed event information EV for performed events that are detected from the start of the culture process to the first culture medium change from among the performed event information EV stored in the performed event storage unit 432. The initial information generation unit 4261 acquires environment log information EL that is detected from the start of the culture process to the first culture medium change from among the environment log information EL stored in the environment log storage unit 433. Here, the start of the culture process refers to, for example, immediately after seeding of cells in the culture process. The initial information generation unit 4261 supplies the acquired image P, the held event information EV, and the environment log information EL to the analysis result presentation unit 428 as initial information IS.
[0040] The current information generation unit 4262 acquires images P acquired from immediately after sowing to the present from among the images P stored in the image storage unit 431. The current information generation unit 4262 acquires implemented event information EV for implemented events detected from immediately after sowing to the present from among the implemented event information EV stored in the implemented event storage unit 432. The current information generation unit 4262 acquires environmental log information EL detected from immediately after sowing to the present from among the environmental log information EL stored in the environmental log storage unit 433. The current information generation unit 4262 supplies the acquired image P, the held event information EV, and the environment log information EL to the analysis result presentation unit 428 as current information PS.
[0041] The intermediate information generating unit 4263 acquires the intermediate information MS from the operation detecting unit 423, and supplies it to the analysis result presenting unit 428. The intermediate information MS is input from the operation unit 45. The intermediate information generating unit 4263 does not necessarily have to be provided.
[0042] The end information generating unit 4264 acquires the end information FS from the operation detecting unit 423, and supplies it to the analysis result presenting unit 428. The end information FS is input from the operation unit 45.
[0043] The end-stage information FS is selected based on the purpose or goal of the culture, such as expansion, drug efficacy, transplantation, biomaterial production, cell experiments, or culture process improvement. Examples of end-stage information FS include the amount of cells, the name of the cell or organ, sample images of the cells to be cultured, the name and literature of the differentiation induction method, the desired drug, procedure, and conditions, the name of the transplant procedure for the cell transplant, the name and condition of the patient to be transplanted, or their age, gender, height, weight, or medical record, the name and ingredients of the drug to be tested for efficacy and pharmacology, the name of the disease to be elucidated, the biomaterials and proteins to be obtained through culture, and the type of cell experiment in which the cells obtained through culture will be used. The amount of cells is specified based on the target cell growth rate, area, and number of post-culture passages. Of the end information FS, a sample image of cells to be obtained by culturing is called an end image PF. Of the end information FS, the rest is called non-image end information CF.
[0044] The culture mode selection unit 427 selects the culture mode CM. The culture mode CM is information indicating a set of weights W that indicates which of the setting values V used by the analysis result presentation unit 428 when analyzing the culture protocol should be prioritized over other setting values. Therefore, the culture mode CM is information indicating a predetermined set of weights W. Furthermore, the weights W are used to select a culture protocol that indicates the conditions for cell culture.
[0045] Here, the set value V is, for example, past culture results, predicted cell amount, predicted type of cultured cells, prediction accuracy (guaranteed accuracy), culture costs such as amount and period calculated from the culture set value, consumption of medium and equipment calculated from the culture set value, process and degree of cell growth, culture operation time, etc. The set value V may be a value that specifies the upper and lower limits of these amounts. The weight W is a weight associated with each setting value V. Note that the weight W may be associated with a plurality of setting values V.
[0046] Examples of the culture mode CM include a quality-oriented mode, a production volume-oriented mode, and a speed mode. The quality-oriented mode is a culture mode that prioritizes prediction accuracy among the set values V over other set values of the set value V. The production volume-oriented mode is a culture mode that prioritizes the predicted cell amount among the set values V over other set values of the set value V. The speed mode is a culture mode that prioritizes culture cost among the set values V over other set values of the set value V.
[0047] The culture mode selection unit 427 selects a culture mode CM in response to an operation from the operation unit 45. Here, the culture mode CM is stored in advance in the culture mode storage unit 435. That is, the culture mode selection unit 427 selects a culture mode CM from one or more culture modes CM stored in advance in the culture mode storage unit 435.
[0048] In the above example, the weights W acquired by the culture mode selection unit 427 are determined by the culture mode CM selected by the culture mode selection unit 427. However, the present invention is not limited to this. The culture mode selection unit 427 may acquire weights W input from the operation unit 45. When the weights W are input from the operation unit 45, the observer inputs each of the weights W from the operation unit 45.
[0049] The culture mode CM may also be information indicating a set of one or more set values V instead of a set of weights W. The culture mode CM may also be information indicating a set of weights W and a set of one or more set values V. The quality-oriented mode may, for example, include prediction accuracy as a predetermined value. The production volume-oriented mode may, for example, include the predicted amount of cells as a predetermined value. The speed mode may, for example, include the length of the culture period as a predetermined value.
[0050] The analysis result presentation unit 428 presents a culture protocol indicating the conditions for cell culture. The analysis result presentation unit 428 includes an initial information acquisition unit 4281, a current information acquisition unit 4282, an intermediate information acquisition unit 4283, an end information acquisition unit 4284, a parameter acquisition unit 4285, a feature extraction unit 4286, a culture protocol presentation unit 4287, and a state prediction unit 4288.
[0051] The initial information acquisition unit 4281 acquires the initial information IS supplied by the initial information generation unit 4261. Here, the initial information IS is information relating to the culture state of the cells from the start of the culture process to the first medium change. That is, the initial information generation unit 4261 acquires the initial information IS relating to the culture state of the cells from the start of the culture process to the first medium change.
[0052] The current information acquisition unit 4282 acquires the current information PS supplied by the current information generation unit 4262. Here, the current information PS is information relating to the culture state of the cells from immediately after seeding in the cell culture process to the present. In other words, the current information acquisition unit 4282 acquires the current information PS relating to the culture state of the cells from the start of the culture process to the present.
[0053] The intermediate information acquisition unit 4283 acquires the intermediate information MS supplied by the intermediate information generation unit 4263. Here, the intermediate information MS is information relating to the culture state of the cells during the period between the start and end of the culture process. That is, the intermediate information acquisition unit 4283 acquires the intermediate information MS relating to the culture state of the cells during the period between the start and end of the culture process.
[0054] The termination information acquisition unit 4284 acquires termination information FS supplied by the termination information generation unit 4264. Here, the termination information FS is information relating to the culture state of the cells at the end of the cell culture process. That is, the termination information generation unit 4264 acquires termination information FS relating to the culture state of the cells at the end of the culture process after the period from the start of the culture process to the first medium change.
[0055] The parameter acquisition unit 4285 acquires parameters PM used to select a culture protocol indicating the conditions for cell culture. Here, the parameters PM are a weight W indicated by the culture mode CM selected by the culture mode selection unit 427 and a set value V supplied from the operation detection unit 423. The set value V is input by the observer via the operation unit 45.
[0056] The feature extraction unit 4286 extracts image feature values FP from the image P included in the acquired initial information IS using a known machine learning technique. The feature extraction unit 4286 extracts image feature values FP from the image P included in the acquired current information PS using a known machine learning technique.
[0057] Here, the image feature FP includes, for example, the growth state (single, colony-forming, sheet-like), number of cells, cell size, characteristics of the cell's external shape (spherical, bubbly, jagged), intercellular adhesion, intercellular connectivity, number of nuclei, shape of nuclei, size of nuclei, number of nucleoli, size of nucleoli, concentration of nucleoli, amount of cytoplasmic granules, brightness of cytoplasmic granules, size of vacuoles, number of vacuoles, shape of cytoplasmic protrusions (number, length, thickness), occupied area, density, etc. The image feature FP also includes features regarding time changes in the image P.
[0058] The feature extraction unit 4286 extracts non-image feature values F from the completed event information EV and environment log information EL included in the acquired initial information IS. The feature extraction unit 4286 extracts non-image feature values F from the completed event information EV and environment log information EL included in the acquired current information PS.
[0059] The non-image feature F is a feature other than the image feature FP from immediately after seeding to the first culture medium change. The non-image feature F is, for example, the passage history (culture history of undifferentiated cells), coating agent, amount and type of culture medium, type of cell line, name of culture protocol, seeding method, freezing method, reference papers, etc. The non-image feature F may also include culturist information indicating the registered culturist ID and the culturist's years of culturing experience. The culturist information may be input from the operation unit 45.
[0060] The feature extraction unit 4286 supplies the extracted image feature FP and non-image feature F to the culture protocol presentation unit 4287 and the state prediction unit 4288. The culture protocol presenting unit 4287 calculates a culture protocol based on the image feature FP and non-image feature F, the termination information FS, the weight W, the setting value V, and the learning result LR. Here, the culture protocol presenting unit 4287 acquires the learning result LR from the learning result storage unit 434. The state prediction unit 4288 predicts the final state of the culture based on the image feature amount FP, the non-image feature amount F, the termination information FS, the weight W, the setting value V, and the learning result LR. Here, the state prediction unit 4288 acquires the learning result LR from the learning result storage unit 434.
[0061] The analysis result presenting unit 428 supplies the culture protocol EVC indicating the culture protocol calculated by the culture protocol presenting unit 4287 and the final state predicted PFS indicating the final state of the culture predicted by the state predicting unit 4288 to the display control unit 422, and causes them to be displayed on the display unit 44. In addition, the analysis result presenting unit 428 causes the learning result storage unit 434 to store the final state predicted PFS indicating the final state predicted by the state predicting unit 4288.
[0062] In other words, the analysis result presentation unit 428 presents the culture protocol based on the parameters acquired by the parameter acquisition unit 4285, the initial information IS acquired by the initial information generation unit 4261, the termination information FS acquired by the termination information generation unit 4264, and the learning results obtained by learning the correlation between the cell culture results according to the culture protocol and the culture protocol.
[0063] The following describes the processing performed by the control device 41. The processing performed by the control device 41 includes a learning stage processing and an application stage processing. The control device 41 performs the application stage processing based on the results of the learning stage processing.
[0064] [Learning stage] The learning stage processing performed by the control device 41 will now be described with reference to Figures 6 to 8. Figure 6 is a diagram showing an example of the learning stage processing of this embodiment.
[0065] Step S200: The learning unit 425 acquires multiple pieces of data linking time-series cell images, performed events, and environmental logs throughout the culture process from the storage unit 43. Here, the learning unit 425 acquires images P1 to Pn and the imaging date and time information DT1 to DTn added to these images P1 to Pn, respectively, from the image storage unit 431 as time-series cell images. The learning unit 425 acquires performed event information EV1 to EVn from the performed event storage unit 432. Here, a plurality of pieces of data that the learning unit 425 acquires from the storage unit 43 will be described with reference to FIG. 7 shows an example of information stored in the storage unit 43 of this embodiment. In the example shown in the figure, the storage unit 43 stores the culture progress in a culture sequence seq1. In this culture sequence seq1, images P are captured at time t1, time t2, ... time t3. As described above, these images P are associated with image capture date and time information DT and stored in chronological order in the image storage unit 431. Furthermore, each of these images P is associated with event information EV. In this example, image P1 is associated with "culture medium change," "immediate change," and "half change" as event information EV. As described above, this event information EV is stored in the event storage unit 432. Furthermore, environmental log information EL is associated with each of these images P. In this example, the temperature "37°C" and humidity "90%" are associated with image P1 as environmental log information EL. As described above, this environmental log information EL is stored in the environmental log storage unit 433.
[0066] Here we will explain the event information EV in more detail. 8 is a diagram showing an example of the event information EV of this embodiment. In the event information EV, events are categorized into hierarchical levels. In this example, the events are categorized into three hierarchical levels. As an example shown in the figure, events (layer 1) include "medium change," "passage," "cleaning," etc. Events (layer 2) for "medium change" include "perform immediately," "perform after a specified time," etc. Although not shown, information about the observer who will perform each event may also be displayed as event information.
[0067] Returning to FIG. 6, the description of the learning process of learning unit 425 will be continued. Step S201: The learning unit 425 divides the plurality of data acquired from the storage unit 43 into predetermined time periods. Here, as an example, the learning unit 425 divides the plurality of data into the types of events held indicated by the held event information EV. The learning unit 425 may divide the plurality of data into predetermined time lengths. Each of the divided plurality of data includes one or more sets of the image capture date and time, the image, the environmental log, and the held event.
[0068] Step S202: The learning unit 425 extracts the feature amounts of the cell images for each period. The learning unit 425 extracts the feature amounts of the cell images for each data segmented in step S201. When the cell image for a certain period is an image Pi (i = 1, 2, ..., N:N is the number of images included in a certain period), the learning unit 425 extracts the feature amounts fi (i = 1, 2, ..., N:N is the number of images included in a certain period) of each image Pi. When calculating the feature amounts fi, in addition to the corresponding image Pi, previous and subsequent images P-(i-1), P-(i+1), and images Pj (j = 1, 2, ..., N:N is the number of images included in a certain period) may also be used.
[0069] The learning unit 425 extracts morphological features as feature values fi using known image processing. Here, morphological features include, for example, the area, circumference, and seeding variation of the cells captured in the image Pi. The learning unit 425 may extract feature values fi using known machine learning techniques. Here, machine learning is, for example, deep learning. When deep learning is used to extract feature values fi, the learning unit 425 uses previously acquired learning data for deep learning training. For example, the value of the area of an ideal cell is used as this learning data. When deep learning is used to extract feature values fi, the learning unit 425 may extract, as feature values fi, values output from an intermediate layer of a neural network used in deep learning. The learning unit 425 sets the representative value of each component of the extracted feature quantity fi (i=1, 2, . . ., N: N is the number of images included in a certain period) as the feature quantity of the images in a certain period. Here, the representative value is the average value or the median value.
[0070] Step S203: The learning unit 425 learns a model that reproduces the feature quantities of cell images for each period by regression analysis based on the feature quantities of past cell images, the environmental log, and the implemented events. Here, the feature quantities of past cell images are feature quantities extracted from images for periods prior to each period. The model is a function that outputs a value of 1 or greater when the feature quantities of past cell images, the environmental log, and the implemented events are input. The number of values 1 or greater is equal to the number of components of the feature quantities of cell images for each period. This model includes one or more model parameters, which are values that characterize this model. The learning unit 425 adjusts the model parameters so that the value 1 or greater output by the model when the feature quantities of past cell images, the environmental log, and the implemented events are input to this model is close to the feature quantities of cell images for each period.
[0071] Before learning the model, the learning unit 425 may extract heuristic values from the environmental log for each period, and use them as the environmental log for each period before learning the model. Here, the heuristic values include the mean, variance, maximum value, minimum value, etc. Furthermore, the learning unit 425 may also use a deep learning architecture that automatically extracts features in the process of step S203. The learning unit 425 supplies the learned model as a learning result LR to the analysis result presenting unit 428. Here, if the learning unit 425 uses deep learning, the learning result LR is a result obtained by deep learning.
[0072] [Application stage] The application stage processing performed by the control device 41 will be described with reference to FIGS. Fig. 9 is a diagram showing an example of the application stage processing of this embodiment. The processing shown in Fig. 9 is performed after the learning stage processing shown in Fig. 6 is completed and the learning unit 425 supplies the learning result LR to the analysis result presenting unit 428.
[0073] Step S300: The initial information acquisition unit 4281 acquires initial information IS. The initial information IS includes images P0 to Pn acquired from immediately after seeding until the first culture medium change, as well as implemented event information EV and environmental log information EL. The initial information acquisition unit 4281 supplies the acquired initial information IS to the culture protocol presentation unit 4287 and the state prediction unit 4288.
[0074] Here, with reference to FIG. 10, the images P0 to Pn that the initial information acquisition unit 4281 acquires from the image storage unit 431 will be described. 10 is a diagram showing an example of an image P stored in the image storage unit 431 of this embodiment. In this example, the imaging device 34 captures images at time t0, time t1, ... time tn-1, time tn, and time tn+1. For example, image P0 is an image captured at time t0. When the image acquisition unit 421 acquires image P0 from the imaging device 34 at time t0, it adds imaging date and time information DT indicating time t0 to image P0 and stores it in the image storage unit 431. Similarly, for images P1 to Pn+1, the image acquisition unit 421 adds imaging date and time information DT to each of the images P and stores them in the image storage unit 431.
[0075] At time tE between time tn and time tn+1, a culture medium change is performed as an event to be performed. The initial information generating unit 4261 acquires, from the images P stored in the image storage unit 431, images P0 to Pn that were acquired from time t0 immediately after seeding to time tE when the first culture medium change is performed.
[0076] When the initial information acquisition unit 4281 acquires the image Pn+1, the display control unit 422 displays the initial information input screen DI on the display unit 44. Here, the image Pn+1 is a cell image captured immediately after seeding and after the first culture medium change is performed. The initial information input screen DI will be described with reference to FIG. 11.
[0077] 11 is a diagram showing an example of an initial information input screen DI according to this embodiment. The display unit 44 displays the initial information input screen DI. In this example, an initial image PI is displayed in an area AR1 of the initial information input screen DI. This initial image PI is image Pn+1 in FIG. 10. An area AR2 of the initial information input screen DI displays implemented event information EV and environmental log information EL. In this example, the passage history, amount and type of culture medium, type of cell line, and procedure are displayed as non-image features F.
[0078] In this example, the initial information acquisition unit 4281 acquires the initial information IS from the initial information generation unit 4261, but the present invention is not limited to this. The initial information acquisition unit 4281 may acquire the initial information IS from the operation detection unit 423. Furthermore, the initial information acquisition unit 4281 may acquire the initial information IS from both the storage unit 43 and the operation detection unit 423.
[0079] Returning to FIG. 9, the description of the application stage process continues. Step S301: The end information acquisition unit 4284 acquires end information FS supplied by the operation detection unit 423. The end information acquisition unit 4284 supplies the acquired end information FS to the analysis result presentation unit 428. Here, the end information FS includes at least one of an end image PF and non-image end information CF.
[0080] Here, the end information input screen DF will be described with reference to FIG. 12 is a diagram showing an example of the end information input screen DF of this embodiment. The end information input screen DF is displayed on the display unit 44. In this example, an end image PF is displayed in an area AR3 of the end information input screen DF. An input form for inputting non-image end information CF is displayed in an area AR4 of the end information input screen DF. The end image PF and non-image end information CF are input by the observer via the operation unit 45.
[0081] In this example, the termination information acquiring unit 4284 acquires the termination information FS from the current information generating unit 4262, but the present invention is not limited to this. The termination information acquiring unit 4284 may acquire the termination information FS from an image P stored in the storage unit 43. This image P is, for example, a cell image stored in the image storage unit 431 during the learning stage of the control device 41.
[0082] Returning to FIG. 9, the description of the application stage process continues. Step S302: The analysis result presentation unit 428 analyzes the culture process. The culture process analysis processing will now be described with reference to Fig. 13. Note that in the processing of step S302, the present refers to the time immediately after the first culture medium exchange has been carried out since immediately after seeding.
[0083] FIG. 13 is a diagram showing an example of the processing of the culture process analysis of this embodiment. Step S400: The feature extraction unit 4286 extracts an image feature FP from the current image P. Here, the feature extraction unit 4286 extracts image features FP0 to FPn from images P0 to Pn included in the initial information IS, respectively. The feature extraction unit 4286 extracts the image features FP0 to FPn using a known machine learning method. The image features FP0 to FPn extracted by the feature extraction unit 4286 may be selected according to the type of cell. Furthermore, the feature extraction unit 4286 extracts features regarding time changes in images P0 to Pn, and includes them in the image features FP0 to FPn. The feature extraction unit 4286 supplies the extracted image feature amounts FP0 to FPn to the culture protocol presentation unit 4287 and the state prediction unit 4288.
[0084] The feature extraction unit 4286 may extract morphological features from images P0 to Pn as image features FP0 to FPn using known image processing. Here, morphological features include, for example, the area, circumference, and seeding variation of the cells captured in images P0 to Pn, respectively. Instead of image features FP0 to FPn extracted using a machine learning technique, the feature extraction unit 4286 may use morphological features extracted using image processing as image features FP0 to FPn, or may combine features extracted using a machine learning technique and morphological features extracted using image processing to generate image features FP0 to FPn.
[0085] Step S401: The feature extraction unit 4286 sets a series of environmental logs and implemented events based on the environmental log information EL and implemented event information EV. Here, the feature extraction unit 4286 sets a series of environmental logs and implemented events by extracting non-image features F from the environmental logs and implemented events from immediately after sowing to the present, and supplies them to the culture protocol presentation unit 4287 and the state prediction unit 4288.
[0086] Step S402: The analysis result presentation unit 428 performs a process of predicting the image feature amount FP of the cell image in the final state obtained by culturing. Here, the process of predicting the image feature amount FP of the cell image in the final state will be described with reference to FIG.
[0087] FIG. 14 is a diagram showing an example of a process for predicting the image feature amount FP of a cell image in a final state according to this embodiment. Step S500: The state prediction unit 4288 acquires the image feature FP and the non-image feature F supplied by the feature extraction unit 4286. Here, the image feature FP supplied by the feature extraction unit 4286 is a feature extracted from the current image P. The non-image feature F supplied by the feature extraction unit 4286 is a feature extracted from a set series of environment logs and implemented events.
[0088] Step S501: The state prediction unit 4288 predicts the image feature FP of the cell image for the next period using the image feature FP extracted from the current image, the non-image feature F extracted from the given environmental log and event, and the learned model (learning result LR). The state prediction unit 4288 pairs the predicted image feature FP of the cell image with the non-image feature F to generate the prediction result PR.
[0089] Step S502: The state prediction unit 4288 predicts the image feature FP of the cell image for the next period using the non-image feature F extracted from the environmental log and events for the next period, the prediction result PR, and the learned model (learning result LR). Each time the state prediction unit 4288 performs the processing of step S502, it pairs the image feature FP of the cell image predicted in the prediction processing of step S502 with the non-image feature F extracted from the environmental log and events for the next period used in the prediction processing of step S502, and adds the resulting pair to the current prediction result PR.
[0090] Step S503: The state prediction unit 4288 determines whether the cell image has reached the final state. Here, as an example, the state prediction unit 4288 determines that the cell image has reached the final state when the culture sequence is equal to or longer than a predetermined length. Note that, when the end information FS includes an end image PF, the state prediction unit 4288 may determine that the cell image has reached the final state when the distance between the image feature FP of the predicted cell image and the image feature FP extracted from the end image PF is equal to or less than a predetermined value. Here, the distance between the image feature FPs is, for example, the distance between values of 1 or greater indicating the image feature FPs. For this distance, a Euclidean distance or the like is used.
[0091] If the state prediction unit 4288 determines that the cell image has reached the final state (step S503: YES), it ends the process. On the other hand, if the state prediction unit 4288 determines that the cell image has not reached the final state (step S503: NO), it repeats the process shown in step S502.
[0092] Returning to FIG. 13, the description of the culture process analysis process will be continued. Step S403: The state prediction unit 4288 stores the prediction result PR in the prediction result storage unit 436.
[0093] Returning to FIG. 9, the description of the application stage process continues. Step S303: The parameter acquisition unit 4285 acquires the setting value V from the operation detection unit 423. The parameter acquisition unit 4285 supplies the acquired setting value V to the culture mode selection unit 427. Note that the parameter acquisition unit 4285 may set a predetermined value stored in advance as the setting value V.
[0094] Step S304: The culture mode selection unit 427 presents the culture mode CM. Here, the culture mode selection unit 427 acquires the culture mode CM from the culture mode storage unit 435. The culture mode selection unit 427 acquires the prediction result PR from the prediction result storage unit 436. The culture mode selection unit 427 acquires the setting value V from the parameter acquisition unit 4285.
[0095] The culture mode selection unit 427 selects an environment log and an event for each acquired culture mode CM based on the prediction result PR and the set value V. Here, the culture mode selection unit 427 predicts a predicted value corresponding to the set value V when the final state is reached, based on the environment log and events from the initial state to the selected final state. The predicted value corresponding to the set value V is, for example, a predicted value of the period, cost, and success rate required for culture. The culture mode selection unit 427 selects an environment log and an event whose predicted value satisfies the condition corresponding to the set value V. However, the culture mode selection unit 427 determines which condition of the conditions corresponding to the set value V to prioritize, based on the culture mode CM.
[0096] The culture mode selection unit 427 supplies the culture mode CM and the predicted value corresponding to this culture mode CM to the display control unit 422. The display control unit 422 causes the display unit 44 to display the culture mode CM and the predicted value supplied by the culture mode selection unit 427 as a culture mode selection screen DM.
[0097] Now, with reference to FIG. 15, the culture mode selection screen DM will be described. 15 is a diagram showing an example of the culture mode selection screen DM of this embodiment. The culture mode selection screen DM is displayed on the display unit 44. In this example, quality-focused mode information CM1, production volume-focused mode information CM2, and speed mode information CM3 are displayed as the culture modes CM. In an area AR5 of the culture mode selection screen DM, the quality-focused mode information CM1 is displayed together with a predicted value. In an area AR6 of the culture mode selection screen DM, production volume-focused mode information CM2 is displayed together with a predicted value. In an area AR6 of the culture mode selection screen DM, the speed mode information CM3 is displayed together with a predicted value. The observer selects, from the operation unit 45, one of the culture modes CM displayed on the culture mode selection screen DM according to the culture purpose.
[0098] Returning to FIG. 9, the description of the application stage process continues. Step S305: The culture mode selection unit 427 selects the culture mode CM. Here, the culture mode selection unit 427 acquires information indicating the selected culture mode CM supplied by the operation detection unit 423. The culture mode selection unit 427 selects the culture mode CM based on the acquired information indicating the selected culture mode CM. The culture mode selection unit 427 supplies a set of weights W indicated by the selected culture mode CM to the parameter acquisition unit 4285. The parameter acquisition unit 4285 supplies the weights W supplied by the culture mode selection unit 427 and the set value V supplied by the operation detection unit 423 to the culture protocol presentation unit 4287.
[0099] Step S306: The analysis result presentation unit 428 presents the culture protocol EVC and the final state prediction PFS according to the culture mode CM. Here, the analysis result presentation unit 428 acquires the prediction result PR from the prediction result storage unit 436.
[0100] The cultivation protocol presenting unit 4287 selects a cultivation protocol EVC based on the set value V, weight W, current information PS, termination information FS, and prediction result PR. The cultivation protocol presenting unit 4287 predicts a predicted value corresponding to the set value V when the final state is reached, based on the environmental log and events from the current state to the selected final state. The cultivation protocol presenting unit 4287 selects, as the cultivation protocol EVC, the environmental log and events whose predicted value satisfies the condition corresponding to the set value V.
[0101] However, the culture protocol presenting unit 4287 determines which condition among the conditions corresponding to the set value V should be prioritized based on the weight W. That is, the culture protocol presenting unit 4287 selects a culture protocol EVC according to the culture mode CM. Note that the culture protocol presenting unit 4287 may select a culture protocol EVC without using the weight W.
[0102] The culture protocol presenting unit 4287 supplies the selected culture protocol EVC to the display control unit 422. The display control unit 422 causes the display unit 44 to display the culture protocol EVC supplied by the culture protocol presenting unit 4287 as an analysis result screen DO.
[0103] The culture protocol presenting unit 4287 may select multiple culture protocols EVC. When the culture protocol presenting unit 4287 selects multiple culture protocols EVC, the display control unit 422 displays an analysis result screen DO including the multiple culture protocols EVC supplied by the culture protocol presenting unit 4287 on the display unit 44. The observer may select one of the multiple culture protocols EVC from the analysis result screen DO.
[0104] The state prediction unit 4288 predicts the final state based on the current information PS, the culture protocol EVC selected by the culture protocol presenting unit 4287, and the prediction result PR. Here, the prediction result PR is a result predicted by the state prediction unit 4288 using the non-image feature F extracted from the environmental log and events for a certain period, the prediction result PR, and the learned model (learning result LR) in steps S501 and S502 shown in Fig. 14. In other words, the prediction result PR is a result predicted based on the learning result LR. Therefore, the state prediction unit 4288 predicts the culture state of the cells based on the current information PS acquired by the current information acquisition unit 4282, the culture protocol EVC presented by the culture protocol presentation unit 4287, and the learning result LR.
[0105] The state prediction unit 4288 supplies a final state predicted PFS indicating the predicted final state to the display control unit 422. The state prediction unit 4288 causes the display unit 44 to display the final state predicted PFS supplied by the state prediction unit 4288 as an analysis result screen DO.
[0106] Here, the analysis result screen DO will be described with reference to FIG. 16 is a diagram showing an example of an analysis result screen DO of this embodiment. The analysis result screen DO is displayed on the display unit 44. In this example, speed mode information CM3 is displayed together with a predicted value in an area AR8 of the analysis result screen DO. The culture protocol EVC is displayed in area AR9 on the analysis result screen DO. In this example, the culture protocol EVC indicates instructions for changing the culture medium, points to be careful of, and conditions under which the culture should be stopped. The final state predicted PFS is displayed in area AR10 of the analysis result screen DO. The final state predicted PFS includes the predicted end image PF. In this example, the final state predicted PFS indicates the cell number, quality, and undifferentiated cell rate.
[0107] Returning to FIG. 9, the description of the application stage process continues. Step S307: The operation detection unit 423 detects an operation of the observer to select an event on the operation unit 45. The observer selects an event to be performed from the event selection screen DE displayed on the display unit .
[0108] In the process shown in step S306 above, the culture protocol presenting unit 4287 selects the culture protocol EVC based on the set value V, the weight W, the current information PS, the end information FS, and the prediction result PR, but this is not limiting. The culture protocol presenting unit 4287 may select the culture protocol EVC based on the intermediate information MS in addition to the set value V, the weight W, the current information PS, the end information FS, and the prediction result PR.
[0109] In other words, the culture protocol presentation unit 4287 may present the culture protocol EVC based on the weight W acquired by the parameter acquisition unit 4285, the current information PS acquired by the current information acquisition unit 4282, the termination information FS acquired by the termination information generation unit 4264, the intermediate information MS acquired by the intermediate information generation unit 4263, and the learning result LR.
[0110] In this case, the culture protocol presenting unit 4287 selects the culture protocol EVC for passing through the intermediate state indicated by the intermediate information MS in the process of reaching the final state corresponding to the end image PF indicated by the end information FS.
[0111] Furthermore, in the processing shown in step S306 above, the culture protocol presenting unit 4287 has been described as selecting an environmental log and an event as a culture protocol EVC based on the set value V, weight W, current information PS, termination information FS, and prediction result PR, but this is not limiting. The culture protocol presenting unit 4287 may generate an environmental log and an event as a culture protocol EVC based on the set value V, weight W, current information PS, termination information FS, and prediction result PR. In other words, the culture protocol presenting unit 4287 may generate a culture protocol EVC based on the learning result LR.
[0112] The event selection screen DE will now be described with reference to FIG. 17 is a diagram showing an example of the event selection screen DE of this embodiment. The event selection screen DE is displayed on the display unit 44. In this example, the current image Px is displayed together with the image capture date and time in an area AR11 of the event selection screen DE. A presented event REV and a candidate event OEV are displayed in an area AR12 of the event selection screen DE. The presented event REV is an event indicated by the culture protocol EVC presented by the culture protocol presentation unit 4287. The presented event REV has the text "(recommended)" added to it, indicating that it is an event indicated by the culture protocol EVC. The observer may select a candidate event OEV other than the presentation event REV indicated by the culture protocol EVC.
[0113] Returning to FIG. 9, the description of the application stage process continues. Step S308: The current information acquisition unit 4282 acquires current information PS. The current information PS includes images P acquired from immediately after sowing to the present, implemented event information EV, and environmental log information EL. The current information acquisition unit 4282 supplies the acquired current information PS to the analysis result presentation unit 428.
[0114] Step S309: The state prediction unit 4288 determines whether the image P included in the current information PS has reached the final state. Here, as an example, the state prediction unit 4288 determines that the cell image has reached the final state when the length of the culture sequence is equal to or longer than a predetermined length. Note that the state prediction unit 4288 may also determine that the cell image has reached the final state when the distance between the image feature FP of the image P included in the current information PS and the image feature FP extracted from the final image PF is equal to or shorter than a predetermined value.
[0115] Step S310: The analysis result presentation unit 428 analyzes the culture process. Comparing the process of analyzing the culture process in step S310 with the process of analyzing the culture process in step S302, the difference is that in step S310, the analysis result presentation unit 428 performs the analysis based on current information PS, whereas in step S302, the analysis result presentation unit 428 performs the analysis based on initial information IS. The other processing details in step S310 are the same as in step S302, so a description thereof will be omitted.
[0116] 9, if the culture mode selection unit 427 cannot select an environmental log and an event whose predicted value satisfies the condition corresponding to the set value V, the culture mode selection unit 427 may supply warning information AL0 indicating that it has predicted that a final state that satisfies the set condition cannot be obtained from the cell indicated by the initial image PI to the display control unit 422. The display control unit 422 may cause the warning information AL supplied by the culture mode selection unit 427 to be displayed on the display unit 44.
[0117] 9, if the culture protocol presenting unit 4287 cannot select an environmental log and an event whose predicted value satisfies the condition corresponding to the set value V, the culture protocol presenting unit 4287 may supply warning information AL1 indicating that it has predicted that a final state that satisfies the set condition cannot be obtained from the cell shown in the image P of the current information PS to the display control unit 422. The display control unit 422 may cause the warning information AL1 supplied by the culture protocol presenting unit 4287 to be displayed on the display unit 44.
[0118] In the above embodiment, the termination information generation unit 4264 acquires termination information FS from the operation detection unit 423, that is, the termination information FS is input from the operation unit 45. However, the present invention is not limited to this. The termination information generation unit 4264 may acquire one or more pieces of termination information FSM-1 to FSM-n from the storage unit 43. When the termination information generation unit 4264 acquires termination information FSM-1 to FSM-n from the storage unit 43, the storage unit 43 stores termination information FSM-1 to FSM-n.
[0119] The termination information generation unit 4264 acquires termination information FSM-1 to termination information FSM-n from the storage unit 43 and supplies it to the feature extraction unit 4286. The culture protocol presentation unit 4287 selects a culture protocol EVCM-1 to EVCM-n for each of the termination information FSM-1 to FSM-n supplied by the termination information generation unit 4264. The display control unit 422 causes the display unit 44 to display an analysis result screen DO including the culture protocol EVCM-1 to EVCM-n selected by the culture protocol presentation unit 4287.
[0120] In this example, even if the culture purpose or goal is not input as termination information FS from the operation unit 45, the control device 41 can temporarily set a plurality of culture purposes or goals and present culture protocols EVCM-1 to EVCM-n according to each culture purpose or goal. For example, if the termination information FSM-1 to FSM-n stored in the memory unit 43 include the cell quantity, the control device 41 can present the culture protocols EVCM-1 to EVCM-n as a list of protocols for each cell quantity. After the application stage processing is completed, the control device 41 may execute the learning stage processing using the information obtained in the application stage processing. In this way, the control device 41 executes the learning stage processing every time information on the application stage processing is obtained, thereby always obtaining the best model.
[0121] As described above, the culture support device (control device 41) of this embodiment includes a first information acquisition unit (initial information acquisition unit 4281 or current information acquisition unit 4282), a second information acquisition unit (termination information acquisition unit 4284), a parameter acquisition unit 4285, and a culture protocol presentation unit 4287. The first information acquisition unit (initial information acquisition unit 4281 or current information acquisition unit 4282) acquires first information (initial information IS or current information PS) relating to the cell culture state at a first period in the cell culture process. The second information acquisition unit (end information acquisition unit 4284) acquires second information (end information FS) on the culture state of the cells at a second time period that is later than the first time period. The parameter acquisition unit 4285 acquires a parameter (a set value V or a weight W) used to select a culture protocol EVC indicating the conditions for cell culture. The culture protocol presentation unit 4287 presents the culture protocol EVC based on the parameters (set value V or weight W) acquired by the parameter acquisition unit 4285, the first information (initial information IS or current information PS) acquired by the first information acquisition unit (initial information acquisition unit 4281 or current information acquisition unit 4282), the second information (termination information FS) acquired by the second information acquisition unit (termination information acquisition unit 4284), and the learning result LR in which the correlation between the cell culture results using the culture protocol EVC and the culture protocol EVC has been learned.
[0122] With this configuration, the culture support device (control device 41) of this embodiment can present a culture protocol EVC for obtaining the cell culture state indicated by the second information based on the learning result LR at the first period of the cell culture process, so that the culture techniques and environmental conditions can be fine-tuned successively while checking the culture status according to the purpose and goal of the culture.
[0123] Furthermore, in the culture support device (control device 41) according to this embodiment, the learning result LR is a result obtained by deep learning. With this configuration, the culture assistance device (control device 41) according to this embodiment can present the culture protocol EVC based on a learning result with higher accuracy than when deep learning is not used.
[0124] Moreover, the culture support device (control device 41) according to this embodiment further includes a state prediction unit 4288. The state prediction unit 4288 predicts the culture state of the cells based on the first information (initial information IS or current information PS), the first information (initial information IS or current information PS) acquired by the second information acquisition unit (end information acquisition unit 4284), the culture protocol EVC presented by the culture protocol presentation unit 4287, and the learning result LR.
[0125] With this configuration, the culture support device (control device 41) of this embodiment can predict the culture state of the cells indicated by the second information based on the learning result LR at the first period of the cell culture process, allowing the observer to make fine adjustments successively while checking the predicted culture state.
[0126] Moreover, the culture support device (control device 41) according to this embodiment further includes an intermediate information acquisition unit 4283. The intermediate information acquisition unit 4283 acquires intermediate information MS relating to the culture state of the cells in a period between the first period and the second period. The culture protocol presentation unit 4287 presents the culture protocol EVC based on the parameters (setting value V or weight W) acquired by the parameter acquisition unit 4285, the first information (initial information IS or current information PS) acquired by the first information acquisition unit (initial information acquisition unit 4281 or current information acquisition unit 4282), the second information (termination information FS) acquired by the second information acquisition unit (termination information acquisition unit 4284), the intermediate information MS acquired by the intermediate information acquisition unit 4283, and the learning result LR.
[0127] With this configuration, the culture support device (control device 41) of this embodiment can present a culture protocol EVC at the first period of the cell culture process based on the learning result LR, for obtaining the cell culture state indicated by the second information via the cell culture state at the period between the first period and the second period indicated by the intermediate information MS, thereby making it possible to specify the intermediate state of the culture and perform the culture.
[0128] Moreover, the culture support device (control device 41) according to this embodiment further includes a culture mode storage unit 435 and a culture mode selection unit 427. The culture mode storage unit 435 stores in advance one or more culture modes CM, which are information indicating a predetermined set of parameters. The culture mode selection unit 427 selects a culture mode CM from the one or more culture modes CM stored in the culture mode storage unit 435. The parameter acquisition unit 4285 acquires a parameter (set value V or weight W) indicated by the culture mode CM selected by the culture mode selection unit 427. With this configuration, in the culture support device (control device 41) of this embodiment, there is no need to sequentially specify the parameters (setting value V or weight W) used to present the culture protocol EVC, so that the parameters (setting value V or weight W) used to present the culture protocol EVC can be easily specified.
[0129] The first time point may be the beginning of the culture process, and the second time point may be the end of the culture process. With this configuration, the culture support device (control device 41) according to this embodiment can present a culture protocol EVC for obtaining the cell culture state at the end of the culture process based on the learning result LR at the start of the cell culture process, allowing the observer to fine-tune the culture technique and environmental conditions successively while checking the culture status from the start to the end of the cell culture process. Note that the first time point is not limited to the start of the culture process and may be a later time point. Furthermore, the second time point is not limited to the end of the culture process and may be a time point before the end of the culture process.
[0130] The observation device according to this embodiment includes a temperature-controlled room 15, an imaging device 34, and a culture support device (control device 41). The temperature-controlled room 15 can house a culture vessel 19 for culturing cells, and can maintain the interior at predetermined environmental conditions. The imaging device 34 images the cells housed in the culture vessel 19 within the temperature-controlled room 15 at predetermined time intervals.
[0131] [Variations] As a modification of the above embodiment, a case where there are multiple incubators 11-1 to 11-n will be described. This modification can be applied to a case where, even if it is predicted that cells cultured in one of the incubators, incubator 11-i, will not reach the target final state, cells cultured in another incubator 11-j will be able to reach the final state indicated by the termination information FS.
[0132] The culture system S includes a management device AD and a plurality of incubators 11-1 to 11-n. The incubators 11-1 to 11-n include control devices 41-1 to 41-n, respectively. The management device AD communicates with each of the control devices 41-1 to 41-n. The management device AD and the control devices 41-1 to 41-n each include a communication module, and communicate with each other via wired or wireless communication.
[0133] Among the control devices 41-1 to 41-n, when the control device 41-i is unable to select an environmental log and an event whose predicted value satisfies the condition corresponding to the set value V in the process shown in step S304 of FIG. 9 or the process shown in step S306, it supplies warning information AL (warning information AL0 or warning information AL1) to the management device AD.
[0134] When the management device AD acquires warning information AL supplied by the control device 41-i, it acquires a setting value V and termination information FS from the control device 41-i. Here, any one of the control devices 41-1 to 41-n other than the control device 41-i is designated as a representative control device 41-j. The management device AD supplies the setting value Vi and termination information FS-i acquired from the control device 41-i to the control device 41-j.
[0135] The control device 41-j determines whether the final state indicated by the end information FS-i can be reached from the state of the cell image indicated by the current information PS-j based on the setting value Vi acquired from the control device 41-i, the current information PS-j of the control device 41-j, and the end information FS-i acquired from the control device 41-i. If the control device 41-j determines that the final state indicated by the end information FS-i can be reached from the state of the cell image indicated by the current information PS-j, it supplies to the management device AD culturability information PC indicating that the cells cultured in the incubator 11-j can reach the final state.
[0136] The management device AD supplies the cultivable information PC supplied by the control device 41-j to the control device 41-i. The control device 41-i displays the cultivable information PC supplied by the management device AD on its own display unit 44-i. In addition, instead of including the management device AD, the culture system S may be configured so that direct communication is performed between the control devices 41-1 to 41-n.
[0137] In this modified example, even if it is predicted that cells cultured in one incubator 11-i will not be able to reach the final state indicated by the termination information FS-i, the observer can know that cells cultured in another incubator 11-j of the culture system S are predicted to reach the final state indicated by the termination information FS-i.
[0138] Note that some of the control device 41 in the above-described embodiment, such as the image acquisition unit 421, display control unit 422, operation detection unit 423, memory control unit 424, learning unit 425, analysis information generation unit 426, culture mode selection unit 427, and analysis result presentation unit 428, may be implemented by a computer. In this case, a program for implementing this control function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be loaded into a computer system and executed. Note that the term "computer system" as used herein refers to a computer system built into the control device 41, including hardware such as an OS and peripheral devices. Furthermore, the term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into a computer system. Furthermore, the term "computer-readable recording medium" may also include those that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks such as the Internet or communication lines such as telephone lines, or those that store programs for a certain period of time, such as volatile memory within the computer system serving as the server or client in such cases. Furthermore, the above program may be one that realizes part of the functions described above, or may be one that can realize the functions described above in combination with a program already recorded in the computer system. Furthermore, part or all of the control device 41 in the above-described embodiment may be realized as an integrated circuit such as an LSI (Large Scale Integration). Each functional block of the control device 41 may be individually implemented as a processor, or part or all of the blocks may be integrated into a processor. The integrated circuit implementation method is not limited to LSI, and may be implemented using a dedicated circuit or a general-purpose processor. Furthermore, if an integrated circuit implementation technology that can replace LSI emerges due to advances in semiconductor technology, an integrated circuit based on that technology may be used.
[0139] One embodiment of the present invention has been described in detail above with reference to the drawings, but the specific configuration is not limited to that described above, and various design changes and the like are possible within the scope that does not deviate from the gist of the present invention. [Explanation of symbols]
[0140] 11... incubator (observation device), 41... control device (culture support device), 43... memory unit, 44... display unit, 45... operation unit, 421... image acquisition unit, 422... display control unit, 423... operation detection unit, 424... memory control unit, 425... learning unit, 426... analysis information generation unit, 427... culture mode selection unit, 428... analysis result presentation unit
Claims
[Claim 1] a first information acquiring unit that acquires first information regarding a first culture state of the cells at a first time point in a culture process of the cells; a second information acquisition unit that acquires second information regarding a second culture state of the cells at a second time point that is later than the first time point; a state prediction unit that repeatedly performs a process of predicting the culture state after a predetermined period based on a learning result obtained by learning the correlation between a culture protocol indicating the culture conditions of the cells in the culture process and a change in the culture state of the cells due to the culture protocol, information about the culture state, and the culture protocol, until the culture state reaches the second culture state from the first culture state; a culture protocol selection unit that selects the culture protocol used in the series of prediction processes performed by the state prediction unit as a culture protocol that causes the cell to reach the second culture state from the first culture state; Equipped with When the culture state output by the prediction process has not reached the second culture state, the state prediction unit performs a prediction process based on information about the culture state output by the immediately preceding prediction process, the culture protocol, and the learning result. Culture support equipment.
Citation Information
Patent Citations
Device for automatically detecting optimum culturing condition and culturing device
JP1991259078A
Automatic culture apparatus
JP2010104299A
Apparatus and method for processing cell culture data
JP2013505012A
Cell evaluation apparatus, incubator, cell evaluation method, program and cell cultivation method
JP2016123366A
Culturing apparatus
JP2004229619A