Strain morphology screening system and method based on image technology

Abstract

The present application relates to the technical field of microbial culture, and particularly relates to a strain morphology screening system and method based on image technology. The method comprises the following steps: after diluting the strain to be screened to a predetermined cell concentration, the strain is injected into a microcavity array of a microfluidic chip; images of the microcavity array are periodically collected, morphology parameters of cells in each microcavity are measured, and the collection time is recorded; the morphology change intensity index of each morphology parameter between adjacent two collection time points of the same microcavity is calculated, the change amount of each morphology parameter is weighted and summed according to a preset weight, and then divided by the time interval to obtain the morphology change speed of each microcavity; when the morphology change speed of the microcavity exceeds a preset speed threshold, the microcavity is added to a key monitoring list. The present application realizes efficient screening of target strains by dynamically monitoring and accurately identifying the strain morphology transformation process through the microcavity array of the microfluidic chip combined with morphology parameter extraction and machine learning classification.

Description

Technical Field

[0001] This invention relates to the field of microbial culture technology, and in particular to a strain morphology screening system and method based on image technology. Background Technology

[0002] In biological research and industrial microbiology applications, precise screening and monitoring of strain morphology is crucial. Essentially, this involves systematically observing and analyzing the morphological characteristics of microorganisms at different scales to establish a correlation between morphological phenotype and target function, thereby achieving efficient strain selection. This technology system encompasses macroscopic colony morphology analysis (including parameters such as size, shape, edge, surface texture, color, elevation, and special structures, with quantitative assessment achieved through visual observation, digital imaging combined with image analysis software, or automated colony screening systems), monitoring of liquid culture characteristics (growth patterns, color changes, and viscosity changes), and microscopic cell morphology identification, among others.

[0003] Traditional strain morphology screening relies on manual microscopic observation or batch culture, which cannot capture dynamic changes at the single-cell level (such as the morphology change from hyphae to yeast-like), leading to the neglect of subtle adaptive differences; existing methods (such as flow cytometry) have limited resolution, cannot provide continuous spatiotemporal monitoring, are complex to operate and are prone to losing heterogeneity information; under high-throughput requirements, the lack of automated identification systems leads to low screening efficiency and poor accuracy. Summary of the Invention

[0004] Therefore, it is necessary for the present invention to provide a method and system for screening strain morphology based on image technology to solve at least one of the above-mentioned technical problems.

[0005] To achieve the above objectives, a method for screening bacterial strains based on image technology includes the following steps:

[0006] Step S1: Dilute the strain to be screened to a predetermined cell concentration and inject it into the microcompartment array of the microfluidic chip; periodically acquire images of the microcompartment array at regular time intervals, measure the morphological parameters of the cells in each microcompartment and record the acquisition time;

[0007] Step S2: Calculate the intensity index of morphological changes of each morphological parameter in the same microchamber between two adjacent acquisition times, and determine the stage of cell morphological change; calculate the rate of morphological change of each microchamber based on the morphological transformation stage;

[0008] Step S3: When the morphological change rate of the microchamber exceeds the preset speed threshold, add the microchamber to the key monitoring list; increase the image acquisition frequency for the microchambers in the key monitoring list;

[0009] Step S4: Input the morphological parameters collected by the microchamber array during the culture cycle into the preset machine learning classification model to identify the evolution trajectory of cell morphology type in each microchamber over time, and calculate the time required for morphological change as the morphological transformation time.

[0010] Step S5: Extract the top 20 microchambers with the shortest morphological transformation time. The microchambers were used to record their coordinates and extract strains as screening results.

[0011] The present invention also provides a strain morphology screening system based on image technology, for performing the strain morphology screening method based on image technology as described above, the strain morphology screening system based on image technology comprising:

[0012] The cell acquisition module is used to dilute the strains to be screened to a predetermined cell concentration and then inject them into the microcompartment array of the microfluidic chip; it periodically acquires images of the microcompartment array, measures the morphological parameters of the cells in each microcompartment, and records the acquisition time;

[0013] The morphological velocity calculation module is used to calculate the intensity index of morphological changes of various morphological parameters of the same microchamber between two adjacent acquisition times, and to determine the stage of cell morphological change; and to calculate the morphological change velocity of each microchamber based on the morphological transition stage.

[0014] The morphology-velocity discrimination module is used to add a micro-cavity to the key monitoring list when the morphological change rate of the micro-cavity exceeds a preset velocity threshold; and to increase the image acquisition frequency for micro-cavities in the key monitoring list.

[0015] The morphological trajectory recognition module is used to input the morphological parameters collected by the microchamber array during the culture cycle into a preset machine learning classification model, identify the evolution trajectory of the morphological type of cells in each microchamber over time, and calculate the time required for morphological change as the morphological transition time.

[0016] The dominant strain screening module is used to extract the top 20 strains with the shortest morphological transformation time from each microcompartment. The microchambers were used to record their coordinates and extract strains as screening results.

[0017] This invention distributes the bacterial strains to be screened in a microcompartment array of a microfluidic chip and periodically acquires images. Starting from multidimensional morphological parameters such as cell area, perimeter, major axis length, minor axis length, roundness, and solidity, a time series of cell morphological parameters is constructed. The changes in morphological parameters between adjacent acquisition times are calculated to obtain the rate of morphological change, thereby enabling continuous dynamic monitoring of the cell morphological change process. Furthermore, based on the response characteristics of different morphological parameters to morphological transformation, this invention classifies morphological parameters into three categories: outer contour response parameters, shape stretching response parameters, and dimensional stability parameters. By analyzing the relationship between the intensity of changes in each type of parameter, the initiation, development, and completion stages of morphological transformation are determined. This allows the calculation of the morphological change rate to match the cell morphological evolution process, improving the accuracy of morphological change identification.

[0018] When the rate of morphological change exceeds a threshold, the corresponding microchamber is added to the key monitoring list, and the image acquisition frequency is increased. This enables dynamic hierarchical monitoring of cells undergoing morphological changes, reducing the overall image acquisition and computational burden while ensuring monitoring accuracy. Furthermore, this invention utilizes a machine learning classification model to classify and identify the time series of morphological parameters within the culture cycle, constructing the evolution trajectory of cell morphology types over time. Using the time taken for morphological transformation as a screening indicator, it can quickly identify target strains with rapid morphological transformation rates in a large number of microchambers. This achieves a high-throughput, automated, and objectively stable strain morphology screening process, improving screening efficiency and accuracy. Attached Figure Description

[0019] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0020] Figure 1 This is a schematic diagram of the steps of a strain morphology screening method based on image technology according to the present invention;

[0021] Figure 2 This is a waterfall chart showing the morphological transformation time ranking according to an embodiment of the present invention;

[0022] Figure 3 This is a target location marker diagram in a micro-cavity array according to an embodiment of the present invention;

[0023] Figure 4 This is a block diagram of a strain morphology screening system based on image technology according to the present invention. Detailed Implementation

[0024] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0025] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0026] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0027] To achieve the above objectives, please refer to Figures 1 to 4 This invention provides a method for screening bacterial strains based on image technology, the method comprising the following steps:

[0028] Step S1: Dilute the strain to be screened to a predetermined cell concentration and inject it into the microcompartment array of the microfluidic chip; periodically acquire images of the microcompartment array at regular time intervals, measure the morphological parameters of the cells in each microcompartment and record the acquisition time;

[0029] Step S2: Calculate the intensity index of morphological changes of each morphological parameter in the same microchamber between two adjacent acquisition times, and determine the stage of cell morphological change; calculate the rate of morphological change of each microchamber based on the morphological transformation stage;

[0030] Step S3: When the morphological change rate of the microchamber exceeds the preset speed threshold, add the microchamber to the key monitoring list; increase the image acquisition frequency for the microchambers in the key monitoring list;

[0031] Step S4: Input the morphological parameters collected by the microchamber array during the culture cycle into the preset machine learning classification model to identify the evolution trajectory of cell morphology type in each microchamber over time, and calculate the time required for morphological change as the morphological transformation time.

[0032] Step S5: Extract the top 20 microchambers with the shortest morphological transformation time. The microchambers were used to record their coordinates and extract strains as screening results.

[0033] Furthermore, step S1 includes:

[0034] The bacterial suspension of the strains to be screened was diluted to a concentration of 1.0× -1.0× CFU / mL;

[0035] In one embodiment, after pre-culturing the bacterial strain to be screened, a bacterial suspension is obtained. This suspension is then serially diluted using sterile diluent to maintain a concentration within the range of 1.0 × 10⁵ to 1.0 × 10⁶ CFU / mL. This ensures a suitable number of cells entering the microfluidic chip's microchamber, preventing either overcrowding or absence of cells. Specifically, the cultured bacterial suspension can be first uniformly agitated to fully disperse the cells in the solution. Then, sterile culture medium or buffer solution is added stepwise according to a preset ratio for dilution. The diluted concentration is then calibrated using colony counting or optical density detection.

[0036] For example, 100 μL of the original bacterial suspension can be added to 900 μL of sterile culture medium for a 10-fold dilution, and then further diluted as needed to bring the final concentration into the above range, so as to obtain a bacterial suspension suitable for loading into a microfluidic chip.

[0037] The diluted bacterial suspension is injected into the microchamber array through the inlet of the microfluidic chip;

[0038] In one embodiment, the diluted bacterial suspension is injected into the internal channels of the microfluidic chip through the inlet, allowing the suspension to enter the microchamber array and distribute itself within each microchamber under the influence of a driving force. Specifically, a micropump, pressure-driven device, or manual micropipette can be used to slowly inject the bacterial suspension into the inlet, allowing the liquid to flow uniformly along the microfluidic channels and gradually fill each microchamber.

[0039] For example, a microsyringe containing bacterial suspension can be connected to the chip inlet to push the bacterial suspension into the chip at a constant flow rate (e.g., a few microliters per minute), so that the chambers in the microchamber array are gradually filled with bacterial suspension, thereby randomly distributing cells into different microchambers.

[0040] After the bacterial suspension is injected, preliminary image acquisition is performed to identify microcavities containing cells and generate a list of monitoring cavities.

[0041] In one embodiment, after the bacterial suspension is injected, a preliminary image of the microfluidic chip's microchamber array is acquired using a microscopic imaging system to identify the presence of cells in each microchamber and generate a list of monitored chambers. Specifically, a microscope paired with a digital camera can be used to scan and image the chip as a whole or in sections, obtaining initial images containing multiple microchambers. Then, image processing algorithms are used to analyze the images to detect the presence of cell outlines in each microchamber.

[0042] For example, a simple grayscale thresholding or contour detection algorithm can be used to process the image. When a clear cell contour or pixel clustering area is detected inside a microcavity, it is determined that the microcavity contains cells, and the microcavity number is added to the monitoring cavity list.

[0043] Periodically acquire images of the micro-cavities in the monitoring chamber list at preset time intervals to obtain a periodic image sequence, and record the current timestamp as the acquisition time at each acquisition.

[0044] In one embodiment, images of microcavities in the monitoring chamber list are periodically acquired at preset time intervals to form a periodic image sequence reflecting changes in cell morphology. The current time is recorded as the acquisition time at each acquisition. In practice, an automatic microscope scanning system or a motorized stage can be used to sequentially locate each monitoring microcavity along a preset path, and a digital camera can be used to photograph the corresponding area. The acquired images are then stored in chronological order.

[0045] For example, the acquisition time interval can be set to 5 minutes or 10 minutes, so that the system automatically acquires an image at each time point and stores the image file in association with the current system time, thereby forming time series data containing images at multiple times.

[0046] Edge detection is performed on the periodic image sequence to identify the contour boundaries of cells, thereby extracting the morphological parameters of cells in each microchamber.

[0047] In one embodiment, edge detection processing is performed on each image in the periodic image sequence to identify the contour boundaries of cells, and the morphological parameters of cells are extracted based on the contour information. In specific implementation, common image processing algorithms (such as the Sobel operator, Canny operator, or other edge detection methods) can be used to process the images to obtain the contour boundaries of cells, and then the area, perimeter, and other morphological parameters of cells are calculated based on the positions of the contour pixels.

[0048] For example, the image can first be converted to grayscale and noise filtered. Then, the Canny edge detection algorithm can be used to extract cell contours. The cell area can be calculated based on the number of pixels inside the contour, and the perimeter can be calculated by accumulating the pixel lengths along the contour path. At the same time, parameters such as the major axis length and minor axis length can be obtained through contour geometric analysis.

[0049] Furthermore, the extraction of morphological parameters of cells in each microcompartment includes:

[0050] The perimeter of the cell outline is calculated by accumulating all pixels along the outline boundary, and the area of ​​the cell region is calculated by counting the number of pixels inside the outline boundary and converting them to micrometers.

[0051] In one embodiment, the cell region can be separated from the background through grayscale processing, noise filtering, and threshold segmentation. The closed contour boundary of the cell can then be obtained through edge detection or contour extraction algorithms. After obtaining the cell contour, all pixels constituting the contour are accumulated point-by-point along the contour boundary to count the number of pixels at the contour boundary. The number of pixels is then converted to a length based on the pixel size corresponding to the microscopic imaging system, thus obtaining the cell contour perimeter. Simultaneously, pixel statistics are performed on the internal region enclosed by the contour boundary to obtain the number of pixels inside the contour. Combined with the conversion ratio between pixels and actual micrometer lengths, the number of pixels is converted into an actual area value, thereby obtaining the cell region area.

[0052] For example, in a cell image of a microchamber, after the cell region is obtained through image segmentation, the number of pixels in its outline is counted as 120 pixels. Under the condition that the microscopic system is calibrated so that each pixel corresponds to a length of 0.5 micrometers, the corresponding cell outline perimeter can be calculated. At the same time, the number of pixels inside the cell region is counted as 850 pixels, and the area of ​​the cell region is obtained according to the pixel area conversion relationship.

[0053] Determine the direction of the longest axis and the shortest axis of the cell, and measure the length of the axis in each direction to obtain the length of the long axis and the length of the short axis;

[0054] In one embodiment, after obtaining the cell contour region, geometric analysis of the cell morphology is performed to determine the direction of the longest and shortest axes of the cell. Specifically, ellipse fitting or principal axis analysis can be performed on the cell contour. By calculating the principal direction vector of the cell contour region, the principal axis direction of the cell is determined, and the maximum span distance corresponding to this principal axis direction is taken as the length of the cell's long axis. Simultaneously, the maximum span distance of the contour region in the direction perpendicular to the long axis direction is calculated and taken as the length of the cell's short axis.

[0055] For example, in a cell contour image, principal component analysis determines its principal direction as a certain tilt direction in the image coordinate system, and the maximum distance between the two ends of the contour is measured along this direction as 12 micrometers, which is determined as the length of the major axis; the maximum width is measured in a direction perpendicular to this direction as 7 micrometers, which is then taken as the length of the minor axis.

[0056] Multiply the area of ​​the cell region by four times pi, and divide the product by the square of the cell's outline perimeter to obtain the roundness of the cell.

[0057] In one embodiment, after obtaining the cell region area and cell outline perimeter, a roundness parameter of the cell is calculated using geometric relationships to characterize how close the cell outline is to an ideal circle. Specifically, the cell region area is multiplied by four times pi to obtain an intermediate calculation result, which is then divided by the square of the cell outline perimeter to obtain the cell's roundness value. In this way, the closer the cell shape is to a circle, the closer its roundness value is to 1; while when the cell outline shows significant stretching or irregular deformation, its roundness value will decrease accordingly.

[0058] For example, in a cell image, if the calculated cell area is 50 square micrometers and the outline perimeter is 28 micrometers, the corresponding roundness value can be calculated according to the above formula, which is used to characterize the compactness of the cell's morphology.

[0059] The cell density is obtained by calculating the ratio of the cell region area to the area of ​​the smallest bounding rectangle.

[0060] In one embodiment, after obtaining the cell region area, the ratio between the cell region area and the area of ​​the minimum bounding rectangle is further calculated to obtain the cell solidity parameter. Specifically, firstly, the minimum bounding rectangle that can completely enclose the cell region is determined based on the cell outline, and the area of ​​the rectangle is calculated; then, the cell region area is divided by the area of ​​the minimum bounding rectangle to obtain the solidity value. This parameter reflects the degree of filling of the cell outline within its bounding range. When the cell shape is relatively full or regular, the solidity value is relatively high; when the cell boundary has depressions or the shape is relatively irregular, the solidity value is relatively low.

[0061] For example, in a cell image, if the calculated cell region area is 60 square micrometers and the area of ​​the smallest bounding rectangle is 75 square micrometers, then the ratio of the two can be used as an indicator of the cell's solidity.

[0062] The number, acquisition time, and corresponding cell area, cell outline perimeter, major axis length, minor axis length, roundness, and solidity of each microchamber are associated and stored as morphological parameters.

[0063] In one embodiment, after calculating the aforementioned morphological parameters, the cell morphological parameters corresponding to each microchamber are uniformly organized and stored to form a data record that can be used for subsequent analysis. Specifically, the number of each microchamber is used as a unique identifier, and the acquisition time of the image corresponding to that microchamber is recorded. At the same time, the calculated morphological parameters such as cell area, cell outline perimeter, major axis length, minor axis length, roundness, and solidity are associated and stored to form a structured data record.

[0064] For example, when an image of a microchamber numbered A05 is acquired at a certain time point, information such as "A05 - acquisition time - area - contour perimeter - major axis length - minor axis length - roundness - solidity" can be stored as a data record in a database or data table, and the morphological parameters corresponding to the same microchamber can be recorded at subsequent time points, thereby forming a morphological data sequence that changes over time.

[0065] Furthermore, step S2 includes the following steps:

[0066] Define morphological change intensity indices, including:

[0067] Rate of change of outer contour ,in This represents the change in the cell's perimeter. The perimeter of the cell outline. For roundness, This represents the change in roundness.

[0068] Shape stretching ,in The length of the major axis. This represents the change in the major axis. For the minor axis length, This represents the change in the minor axis;

[0069] Dimensional stability ,in The area of ​​the cell region. This represents the change in area. For the degree of reality, This represents the change in realness.

[0070] In one embodiment, after obtaining the morphological parameters of cells in each microchamber at different acquisition times, the changes in morphological parameters between two adjacent acquisition times are calculated to construct a morphological change intensity index for characterizing the degree of cell morphological change.

[0071] Specifically, for each cell in a microchamber, parameters such as cell contour perimeter, roundness, major axis length, minor axis length, cell area, and solidity are first obtained based on two consecutively acquired image data sets. The changes in each parameter between adjacent time points are then calculated. Subsequently, the change in contour perimeter is normalized to the corresponding contour perimeter, and the change in roundness is normalized to the roundness. The squares of these two values ​​are then summed and the square root is taken to obtain the outer contour change rate. , used to characterize the degree of change in the overall outline morphology of the cell;

[0072] The change in major axis length is normalized to the major axis length itself, and the change in minor axis length is also normalized to the minor axis length. The squares of both are then summed and the square root is taken to obtain the shape stretching ratio. This is used to characterize changes in cell morphology along its length due to stretching or compression; the change in cell area is normalized to the area itself.

[0073] The change in actuality and the actuality are normalized, and the squares of the two are summed and then the square root is taken to obtain the dimensional stability index. It is used to characterize changes in the overall size and morphological filling degree of cells.

[0074] For example, in a microchamber, if the cell contour perimeter is obtained at two adjacent acquisition times of 30 micrometers and 33 micrometers respectively, the change in contour perimeter can be calculated as 3 micrometers; simultaneously, if the corresponding roundness is 0.85 and 0.80 respectively, the change in roundness can be calculated as 0.05. By substituting these values ​​into the formula, the rate of change of the outer contour during that time period can be calculated. .

[0075] Determining the morphological transformation stage based on morphological change intensity indicators:

[0076] when At that time, it is determined to be the start-up phase, in which This is the activation threshold coefficient, with a value ranging from 1.2 to 2.0.

[0077] when Reaching a local maximum and At that time, it was determined to be in the development stage;

[0078] when , , When all three show a downward trend, it is determined to be the completion stage;

[0079] In one embodiment, after calculating the rate of change of the outer contour... Shape stretching ratio and dimensional stability indicators Subsequently, the morphological transformation process of cells was divided into stages based on the relative relationships among the three types of morphological change indicators. The calculated rate of change in the outer contour was then used to further analyze the process. Greater than the sum of shape elongation and dimensional stability indices multiplied by the activation threshold coefficient. When the cell's outer contour boundary changes, it is considered that this stage is the initiation stage of morphological transformation, and the initiation threshold coefficient is determined to be... It can be set between 1.2 and 2.0 to adjust the sensitivity of the judgment during the startup phase; when the shape stretching rate It reaches a local maximum in the time series, and this value is greater than the rate of change of the outer contour at the corresponding time point. When the change in cell morphology is mainly manifested as axial stretching or shape extension, it is considered to be a stage of morphological transformation; when the rate of change in the outer contour is... Shape stretching ratio and dimensional stability indicators When all three show a downward trend at consecutive time points, it indicates that the changes in cell morphology are gradually weakening and stabilizing, thus indicating the completion of the morphological transformation stage.

[0080] For example, in the time series analysis of a certain cell, if the calculation is performed in the early stages... Significantly greater than and The sum of these values ​​indicates that this stage is the initiation stage of morphological change; if at subsequent time points... Gradually increase and reach peak and exceed If the three indicators gradually decrease, it can be determined that this stage is the main stage of morphological development; while when all three indicators gradually decrease, it indicates that the cell morphology is gradually stabilizing, and this stage can be regarded as the stage where morphological transformation is completed.

[0081] The morphological change rate of each microchamber was calculated based on the morphological transformation stage. :

[0082] During the startup phase ;

[0083] During the development stage, ;

[0084] In the completion phase, The weighting coefficient , , Preset weighting coefficients, and This represents the time interval between two consecutive data collections.

[0085] In one embodiment, after determining the morphological transformation stage of the cell, the corresponding morphological change rate is calculated based on the morphological change characteristics of different stages. This is used to characterize the degree of morphological change in cells per unit time. Specifically, in the initiation phase, since cell morphological changes are mainly manifested as changes in the outer contour boundary, the outer contour change rate is used. Divide by the time interval between two consecutive image acquisitions This allows us to obtain the rate of morphological change at this stage. In the development stage, since cell morphological changes are not only reflected in changes in the outer contour but also accompanied by significant axial stretching, the rate of change in the outer contour is thus determined. With shape stretching Superimpose and then divide by the time interval To obtain the rate of morphological change at this stage; in the completion stage, since the cell morphological changes gradually tend to stabilize, a weighted combination of the rate of change of outer contour, the rate of shape stretching, and the dimensional stability index is used, and the weighted result is divided by the time interval. This yields the overall rate of change in morphology, where the weighting coefficients are... , , This is used to reflect the degree of influence of different morphological change indicators during the completion stage.

[0086] For example, if a microchamber is calculated during the startup phase... If the value is 0.18 and the time interval between two consecutive image acquisitions is 10 minutes, then the rate of morphological change in this stage can be obtained by dividing 0.18 by 10 minutes; while in the completion stage, if , , The values ​​are 0.04, 0.03, and 0.02 respectively, and the weighting coefficients are set to 0.4, 0.35, and 0.25 respectively. The overall morphological change rate can be obtained by weighted summation and then divided by the time interval.

[0087] It should be noted that the weighting coefficients , , Pre-settings can be made based on the morphological change characteristics of different cell types during the experiment, so that the calculated morphological change rate is more consistent with the actual biological change patterns.

[0088] Furthermore, step S3 includes the following steps:

[0089] Step S31: After extracting the morphological parameters at each acquisition moment and calculating the morphological change rate of each microchamber, compare the morphological change rate of each microchamber with the preset velocity threshold one by one.

[0090] In one embodiment, the data processing module reads the morphological change rate data of each microchamber at the current acquisition time, and traverses it sequentially according to the microchamber number. Each morphological change rate value is compared with a preset rate threshold. When the morphological change rate reaches or exceeds the threshold, it is marked as "significant change state".

[0091] For example, at a certain acquisition time, the system calculates that the morphological change rates of microchambers A01, A02, and A03 are 0.05, 0.22, and 0.11, respectively, while the preset rate threshold is 0.15. Therefore, the system judges that the cell morphological change of microchamber A02 is significant, while A01 and A03 are still in the normal change state.

[0092] Step S32: When the morphological change rate of a certain microchamber at the current acquisition time is greater than the speed threshold, record the acquisition time as the trigger time and add the number of the microchamber to the key monitoring list;

[0093] In one embodiment, a key monitoring list structure is established in the data recording module. When a microchamber that meets the conditions is detected, the microchamber number and the trigger time are written into the list, and the list is read first in subsequent acquisition processes to adjust the acquisition strategy.

[0094] For example, when the microchamber numbered B12 detects a morphological change rate of 0.21 at the acquisition time of 12:30, which exceeds the threshold of 0.15, the system records 12:30 as the trigger time of the microchamber and adds B12 to the key monitoring list, thereby marking the microchamber as an object that needs to be observed.

[0095] It should be noted that once a microchamber is included in the key monitoring list, images will be acquired using a different acquisition strategy than those for conventional microchambers in subsequent time periods, in order to improve the observation accuracy of morphological transformation processes.

[0096] Step S33: Adjust the image acquisition strategy for the micro-chambers in the key monitoring list, shortening the acquisition time interval to 50% of the normal time interval to form a high-frequency acquisition sequence.

[0097] In one embodiment, after generating a key monitoring list, the image acquisition strategy for each microchamber is dynamically adjusted based on its recorded microchamber number. The image acquisition interval for these microchambers is shortened to half the regular acquisition interval, thereby forming a high-frequency acquisition sequence with higher temporal resolution to more precisely record cell morphological changes. During the acquisition task, the key monitoring list is read first, and images are acquired from microchambers in the list using the shortened acquisition interval, while the original regular acquisition interval is maintained for microchambers not in the list.

[0098] For example, if the system's regular image acquisition interval is 10 minutes, the acquisition interval for key monitored micro-cavities will be adjusted to 5 minutes to obtain a denser sequence of morphological change data.

[0099] Furthermore, step S3 also includes the following steps:

[0100] Step S34: Continuously calculate the morphological change rate of the high-frequency acquisition sequence as the high-frequency morphological change rate;

[0101] In one embodiment, at each high-frequency acquisition moment, the morphological parameters of the microchamber cell, such as area, perimeter, major axis length, minor axis length, roundness, and solidity, are acquired and the difference between these parameters and the corresponding parameters at the previous high-frequency acquisition moment is calculated. Then, the parameters are weighted and summed according to predetermined morphological parameter weights and divided by the current high-frequency acquisition time interval to obtain the high-frequency morphological change rate at that moment.

[0102] For example, if the original routine sampling interval for a microchamber was 10 minutes, but was adjusted to 5 minutes after being added to the key monitoring list, the system will continuously calculate the rate of morphological change at 5-minute intervals, thus obtaining a finer-grained sequence of change rates. For instance, high-frequency morphological change rate values ​​of 0.19, 0.17, and 0.16 can be calculated at times such as 13:05, 13:10, and 13:15, respectively. This continuous calculation method can more accurately reflect the dynamic trend of cell morphology changes over a short period of time.

[0103] Step S35: Continue to acquire images and calculate the rate of morphological change for micro-chambers that are not included in the key monitoring list at regular time intervals;

[0104] In one embodiment, for microcavities not included in the key monitoring list, the system continues to acquire images and calculate morphological change rates according to a pre-set regular acquisition time interval to maintain the basic monitoring status of the entire microcavity array. Specifically, when performing an acquisition task, the key monitoring list is first read, and a high-frequency acquisition strategy is executed for the microcavities in the list. For unmarked microcavities, the original acquisition cycle is maintained, for example, acquiring images every 10 minutes, and calculating their morphological change rate based on the change in morphological parameters between two adjacent acquisition times.

[0105] For example, in a microchamber array containing 200 microchambers, only 15 are included in the key monitoring list, while the remaining 185 microchambers are still image-acquired and velocity-calculated at 10-minute intervals. For instance, if the morphological change velocity of a certain microchamber is calculated to be 0.07 between 14:00 and 14:10, its change status is continuously updated. In this way, key monitoring of microchambers potentially undergoing morphological changes can be carried out without significantly increasing the overall data acquisition burden of the system, while maintaining basic monitoring of the remaining microchambers.

[0106] Step S36: During the high-frequency acquisition process, if the high-frequency morphological change rate of a key monitoring micro-chamber is lower than the rate threshold in multiple consecutive acquisitions, it is determined that the micro-chamber has completed the morphological transformation process. The micro-chamber is then removed from the key monitoring list and its regular time interval acquisition is resumed; in order to achieve dynamic hierarchical monitoring of the morphological transformation process of each micro-chamber.

[0107] In one embodiment, during high-frequency acquisition of key monitored microcavities, the system continuously judges the rate of high-frequency morphological change. When it is detected that the rate of high-frequency morphological change of a key monitored microcavity is lower than a preset rate threshold after multiple consecutive acquisitions, it is considered that the cell morphological change in the microcavity has stabilized, that is, the morphological transformation process is basically completed. Therefore, the microcavity is removed from the key monitoring list, and its regular image acquisition time interval is restored. In specific implementation, a threshold for the number of consecutive judgments can be set; for example, if the high-frequency acquisition results are lower than the rate threshold after three or five consecutive acquisitions, the judgment is triggered to exit.

[0108] For example, when the morphological change rate of a key monitored microchamber is calculated to be 0.08, 0.07, 0.06, and 0.05 in four consecutive high-frequency acquisition times, respectively, while the preset rate threshold is 0.15, the system determines that the morphological change of the microchamber has slowed down significantly, thereby removing it from the key monitoring list and restoring its acquisition strategy to the routine image acquisition every 10 minutes. This dynamic exit mechanism allows the system to automatically adjust the monitoring level according to the cell morphological change status during the culture cycle, thus forming a dynamic hierarchical monitoring mechanism. This means high-frequency monitoring is performed on microchambers with active morphological changes, while routine monitoring is restored for morphologically stable microchambers, thereby improving overall monitoring efficiency and data acquisition accuracy.

[0109] Of particular importance, the preset morphological parameter weights are determined in the following way:

[0110] Preliminary experiments were conducted on the reference strain, and the morphological changes of multiple microcompartments under stress conditions were continuously monitored.

[0111] For microcavities that undergo morphological transformation during the morphological change process, extract the change data of each morphological parameter before the transformation occurs;

[0112] The correlation between the change in each morphological parameter and the occurrence of morphological transformation is calculated separately. The morphological parameter with the highest correlation is taken as the benchmark value. The correlation of the remaining morphological parameter changes is compared with the benchmark value to obtain the morphological parameter weights.

[0113] Furthermore, step S4 includes the following steps:

[0114] Step S41: After the culture cycle is completed, extract the morphological parameters recorded in each microchamber throughout the entire culture cycle, and construct the time series of morphological parameters of each microchamber according to the order of acquisition time.

[0115] In one embodiment, after the culture cycle ends, morphological parameters calculated at each acquisition time are read from the image analysis module or database. These parameters include cell area, perimeter, major axis length, minor axis length, aspect ratio, roundness, solidity, and number of branches. The data are then sorted according to the acquisition timestamps, arranging them sequentially from the start to the end of the culture. Subsequently, using each acquisition time as a time node, the corresponding morphological parameters are combined to form a multidimensional parameter set, which is then concatenated in chronological order to form the morphological parameter time series of the microchamber.

[0116] For example, if a microchamber acquires 60 images during the culture process, the system will extract the morphological parameters at these 60 moments and construct a time series such as {P1, P2, P3, ..., P60} in chronological order, where each Pi contains multiple parameters such as area, roundness, and aspect ratio. By constructing this time series, the morphological changes of cells throughout the entire culture cycle can be fully reflected.

[0117] Step S42: Input the time series of morphological parameters of each microchamber at each acquisition time as feature vectors into the preset machine learning classification model, classify the feature vectors at all acquisition times, and obtain the morphological type label corresponding to each acquisition time.

[0118] In one embodiment, the morphological parameters are standardized or normalized as necessary to eliminate the influence of different parameter dimensions on the model input. Then, a vector containing multiple morphological parameters is used as the model input features, such as parameters [area, roundness, aspect ratio, number of branches], to form a feature vector. This feature vector is input to a preset classification model, such as a support vector machine model, a random forest model, or a convolutional neural network classification model. The model outputs the corresponding morphological type label based on the learned morphological feature patterns, such as "mycelial morphology" or "yeast morphology".

[0119] For example, when the morphological parameters at a certain sampling time are an area of ​​120 μm², a roundness of 0.35, and an aspect ratio of 4.8, the model may classify it as a mycelial morphology. However, when the roundness increases to 0.82 and the aspect ratio decreases to 1.3, the model may classify it as a yeast-like morphology. By classifying all sampling times, a morphological type label corresponding to each time point can be obtained.

[0120] Furthermore, step S4 also includes the following steps:

[0121] Step S43: Arrange the morphological type labels of each microchamber according to the acquisition time sequence to form a morphological type time series;

[0122] In one embodiment, the category labels are associated with their corresponding acquisition times and sorted according to the timestamps. Then, the labels are arranged in chronological order, for example, forming a sequence such as {C1, C2, C3, ..., Cn}, where each Ci represents the morphology type of the i-th acquisition time.

[0123] For example, if a microchamber is identified as having a mycelial morphology in the first 20 sampling times, but is identified as having a yeast-like morphology after the 21st sampling time, then the morphological type time series of this microchamber might appear as {mycelial, mycelial, mycelial, ..., yeast-like, yeast-like}. This time series format allows for a clear observation of the changing trend of morphological type during the cultivation process, thus providing a data foundation for subsequent identification of morphological transition times.

[0124] Step S44: Identify the time point in the morphological type time series where the morphological type first changes from mycelial to yeast-like, and record the corresponding collection time as the time of category change;

[0125] In one embodiment, the category labels of adjacent time points are compared one by one starting from the beginning of the time series. When it is detected that the label before a certain time point is a mycelial morphology and the label at the current time point changes to a yeast morphology, the time point is determined to be the starting point of the morphology type change.

[0126] For example, in a categorical time series of a microchamber, the first 18 collections all show a mycelial morphology, while the 19th collection is identified as a yeast-like morphology. The system then records the timestamp corresponding to the 19th collection as the time of categorical change. By identifying this change point, the key time point in the cell's transition from a mycelial to a yeast-like morphology can be determined.

[0127] Step S45: Compare the time of category change with the time of incubation start of the microchamber, and calculate the time difference between the two as the time of morphological transformation of the microchamber.

[0128] In one embodiment, the initial timestamp of when the microchamber begins culture is read, such as the time when cells are seeded into the microchamber or when culture conditions change, and then the difference between the time of class change and the starting time is calculated to obtain a time interval data.

[0129] For example, if the culture start time of a microchamber is 08:00, and the morphological type change time identified by the system is 12:30, then the morphological transformation time of that microchamber is 4 hours and 30 minutes. By calculating this time difference for all microchambers, morphological transformation time data for multiple samples can be obtained, which can then be used for further statistical analysis, such as calculating the average transformation time or analyzing the differences in morphological transformation under different experimental conditions.

[0130] Of particular importance, determining the time point of the first change also includes the following steps:

[0131] When the morphology of a certain microchamber changes from a mycelial morphology to a yeast morphology for the first time at a certain acquisition time in the morphology time series, the acquisition time is marked as a candidate transition time.

[0132] After the candidate transition time, continue to read the morphology type labels corresponding to the subsequent acquisition times of the microchamber in the order of acquisition time; determine whether the morphology type labels of the subsequent acquisition times continue to remain as yeast-like morphology type;

[0133] When the morphological type label of several subsequent collection times is all yeast-like morphological type, the candidate transition time is confirmed as the category change time.

[0134] If, during subsequent data collection, the morphological type changes back to a mycelial morphological type or to another morphological type, the candidate transition time is removed from the list, and the search for the next candidate transition time continues in the morphological type time series.

[0135] Furthermore, step S5 includes the following steps:

[0136] Step S51: Associate the number of each microchamber with its morphological transformation time and form a time record table;

[0137] In one embodiment, the system reads the morphological transformation time of each microchamber calculated in step S45 and pairs it with the unique number of the microchamber for recording. Each record contains at least two information fields: microchamber number and corresponding morphological transformation time, for example, {microchamber number: A01, transformation time: 4 hours 30 minutes}.

[0138] For example, if there are 100 microcavities in a microfluidic chip, the time log will contain 100 records, each corresponding to a microcavity and its time data.

[0139] Step S52: Sort the time consumption record table according to the order of morphological transformation time from small to large, and generate a sorted list;

[0140] In one embodiment, each record in the time consumption record table is read and sorted in ascending order according to the time consumption value, so that the microcavities with the fastest morphological transformation are placed at the front of the list, while the microcavities with slower transformation or no transformation are placed at the back.

[0141] For example, for 100 records, the system sorts them by time to obtain a sorted list {A23, 1h10min;B07, 1h25min;…;C12, >12h}, with the microchambers at the front of the list being the ones that change the fastest.

[0142] Step S53: Calculate the top 20 items in the sorted list The number of microchambers, and the top 20 in the sorted list. Microcavities; extract the microcavity numbers and their coordinate positions in the microfluidic chip within the marked area to form a list of target microcavities;

[0143] In one embodiment, the total number of microchambers N is first counted, and then the number of the top 20% is calculated. And select the top from the head of the sorted list The system then extracts the numbers of these microcavities and their coordinate information on the microfluidic chip, such as row and column numbers or absolute coordinates, and generates a list of target microcavities for subsequent micromanipulation positioning.

[0144] For example, if there are 100 microcavities in the chip, the first 20% of the microcavities will be 20. The system will mark the 20 microcavities with the shortest processing time in the sorted list and record their numbers and coordinates to form a target list {number A23, coordinates (2,3); number B07, coordinates (1,7);...}.

[0145] Step S54: According to the target microcompartment list, the cells in the corresponding microcompartments are located and extracted using a micromanipulation device. The extracted cells are then transferred to a culture medium for amplification culture to obtain the screened target strain.

[0146] In one embodiment, the micromanipulation device utilizes the coordinate information of a microfluidic chip to move a micropipette or micropipette to the target microchamber location for micro-aspiration or micromanipulation extraction of single cells. After extraction, the operating system releases the cells into a pre-prepared culture medium, such as liquid or solid agar medium, and places them under suitable culture conditions for amplification.

[0147] For example, in the microchamber numbered A23, a micromanipulation device aspirates a single cell into a micropipette under microscopic observation and releases it into a culture dish for incubation.

[0148] like Figure 2 The image shows a waterfall plot ranking of morphological transformation times. The plot contains 50 vertical bars, each representing a microchamber. The height of the bar indicates the time required for cells within that microchamber to transform from one morphology to another. All bars are arranged in ascending order of morphological transformation time, creating a gradient visual effect. The first 10 bars on the left (representing 20% ​​of the total) are filled with a dark color, representing the dominant strains with the fastest morphological transformation speed; the 40 bars on the right are filled with a light color, representing strains with slower transformation. A red dashed box marks the top 20% of dominant strains. This visualization method allows for intuitive identification of performance differences and rapid determination of screening targets.

[0149] like Figure 3 The image shows a microcavity array and target location markers for a microfluidic chip. The figure displays an 8×8 regularly arranged circular microcavity grid, with each circle representing an independent microcavity. According to... Figure 2 The screening results identified the top 20% of microchambers (10 in total) with the shortest morphological transformation time, which were marked with orange asterisks in the array. These asterisks were distributed across different locations on the chip, indicating the randomness of the spatial distribution of the dominant strains. By recording the row and column coordinates of the asterisk-marked microchambers, a precise correspondence between the data analysis results and their physical locations could be established, providing location information for subsequent targeted extraction operations.

[0150] See Figure 4 The present invention also provides a strain morphology screening system 100 based on image technology, used to perform the strain morphology screening method based on image technology as described above, the strain morphology screening system 100 based on image technology includes:

[0151] The cell acquisition module 101 is used to inject the strain to be screened into the microcompartment array of the microfluidic chip after diluting it to a predetermined cell concentration; periodically acquire images of the microcompartment array, measure the morphological parameters of the cells in each microcompartment, and record the acquisition time;

[0152] The morphological velocity calculation module 102 is used to calculate the intensity index of morphological changes of each morphological parameter in the same microchamber between two adjacent acquisition times, and to determine the stage of cell morphological change; and to calculate the morphological change velocity of each microchamber based on the morphological transformation stage.

[0153] The morphology-velocity discrimination module 103 is used to add the micro-cavity to the key monitoring list when the morphological change rate of the micro-cavity exceeds a preset velocity threshold; and to increase the image acquisition frequency for the micro-cavities in the key monitoring list.

[0154] The morphological trajectory recognition module 104 is used to input the morphological parameters collected by the microchamber array during the culture cycle into a preset machine learning classification model, identify the evolution trajectory of the morphological type of cells in each microchamber over time, and calculate the time required for morphological change as the morphological transformation time.

[0155] The superior strain screening module 105 is used to extract the top 20 strains with the shortest morphological transformation time in each microcompartment. The microchambers were used to record their coordinates and extract strains as screening results.

[0156] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0157] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for screening bacterial strains based on image technology, characterized in that, Includes the following steps: Step S1: Dilute the strain to be screened to a predetermined cell concentration and inject it into the microcompartment array of the microfluidic chip; Images of the microchamber array were periodically acquired at regular time intervals, the morphological parameters of cells in each microchamber were measured, and the acquisition time was recorded. Step S2: Calculate the intensity index of morphological changes of each morphological parameter in the same microchamber between two adjacent acquisition times, and determine the stage of cell morphological change; calculate the rate of morphological change of each microchamber based on the morphological transformation stage; Step S3: When the morphological change rate of the microchamber exceeds the preset speed threshold, add the microchamber to the key monitoring list; increase the image acquisition frequency for the microchambers in the key monitoring list; Step S4: Input the morphological parameters collected by the microchamber array during the culture cycle into the preset machine learning classification model to identify the evolution trajectory of cell morphology type in each microchamber over time, and calculate the time required for morphological change as the morphological transformation time. Step S5: Extract the top 20 microchambers with the shortest morphological transformation time. The microchambers were used to record their coordinates and extract strains as screening results.

2. The method for screening bacterial strains based on image technology according to claim 1, characterized in that, Step S1 includes: The bacterial suspension of the strains to be screened was diluted to a concentration of 1.0× -1.0× CFU / mL; The diluted bacterial suspension is injected into the microchamber array through the inlet of the microfluidic chip; After the bacterial suspension is injected, preliminary image acquisition is performed to identify microcavities containing cells and generate a list of monitoring cavities. Periodically acquire images of the micro-cavities in the monitoring chamber list at preset time intervals to obtain a periodic image sequence, and record the current timestamp as the acquisition time at each acquisition. Edge detection is performed on the periodic image sequence to identify the contour boundaries of cells, thereby extracting the morphological parameters of cells in each microchamber.

3. The method for screening bacterial strains based on image technology according to claim 2, characterized in that, The extraction of morphological parameters of cells in each microcompartment includes: The perimeter of the cell outline is calculated by accumulating all pixels along the outline boundary, and the area of ​​the cell region is calculated by counting the number of pixels inside the outline boundary and converting them to micrometers. Determine the direction of the longest axis and the shortest axis of the cell, and measure the length of the axis in each direction to obtain the length of the long axis and the length of the short axis; Multiply the area of ​​the cell region by four times pi, and divide the product by the square of the cell's outline perimeter to obtain the roundness of the cell. The cell density is obtained by calculating the ratio of the cell region area to the area of ​​the smallest bounding rectangle. The number, acquisition time, and corresponding cell area, cell outline perimeter, major axis length, minor axis length, roundness, and solidity of each microchamber are associated and stored as morphological parameters.

4. The method for screening bacterial strains based on image technology according to claim 3, characterized in that, Step S2 includes the following steps: Define morphological change intensity indices, including: Rate of change of outer contour ,in This represents the change in the cell's perimeter. The perimeter of the cell outline. For roundness, This represents the change in roundness. Shape stretching ,in The length of the major axis. This represents the change in the major axis. For the minor axis length, This represents the change in the minor axis; Dimensional stability ,in The area of ​​the cell region. This represents the change in area. For the degree of reality, This represents the change in realness. Determining the morphological transformation stage based on morphological change intensity indicators: when At that time, it is determined to be the start-up phase, in which This is the activation threshold coefficient, with a value ranging from 1.2 to 2.

0. when Reaching a local maximum and At that time, it was determined to be in the development stage; when , , When all three show a downward trend, it is determined to be the completion stage; The morphological change rate of each microchamber was calculated based on the morphological transformation stage. : During the startup phase ; During the development stage, ; In the completion phase, The weighting coefficient , , Preset weighting coefficients, and This represents the time interval between two consecutive data collections.

5. The method for screening bacterial strains based on image technology according to claim 4, characterized in that, Step S3 includes the following steps: Step S31: After extracting the morphological parameters at each acquisition moment and calculating the morphological change rate of each microchamber, compare the morphological change rate of each microchamber with the preset velocity threshold one by one. Step S32: When the morphological change rate of a certain microchamber at the current acquisition time is greater than the speed threshold, record the acquisition time as the trigger time and add the number of the microchamber to the key monitoring list; Step S33: Adjust the image acquisition strategy for the micro-chambers in the key monitoring list, shortening the acquisition time interval to 50% of the normal time interval, so as to increase the image acquisition frequency and form a high-frequency acquisition sequence.

6. The method for screening bacterial strains based on image technology according to claim 5, characterized in that, Step S3 also includes the following steps: Step S34: Continuously calculate the morphological change rate of the high-frequency acquisition sequence as the high-frequency morphological change rate; Step S35: Continue to acquire images and calculate the rate of morphological change for micro-chambers that are not included in the key monitoring list at regular time intervals; Step S36: During the high-frequency acquisition process, if the high-frequency morphological change rate of a key monitoring micro-chamber is lower than the rate threshold in multiple consecutive acquisitions, it is determined that the micro-chamber has completed the morphological transformation process. The micro-chamber is then removed from the key monitoring list and its regular time interval acquisition is resumed; in order to achieve dynamic hierarchical monitoring of the morphological transformation process of each micro-chamber.

7. The method for screening bacterial strains based on image technology according to claim 6, characterized in that, Step S4 includes the following steps: Step S41: After the culture cycle is completed, extract the morphological parameters recorded in each microchamber throughout the entire culture cycle, and construct the time series of morphological parameters of each microchamber according to the order of acquisition time. Step S42: Input the time series of morphological parameters of each microchamber at each acquisition time as feature vectors into the preset machine learning classification model, classify the feature vectors at all acquisition times, and obtain the morphological type label corresponding to each acquisition time.

8. The method for screening bacterial strains based on image technology according to claim 7, characterized in that, Step S4 also includes the following steps: Step S43: Arrange the morphological type labels of each microchamber according to the acquisition time sequence to form a morphological type time series; Step S44: Identify the time point of the first morphological change in the morphological type time series and record the corresponding acquisition time as the category change time; Step S45: Compare the time of category change with the time of incubation start of the microchamber, and calculate the time difference between the two as the time of morphological transformation of the microchamber.

9. The method for screening bacterial strains based on image technology according to claim 8, characterized in that, Step S5 includes the following steps: Step S51: Associate the number of each microchamber with its morphological transformation time and form a time record table; Step S52: Sort the time consumption record table according to the order of morphological transformation time from small to large, and generate a sorted list; Step S53: Calculate the top 20 items in the sorted list The number of microchambers, and the top 20 in the sorted list. Microcavities; extract the microcavity numbers and their coordinate positions in the microfluidic chip within the marked area to form a list of target microcavities; Step S54: According to the target microcompartment list, the cells in the corresponding microcompartments are located and extracted using a micromanipulation device. The extracted cells are then transferred to a culture medium for amplification culture to obtain the screened target strain.

10. A bacterial strain morphology screening system based on image technology, characterized in that, For performing the image-based strain morphology screening method as described in claim 1, the image-based strain morphology screening system comprises: The cell acquisition module is used to dilute the strains to be screened to a predetermined cell concentration and then inject them into the microcompartment array of the microfluidic chip; it periodically acquires images of the microcompartment array, measures the morphological parameters of the cells in each microcompartment, and records the acquisition time; The morphological velocity calculation module is used to calculate the intensity index of morphological changes of various morphological parameters of the same microchamber between two adjacent acquisition times, and to determine the stage of cell morphological change; and to calculate the morphological change velocity of each microchamber based on the morphological transition stage. The morphology-velocity discrimination module is used to add a micro-cavity to the key monitoring list when the morphological change rate of the micro-cavity exceeds a preset velocity threshold; and to increase the image acquisition frequency for micro-cavities in the key monitoring list. The morphological trajectory recognition module is used to input the morphological parameters collected by the microchamber array during the culture cycle into a preset machine learning classification model, identify the evolution trajectory of the morphological type of cells in each microchamber over time, and calculate the time required for morphological change as the morphological transition time. The dominant strain screening module is used to extract the top 20 strains with the shortest morphological transformation time from each microcompartment. The microchambers were used to record their coordinates and extract strains as screening results.

Images