Imaging system and method of use thereof

Abstract

To provide an imaging system and a controller that generate and / or analyze a variety of objects, such as biological cells to determine clonality.SOLUTION: An imaging system comprises: an imaging device; and a controller in operable connection to the imaging device. The controller performs the steps of: generating a plurality of chronological images of an image area; identifying a target object within the image area; using a prior image of the image area, and cropping the prior image to generate a cropped image area sized to the perimeter of the target object image area; generating a location region of the cropped image area within the image area of the most recent image; analyzing the location region of the most recent image; and inductively iterating the steps for the plurality of chronological images to determine clonality of the target object.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of priority under 35 U.S.C. §119(e) to U.S. Provisional Patent Application No. 62 / 910,951, filed October 4, 2019; U.S. Provisional Patent Application No. 62 / 971,017, filed February 6, 2020; and U.S. Provisional Patent Application No. 63 / 051,310, filed July 13, 2020; the entire contents of which are incorporated herein by reference.

[0002] FIELD OF THE INVENTION The present invention relates generally to imaging, and more particularly to systems and methods for generating images of target objects and analyzing the target objects within the generated images. [Background technology]

[0003] Background information Isolation and subsequent expansion of a single cell from a cultured population establishes monoclonality, which is often considered an essential step in developing high-quality cell lines. This procedure is intended to minimize or eliminate genomic and phenotypic heterogeneity in an attempt to maximize cell line uniformity. For example, a newly genomically engineered cell population may contain a mixture of cells with diverse alleles, zygosity, and epigenetic characteristics. Therefore, a homogeneous cell line can only be re-established by ensuring that all cells in the population are derived from a single ancestral cell isolated downstream of a variation-prone event. This step is referred to as monoclonalization.

[0004] One example of a cell culture process where cloning is often considered critical is that of human induced pluripotent stem cells (iPSCs). This cell type's ability to self-renew indefinitely and differentiate through any lineage makes it highly promising for modeling disease states in vitro, enabling non-invasive genetic association studies, and especially those related to drug response. Such efforts necessarily require large population-level cohorts. Thus, the throughput of cell line derivation is the greatest limiting factor in unlocking the vast promise that iPSC technology holds for fields including functional genomics and precision medicine. The iPSC reprogramming process imposes significant stress on cells, resulting in highly heterogeneous populations with respect to variables such as the residual load of the reprogramming viral vector and induced chromosomal aberrations, necessitating the need for cloning. Although fully automated methods for iPSC generation have been described, the need for a cloning workflow remains, especially when using viral vectors for iPSC generation. This step has historically been a significant bottleneck in automated, high-throughput iPSC derivation, making this cell type an attractive case for investigating single-cloning methodologies.

[0005] Isolation of single cells is typically achieved via fluorescence-activated cell sorting (FACS), a form of flow cytometry. While this process allows for rapid sorting of individual cells, there are several ways in which undesirable outcomes can result. Sorted cells may not survive, leaving empty wells; alternatively, a malfunction in the sorting process may result in more than one cell being mistakenly transferred to a desired well, resulting in polyclonality. Furthermore, for any given cell type, various morphological or physiological changes may occur during development that alter the quality of the cell line. For example, in the case of stem cells (SCs), there are several known morphological markers that indicate loss of pluripotency, a common defect in newly reprogrammed iPSCs. As a result of these factors, the presence, clonality, and quality of cell aggregates within presumed monoclonal wells must be validated post hoc.

[0006] Currently, the only method for validating monoclonality is by manually inspecting microscopic images taken at regular intervals to track colony growth after sorting. This is extremely time-consuming, and technicians often spend several hours per day classifying wells based on colony presence, clonality, and morphology. More importantly, however, relying on human judgment introduces significant sources of bias and technical variability, especially when such protocols are distributed among multiple researchers and research groups. As a result of this lack of standardization, monoclonalization cannot be reliably upscaled without exacerbating the technical variability of the cell lines. All of these factors make monoclonalization a highly desirable target for automation, which would allow colony selection protocols to be indefinitely scalable and distributed on a large scale with minimal technical variability.

[0007] Deep learning based on the use of convolutional neural networks (CNNs) has enabled significant advances in computer vision over the past few years and has become an invaluable tool for automating the analysis of various types of biomedical images. These techniques have already been applied to numerous processes in SC research, including automating differentiation inference and functional prediction in iPSC-derived cell types. However, CNNs have not yet been used to automatically identify clonality during single-cloning protocols for any cell type.

[0008] Deep learning models often rival or surpass the image analysis performance of human researchers in domain-specific tasks. Dedicated neural network architectures exist for specific tasks, such as image classification and segmentation. Specifically, detection networks trained to detect and localize each instance of a given object class within an image offer clear and promising opportunities for automated verification of monoclonality, which ultimately relies on counting individual cells. Implementations of detection networks in other scientific endeavors have previously been shown to be highly successful. These typically adhere to standardized procedures for training and inference; these involve annotating images with object bounding boxes for training, followed by fitting the labeled data through predefined network architectures such as region-based convolutional neural networks (RCNNs) and you-only-look-once (YOLO).

[0009] Several key nuances inherent in monoclonalization make the task resistant to automation through standardized and widely adopted deep learning practices. For example, identifying monoclonal wells requires enumerating individual starting cells. These typically occupy <0.01% of the well's field of view and are often too small for human researchers to see without manually zooming in on the image at the cell's precise location. Grayscale imaging typically exhibits a large amount of noise, exacerbating this difficulty. Debris particles very often appear subjectively indistinguishable from starting cells, and researchers often rely on information in subsequent images, such as growth, to confirm whether a particular particle is a cell or a non-biological artifact.

[0010] Regardless of the above, verifying clonality necessarily relies on the interaction between images taken at different time points. For example, enumerating individual cells in a day 0 image to validate that the sorting process was successful in isolating exactly one starting cell provides no information about whether the cells will subsequently survive, proliferate, or retain desirable morphological traits. Conversely, verifying that only a single colony is visible upon inspection is not sufficient to confirm monoclonality; multiple starting cells may give rise to a single polyclonal cell mass that superficially resembles a monoclonal colony. In other words, as far as human researchers can assess, no single image can contain all the information necessary to infer the clonality of a well. For this reason, constructing a traditional training set consisting simply of images and their corresponding semantic labels is not feasible.

[0011] iPSCs are an attractive cell source for therapeutic applications, medical research, pharmaceutical testing, etc. However, to meet these and other needs, there remains a long-felt need in the art for an automated system for the rapid generation and isolation of reproducible iPSC cell lines under standard conditions. Summary of the Invention

[0012] The present disclosure provides systems and methods for image analysis based on a computational workflow, referred to herein as "Monoqlo" or the System of the Invention, that integrates trained neural networks. While the systems and methods are applicable to the generation and analysis of many types of images, in one aspect they are useful for identifying and analyzing biological cells, for example, to determine cellular characteristics such as physical attributes, clonality, karyotype, phenotype, abnormalities, and disease states.

[0013] Thus, in one aspect, the present invention provides an imaging system including an imaging device and a controller operatively connected to the imaging device, the controller operable to generate images via the imaging device and to analyze the generated images via a processor. In various aspects, the processor includes functionality for performing one or more of the following operations: i) generating a plurality of time-lapse images of an image area via the imaging device; ii) identifying a target object within the image area of ​​a most recent image of the plurality of time-lapse images; iii) generating a target object image area within the image area of ​​the most recent image that includes the identified target object, the target object area having a perimeter within the image area of ​​the most recent image; iv) using a previous image of the image area to generate a cropped image area sized to the perimeter of the target object image area; v) generating a location region for the cropped image area within the image area of ​​the most recent image; and, optionally, vi) analyzing the location region of the most recent image.

[0014] In another aspect, the present invention provides a method for performing image analysis. The method includes identifying and optionally analyzing a target object in an image using a system of the present invention. In some aspects, analyzing the target object includes classifying the target object based on attributes of the target object, such as physical characteristics of the target object, including size and / or shape. In some aspects, the target object is a cell or a cell colony, and the physical attribute is a morphological characteristic of the cell, such as size and / or shape. In some aspects, the attribute is a property of the cell, such as clonality, karyotype, phenotype, abnormality, and / or disease state.

[0015] In yet another embodiment, the present invention provides an automated system for generating iPSCs or generating differentiated cells from iPSCs or SCs. The system includes: a) an induction unit for automatically reprogramming iPSCs or differentiating SCs or iPSCs, the induction unit operable to contact cells with a reprogramming or differentiation factor; b) an imaging system operable to identify iPSCs or differentiated cells, the imaging system including a non-transitory computer-readable medium having instructions for identifying a monoclonal or polyclonal cell population; and, optionally, c) a sorting unit for isolating the identified cells. In some aspects, one or more CNNs are used to process images taken by the imaging system of the cells generated in a) and cultured for a period of time, thereby generating a set of images of the cells and identifying a monoclonal or polyclonal cell population.

[0016] In another embodiment, the present invention provides an automated method for generating iPSCs or generating differentiated cells from iPSCs or SCs. The method includes the steps of: a) generating iPSCs or generating differentiated cells from SCs or iPSCs; b) identifying iPSCs or differentiated cells using an imaging system including a non-transitory computer-readable medium having instructions for identifying a monoclonal or polyclonal cell population; and, optionally, c) isolating the monoclonal or polyclonal cells via a sorting unit. In some aspects, one or more CNNs are used to process images taken by the imaging system of the cells generated in a) and cultured for a period of time, thereby generating a set of images of the cells and identifying a monoclonal or polyclonal cell population.

[0017] In another aspect, the invention provides a non-transitory computer-readable medium having instructions for identifying monoclonal or polyclonal cell populations. In various aspects, the non-transitory computer-readable medium is electronically coupled to an imaging system.

[0018] In yet another aspect, the present invention provides a method for determining the clonality of a cell population, the method comprising: a) culturing cells for a duration of time to generate a cell population; and b) analyzing the cell population over said duration of time using an imaging system electronically linked to a non-transitory computer-readable medium of the present invention, thereby determining whether the cell population is monoclonal or polyclonal.

[0019] The present invention also provides an automated system for analyzing cells or cell populations, comprising: a) a cell culture unit for culturing cells or cell populations; b) an imaging system operable to analyze cells or cell populations, the imaging system comprising a non-transitory computer-readable medium having instructions for identifying morphological characteristics of cells or for identifying monoclonal or polyclonal cell populations; and, optionally, c) a sorting unit for isolating cells of interest from the cell culture unit.

[0020] In yet another aspect, the present invention provides an automated method for analyzing a cell or cell population, comprising: a) culturing the cell or cell population; b) analyzing the cell or cell population using an imaging system comprising a non-transitory computer-readable medium having instructions for trained identification of morphological characteristics of the cells or identification of monoclonal or polyclonal cell populations; and, optionally, c) isolating a cell of interest from the cultured cells.

[0021] In another aspect, the present invention provides a method comprising the steps of: a) culturing cells in a sample well; and b) analyzing the cells using an imaging system of the present invention, wherein the target object is a cell.

[0022] In yet another embodiment, the present invention provides an automated method for generating iPSCs or generating differentiated cells from iPSCs or SCs, the method comprising: a) generating iPSCs or generating differentiated cells from SCs or iPSCs; b) identifying iPSCs or differentiated cells using the imaging system of the present invention, wherein a controller identifies a monoclonal or polyclonal cell population; and optionally c) isolating the monoclonal or polyclonal cells via a sorting unit.

[0023] In another aspect, the present invention provides a method for determining the clonality of a cell population, the method comprising: a) culturing cells for a duration to generate a cell population; and b) analyzing the cell population over said duration using an imaging system of the present invention, wherein a controller identifies monoclonal or polyclonal cell populations, thereby determining whether the cell population is monoclonal or polyclonal.

[0024] In yet another aspect, the present invention provides an automated system for analyzing cells or cell populations, comprising: a) a cell culture unit for culturing cells or cell populations; b) an imaging system of the present invention, wherein the controller is operable to analyze the cells or cell populations by identifying morphological characteristics of the cells or identifying monoclonal or polyclonal cell populations; and, optionally, c) a sorting unit for isolating cells of interest from the cell culture unit.

[0025] In another aspect, the present invention provides an automated method for analyzing a cell or cell population, the method comprising: a) culturing a cell or cell population; b) analyzing the cell or cell population using an imaging system of the present invention, wherein the controller is operable to analyze the cell or cell population by identifying morphological characteristics of the cell or by identifying a monoclonal or polyclonal cell population; and optionally c) isolating a cell of interest from the cultured cells. [The present invention 1001] An imaging system, including: (a) an imaging device; and (b) a controller operatively connected to the imaging device, the controller is operable to generate an image via the imaging device and to analyze the generated image via a processor; The processor: (i) generating a plurality of time-lapse images of an image area via the imaging device; (ii) identifying a target object within the image area of ​​a most recent image of the plurality of time-lapse images; (iii) generating a target object image area within the image area of ​​the most recent image that includes the identified target object, the target object area having a perimeter within the image area of ​​the most recent image; (iv) using a prior image of the image area, cropping the prior image to generate a cropped image area that matches the size of the perimeter of the target object image area; (v) generating a location region for the cropped image area within the image area of ​​the most recent image; and (vi) analyzing the location area of ​​the most recent image. including functionality to controller. [The present invention 1002] The system of the present invention 1001, wherein (i) to (vi) are repeated for each successive image of the plurality of time-lapse images. [The present invention 1003] The system of the present invention 1002, wherein the plurality of time-lapse images comprises more than 10, 100, 1,000, 10,000, 100,000 or more images. [The present invention 1004] The system of the present invention 1002, wherein (i) to (vi) are repeated when only one target object is identified within the image area. [The present invention 1005] The system of the present invention 1001 further comprising identifying a target object within the location region of the most recent image. [The present invention 1006] The system of the present invention 1005 further comprising analyzing the target object. [The present invention 1007] The system of the present invention 1006, wherein analyzing the target object includes classifying the target object based on attributes of the target object. [The present invention 1008] The system of the present invention 1007, wherein the attribute is a physical characteristic of the target object. [The present invention 1009] The system of the present invention 1008, wherein the physical characteristic is a size or a shape. [The present invention 1010] The system of the present invention 1001, wherein (i) to (vi) are performed via one or more convolutional neural networks (CNNs). [The present invention 1011] The system of the present invention 1001, wherein the target object is a cell or a cell colony. [The present invention 1012] The system of the present invention, wherein the cells of the cell colony are monoclonal. [The present invention 1013] A method of performing image analysis, comprising identifying a target object in an image using a system according to any one of the present inventions 1001 to 1012, and optionally analysing the same. [The present invention 1014] The method of claim 1013, wherein analyzing the target object comprises classifying the target object based on attributes of the target object. [The present invention 1015] The method of the present invention 1014, wherein the attribute is a physical characteristic of the target object. [The present invention 1016] The method of claim 1015, wherein the physical characteristic is size or shape. [The present invention 1017] The method of the present invention 1015, wherein the target object is a cell or a cell colony, and the physical attribute is a morphological feature of the cell. [The present invention 1018] 1. An automated system for generating induced pluripotent stem cells (iPSCs) or for generating differentiated cells from iPSCs or stem cells (SCs), comprising: (a) an induction unit for automated reprogramming of iPSCs or differentiation of SCs or iPSCs, the induction unit operable to contact cells with reprogramming or differentiation factors; (b) an imaging system operable to identify iPSCs or differentiated cells, the imaging system comprising a non-transitory computer-readable medium having instructions for identifying a monoclonal or polyclonal cell population; and optionally, (c) A sorting unit for isolating identified cells. [The present invention 1019] The system of the present invention 1018, wherein one or more CNNs are used to process images taken by an imaging system of cells generated in (a) and cultured for a period of time, thereby generating an image set of the cells and identifying monoclonal or polyclonal cell populations. [The present invention 1020] The system of the present invention 1019, wherein the images are processed in a time-lapse manner. [The present invention 1021] The system of the present invention 1020 assigns each of the images a topographical timestamp. [The present invention 1022] The system of the present invention 1021 further comprises a step of categorizing the image set based on morphological characteristics of the cells. [The present invention 1023] The system of the present invention 1022 further comprising classifying the cells as polyclonal or monoclonal based on the categorization. [The present invention 1024] The system of the present invention 1024 further comprising a step of isolating the cells classified as monoclonal via a sorting unit. [The present invention 1025] The system of the present invention 1024, wherein sorting is optionally performed via cell sorting techniques. [The present invention 1026] The system of the present invention 1019, wherein the image set includes more than 1, 10, 100, 1,000, 10,000, 15,000, 20,000, 25,000, 30,000, 50,000, or 100,000 images. [The present invention 1027] 1. An automated method for generating iPSCs or for generating differentiated cells from iPSCs or SCs, comprising the steps of: (a) generating iPSCs or generating differentiated cells from SCs or iPSCs; (b) identifying the iPSCs or differentiated cells using an imaging system comprising a non-transitory computer-readable medium having instructions for identifying a monoclonal or polyclonal cell population; and optionally, (c) isolating said monoclonal or polyclonal cells through a sorting unit. [The present invention 1028] The method of the present invention 1027, wherein one or more CNNs are used to process images taken by an imaging system of cells generated in (a) and cultured for a period of time, thereby generating an image set of the cells and identifying a monoclonal or polyclonal cell population. [The present invention 1029] The method of the present invention 1028, wherein the images are processed in a time-lapse manner. [The present invention 1030] The method of the present invention 1029, wherein each of the images is assigned a topographical timestamp. [The present invention 1031] The method of claim 1030, further comprising categorizing the image set based on morphological characteristics of the cells. [The present invention 1032] The method of claim 1031, further comprising classifying the cells as polyclonal or monoclonal based on the categorization. [The present invention 1033] The method of claim 1032, further comprising the step of isolating the cells classified as monoclonal via a sorting unit. [The present invention 1034] 1027. The method of claim 1027, wherein the sorting is optionally performed via cell sorting techniques. [This invention 1035] The method of the present invention 1027, wherein the image set comprises more than 10,000, 15,000, 20,000, 25,000, or 30,000 images. [The present invention 1036] A non-transitory computer readable medium having instructions for identifying a monoclonal or polyclonal cell population. [This invention 1037] A non-transitory computer readable medium of the present invention 1036 electronically coupled to an imaging system. [The present invention 1038] A non-transitory computer readable medium of the present invention 1037, wherein the instructions define generating a set of images of cultured cells over a duration via an imaging system, the set having a plurality of individual images. [This invention 1039] The non-transitory computer readable medium of the present invention 1038, wherein the image set comprises more than 10,000, 15,000, 20,000, 25,000, or 30,000 images. [The present invention 1040] The non-transitory computer readable medium of the present invention 1038, wherein each image is captured in a chronological fashion and assigned a chronological timestamp. [This invention 1041] A non-transitory computer readable medium of the present invention 1040, wherein the instructions define processing a set of images in chronological order using one or more CNNs. [The present invention 1042] A non-transitory computer readable medium of the present invention 1041, wherein the instructions define categorizing the processed image set based on morphological characteristics of the cells. [This invention 1043] The non-transitory computer readable medium of the present invention 1042, wherein the instructions define classifying cells as polyclonal or monoclonal based on the categorization. [This invention 1044] A non-transitory computer readable medium of the present invention 1043, wherein the instructions define a step of isolating cells classified as monoclonal or polyclonal. [This invention 1045] A method for determining the clonality of a cell population, comprising the steps of: (a) culturing the cells for a duration to generate a cell population; and (b) analyzing the cell population over said duration using an imaging system electronically linked to the non-transitory computer readable medium of any of claims 1036-1044, thereby determining whether the cell population is monoclonal or polyclonal. [The present invention 1046] An automated system for analyzing cells or cell populations, including: (a) a cell culture unit for culturing cells or cell populations; (b) an imaging system operable to analyze the cell or cell population, the imaging system comprising a non-transitory computer-readable medium having instructions for identifying morphological characteristics of the cells or for identifying monoclonal or polyclonal cell populations; and optionally, (c) a sorting unit for isolating cells of interest from said cell culture unit. [This invention 1047] The system of the present invention 1046, wherein one or more CNNs are used to process images taken by an imaging system of cells cultured in (a) and cultured for a duration, thereby generating a time-lapse image set of the cells over time to identify monoclonal or polyclonal cell populations. [This invention 1048] The system of the present invention 1046, wherein one or more CNNs are used to process images taken by an imaging system of cells cultured in (a) and cultured for a duration, thereby producing a time-lapse image set of the cells over time, in which morphological features are identified and analyzed. [This invention 1049] 1. An automated method for analyzing a cell or a population of cells, comprising the steps of: (a) culturing a cell or cell population; (b) analyzing the cells or cell populations using an imaging system comprising a non-transitory computer-readable medium having instructions for trained identification of morphological characteristics of cells or identification of monoclonal or polyclonal cell populations; and optionally, (c) isolating a cell of interest from the cultured cells. [The present invention 1050] 1049. The method of claim 1049, wherein one or more CNNs are used to process images taken by an imaging system of cells cultured in (a) and cultured for a duration, thereby generating a time-lapse image set of the cells over time, and identifying monoclonal or polyclonal cell populations. [This invention 1051] The method of the present invention 1049, wherein one or more CNNs are used to process images taken by an imaging system of cells cultured in (a) and cultured for a duration, thereby generating a set of time-lapse images of the cells over time, in which morphological features are identified and analyzed. [This invention 1052] (a) culturing cells in sample wells; and (b) analyzing the cell using any one of the imaging systems 1001 to 1012 of the present invention, wherein the target object is the cell. A method comprising: [This invention 1053] 1. An automated method for generating iPSCs or for generating differentiated cells from iPSCs or SCs, comprising the steps of: (a) generating iPSCs or generating differentiated cells from SCs or iPSCs; (b) identifying the iPSCs or differentiated cells using the imaging system of any one of claims 1001 to 1012, wherein the controller identifies a monoclonal or polyclonal cell population; and optionally, (c) isolating said monoclonal or polyclonal cells through a sorting unit. [This invention 1054] A method for determining the clonality of a cell population, comprising the steps of: (a) culturing the cells for a duration to generate a cell population; and (b) analyzing the cell population over said duration utilizing an imaging system of any of the inventions 1001-1012, wherein the controller identifies monoclonal or polyclonal cell populations, thereby determining whether the cell population is monoclonal or polyclonal. [This invention 1055] (a) a cell culture unit for culturing cells or cell populations; (b) the imaging system of any of claims 1001-1012, wherein the controller is operable to analyze the cell or cell population by identifying morphological characteristics of the cell or identifying a monoclonal or polyclonal cell population; and optionally, (c) a sorting unit for isolating cells of interest from said cell culture unit; 1. An automated system for analyzing a cell or cell population, comprising: [This invention 1056] 1. An automated method for analyzing a cell or a population of cells, comprising the steps of: (a) culturing a cell or cell population; (b) analyzing the cell or cell population using the imaging system of any of claims 1001-1012, wherein the controller is operable to analyze the cell or cell population by identifying morphological characteristics of the cell or identifying a monoclonal or polyclonal cell population; and optionally, (c) isolating a cell of interest from the cultured cells. [This invention 1057] The system of the present invention 1012, further comprising a cell isolation module for isolating monoclonal cells. [This invention 1058] The system of the present invention 1057, further comprising a protein isolation module for isolating proteins produced by the monoclonal cells. [This invention 1059] The system of the present invention 1025, wherein the cell dispensing technique is FACS. [The present invention 1060] 1034. The method of claim 1034, wherein the cell sorting technique is FACS. [This invention 1061] An automated system for analyzing cells or cell populations, including: (a) a cell culture unit for culturing cells or cell populations; (b) an imaging system operable to analyze the cell or cell population, the imaging system comprising a non-transitory computer-readable medium having instructions for identifying a characteristic of the cell or cell population; and optionally, (c) a sorting unit for isolating cells of interest from said cell culture unit. [This invention 1062] The system of the present invention 1061, wherein one or more CNNs are used to process images taken by an imaging system of cells cultured in (a) and cultured for a period of time, thereby generating a time-lapse image set of the cells over time and identifying characteristics. [This invention 1063] The system of either invention 1061 or 1062, wherein the characteristic is a morphological characteristic, clonality, karyotype, phenotype, abnormality, and / or disease state. [This invention 1064] 1. An automated method for analyzing a cell or a population of cells, comprising the steps of: (a) culturing a cell or cell population; (b) analyzing the cell or cell population using an imaging system comprising a non-transitory computer-readable medium having instructions for performing trained identification of a characteristic of the cell or cell population; and optionally, (c) isolating a cell of interest from the cultured cells. [This invention 1065] The method of the present invention 1064, wherein one or more CNNs are used to process images taken by an imaging system of cells cultured in (a) and cultured for a duration, thereby generating a time-lapse image set of the cells over time and identifying characteristics. [The present invention 1066] The method of any of claims 1064 or 1065, wherein the characteristic is a morphological characteristic, clonality, karyotype, phenotype, abnormality, and / or disease state. [This invention 1067] (a) culturing cells in sample wells; and (b) analyzing the cell using any one of the imaging systems 1001 to 1012 of the present invention, wherein the target object is the cell. A method comprising: [The present invention 1068] The method of claim 1068, wherein the analyzing step includes using one or more CNNs to process images taken by an imaging system of cultured cells over a period of time, thereby generating a time-lapse image set of the cells over time to determine characteristics of the identified cells. [The present invention 1069] The method of any of claims 1067 or 1068, wherein the characteristic is a morphological characteristic, clonality, karyotype, phenotype, abnormality, and / or disease state. [Brief explanation of the drawings]

[0026] [Figure 1A] Figure 1 shows an image of a portion of four CNN modules utilized in one embodiment of the present invention. Shown is a simple schematic of two neural network architectures used for detection and classification tasks. [Figure 1B] 1 shows an image of a portion of four CNN modules utilized in one embodiment of the present invention. Shown is the functionality of each of the three detection modules with representative target data and outputs. [Figure 1C]1 shows images of a portion of four CNN modules utilized in one embodiment of the present invention. Shown are examples of four target morphological classes used in training the morphological classification network of the present invention. [Figure 2]

[0023] Figure 1 is a series of images outlining the automated daily workflow that generates data for training and real-time use in one embodiment of the present invention. After cell deposition via FACs, cells are grown for N days with well-level imaging occurring nightly. N is a variable that depends on cell growth rate and decisions regarding passage timing. [Figure 3]

[0023] Figure 1 is a schematic diagram showing a high-level overview of the design and algorithm logic used in one embodiment of the present invention. The arrows represent the processing order in the algorithm's reverse time-course analysis, starting with the most recent scan. If a colony is detected, an area around the colony is cropped from the previous day's scan, and the image is passed to a local detection model. The process is repeated, with progressively smaller fields of view analyzed. If multiple colonies are detected within a scan, the well is declared polyclonal, and no further scans are analyzed. When the earliest "day 0" scan is reached, the resulting image is passed to a local detection model. The clonality of the well is finally declared based on the number of cells detected. [Figure 4]Figure 4A is an image showing validation results generated by the present invention. Shown is the well-level clonality identification performance of the present framework on real-world production run data. The outer color represents the clonality of the ground-truthed wells, with color meanings as indicated in the legend; the inner color represents the clonality identified by the present invention, and therefore double-colored wells represent errors. Figure 4B is an image showing validation results generated by the present invention. Shown is the class-specific clonality identification performance of the present invention on a manually curated, class-balanced test dataset. Figure 4C is an image showing validation results generated by the present invention. Shown is a summary of the clonality performance of the present invention when the analysis is restricted to morphologically healthy monoclonal wells selected by biologists for further passage. [Figure 5] Figure 5A is a graph providing an overview of the training and performance of a classification model. Shown is a trajectory of the training and validation accuracy of a classification CNN plotted against epochs. Figure 5B is a graph providing an overview of the training and performance of a classification model. Shown is the confusion matrix when a fully trained classification CNN is tested on a held-out validation set. [Figure 6] 1 is a graph showing the relationship between the width of colony bounding boxes predicted by our global detection model and the true width measured by biologists with an image overlay of a scale bar. [Figure 7A] 1 is an image showing an example of a non-biological artifact that can cause a false colony detection by the global detection model of the present invention.The image represents a panoramic view of the image report generated by the present invention. [Figure 7B] 7A and 7B are images showing examples of non-biological artifacts that can cause false colony detection by the global detection model of the present invention. The images represent a zoomed version of the same image report shown in FIG. 7A. [Figure 8] 10 is an image showing an example of overlapping colony reporting by our local detection model, where only a single colony is present after ground truthing. [Figure 9] 10 is an image showing an example of overlapping colony reporting by our local detection model, where only a single colony is present after ground truthing. [Figure 10] 10 is an image showing an example of overlapping colony reporting by our local detection model, where only a single colony is present after ground truthing. [Figure 11] This image illustrates the concept of "colony splitting," whereby an apparently single colony was revealed during reverse time-lapse analysis to have originated from multiple colonies that eventually merged. [Figure 12] 1 is a series of graphs depicting the gating strategy utilized during FACS sorting monocloning of iPSCs in one embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0027] Detailed Description of the Invention The present invention is based on an innovative system and method for image analysis.Before describing the composition and method of the present invention, it should be understood that the present invention is not limited to the specific system, method, and / or experimental conditions described herein; because such systems, methods, and conditions can vary.It should also be understood that the terminology used herein is only intended to describe specific embodiments and is not intended to limit the present invention, since the present invention is limited only by the scope of the appended claims.

[0028] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly indicates otherwise. Thus, for example, a reference to "the system" includes one or more systems, a reference to "the method" includes one or more methods and / or steps of the type described herein, etc., as would be apparent to those skilled in the art upon reading this disclosure.

[0029] The present disclosure provides imaging systems and methods for analyzing imaged objects that utilize a computational workflow that integrates multiple CNNs. In some aspects, the invention is based on a system and computational design that overcomes known difficulties by leveraging the chronological directionality inherent in cell culture processes. The system and computational methodology described herein, called Monoqlo, integrates multiple CNNs, each with its own "modular" functionality.

[0030] The present invention encompasses a highly scalable framework capable of analyzing datasets of over 1,000, 10,000, 50,000, 100,000, 500,000, or 1,000,000 images within a manageable time frame of less than 10, 9, 8, 7, 6, 5, 4, 3, 2, or 1 hour. It will be appreciated that the functionality described herein may be applied to any number of conventional imagers. As detailed in Example I, the work described herein demonstrates the first application of machine learning to the identification of monoclonal cell lines from brightfield microscopy.

[0031] Although the present disclosure is illustrated with respect to the imaging and analysis of living cells, it will be understood that the systems and methods of the present invention are applicable to the imaging and subsequent analysis of any target object.

[0032] Thus, in one aspect, the present invention provides an imaging system including an imaging device and a controller operatively connected to the imaging device; the controller is operable to generate images via the imaging device and to analyze the generated images via a processor. In various aspects, the processor includes functionality for performing one or more of the following operations: i) generating a plurality of time-lapse images of an image area via the imaging device; ii) identifying a target object within the image area of ​​a most recent image of the plurality of time-lapse images; iii) generating a target object image area within the image area of ​​the most recent image that includes the identified target object, the target object area having a perimeter within the image area of ​​the most recent image; iv) using a previous image of the image area to generate a cropped image area sized to the perimeter of the target object image area; v) generating a location region for the cropped image area within the image area of ​​the most recent image; and, optionally, vi) analyzing the location region of the most recent image.

[0033] The present invention further provides a method of performing image analysis using the system of the present invention, which method comprises identifying, and optionally analyzing, a target object in an image using the system of the present invention.

[0034] In some aspects, steps i) through vi) are repeated for each successive image in the plurality of time-lapse images. In some aspects, steps i) through vi) are repeated when only one target object is identified within the image area.

[0035] As discussed herein, the present invention is capable of analyzing image datasets of various sizes within a manageable time frame, hi some aspects, a dataset, e.g., a plurality of time-lapse images, includes more than 1, 10, 100, 1,000, 10,000, 100,000, 200,000, 300,000, 400,000, 500,000, 600,000, 700,000, 800,000, 900,000, or 1,000,000 or more images.

[0036] As discussed further herein, the functionality described herein may be applied to any number of conventional imagers. As such, generation of images for use with the systems and methodologies of the present invention may be achieved in a variety of ways and analyzed and / or processed utilizing the functionality described herein. In some aspects, the systems of the present invention include one or more imaging devices operatively linked to a processor and / or other robotic platform components, such as a cell sorting unit, a cell culture unit, an optical analyzer or assembly, a cell reprogramming or differentiation unit, and a cryopreservation unit. As used herein, an imaging device includes any device or detector capable of capturing an image, including, but not limited to, a camera, a microscope, a CCD camera, a photodiode, a photomultiplier tube, and a laser scanner.

[0037] In various aspects, the system includes functionality for identifying a target object within a location region of a recent image and analyzing the target object.

[0038] In some aspects, analyzing the target object includes classifying the target object based on attributes of the target object, which may include physical characteristics of the target object, such as size, shape, and / or color.

[0039] In some aspects, the target object is a cell or cell colony, and the attribute is a physical attribute, including morphological characteristics of the cell, such as size and / or shape, hi some aspects, the attribute is a characteristic of the cell or cell colony, such as clonality, karyotype, phenotype, abnormality, and / or disease state.

[0040] In various aspects, the systems and methods of the present invention integrate neural networks that may be trained for the analysis and / or classification of specific types of target objects. An overview of the laboratory automation workflow that generates the data is shown in Figures 2 and 3, respectively. The algorithm processes images of image areas that typically contain target objects in a reverse-chronological manner. That is, for each image area, the algorithm begins by analyzing the most recently generated scan. In our case, this is an image that has been cropped only to remove the black edges of the image, preserving the entire field of view of the image area. These images are passed to a global detection model; the output of the global detection model is a coordinate vector that bounds the bounding box of the detected target object.

[0041] The algorithm then expands these coordinates until each dimension of the bounding box is twice that of the predicted target object, loads the next most recent image for that image area, and crops the image to be the resulting region. Because physical positioning is preserved between scans, an earlier instantiation of the same target object will be roughly centered in the newly cropped image. This image is then passed to a local detection model; the local detection model reports the bounding box of the earlier target object, indicating its position in the original, uncropped image when summed with the crop coordinates. The algorithm recursively repeats this process until the resulting most recent image is the earliest ("day 0") scan.

[0042] In some aspects, the training set is stratified based on chronological timestamps and zoom and crop levels, and separate neural networks with their own "modular" functionality are trained. First, the term "global detection" is assigned to the task of detecting the presence or absence of a target object within an image area. Second, the task of detecting target objects within crops of various image areas at various zoom factors is called "local detection." Third, the task of enumerating individual target objects within a fully zoomed crop is called "single-cell detection." We aimed to achieve all three of these tasks through the use of the RetinaNet™ detection architecture with focal loss (Lin et al., In Proceedings of the IEEE International Conference on Computer Vision; pp. 2980-2988 (2017)). Furthermore, it was believed that a model for categorizing images cropped around colony regions into specific classes based on shape and / or size, such as morphological classes for cells; this is referred to herein as "morphological classification."

[0043] As illustrated in Example I, in various aspects, the systems and methodologies of the present invention identify the clonality of a cell or cell population, for example, a monoclonal or polyclonal cell or cell population. Monoclonalization refers to the isolation and proliferation of a single cell from a cultured population. This is typically performed to minimize the technical variability of cell lines downstream of cell-modifying events, such as reprogramming or gene editing, and to develop monoclonal antibodies. Due to the lack of an automated and standardized method for post-assessment of clonality, methods involving monoclonalization cannot be reliably upscaled without exacerbating the technical variability of cell lines.

[0044] The present invention provides a deep learning workflow that automatically detects the presence of colonies and identifies clonality from cell images. As discussed in Example I, the workflow of the present invention integrates multiple convolutional neural networks and, crucially, leverages the temporal anisotropy of the cell culture process. The systems and methodologies described herein provide a fully scalable and highly interpretable framework capable of analyzing industrial data volumes in less than one hour using commodity hardware. In some aspects, the present invention standardizes the single-cloning process, enabling infinite upscaling of colony selection protocols while minimizing technical variability.

[0045] As such, in another embodiment, the invention provides a non-transitory computer readable medium having instructions for identifying monoclonal or polyclonal cell populations. In various aspects, the non-transitory computer readable medium is electronically coupled to an imaging system.

[0046] In some aspects, the instructions specify generating an image set of cultured cells over a period of time via an imaging system, the set having a plurality of individual images. In some aspects, the individual images are taken in a time-lapse manner and assigned time-lapse timestamps. In some aspects, the instructions further specify processing the image set in time-lapse order using one or more CNNs, and categorizing the processed image set based on morphological characteristics of the cells and further classifying the cells as polyclonal or monoclonal based on the categorization.

[0047] In a related embodiment, the present invention further provides a method for determining the clonality of a cell population, comprising: a) culturing cells for a duration of time to generate a cell population; and b) analyzing the cell population over said duration of time using an imaging system electronically linked to a non-transitory computer-readable medium of the present invention, thereby determining whether the cell population is monoclonal or polyclonal.

[0048] The present invention also provides an automated system for analyzing cells or cell populations. The system includes: a) a cell culture unit for culturing cells or cell populations; b) an imaging system operable to analyze cells or cell populations, the imaging system including a non-transitory computer-readable medium having instructions for identifying morphological characteristics of the cells or identifying monoclonal or polyclonal cell populations; and, optionally, c) a sorting unit for isolating cells of interest from the cell culture unit. In some aspects, one or more CNNs are used to process images taken by the imaging system of the cells generated in (a) and cultured for a duration of time, thereby generating a time-lapse image set of the cells over time to identify a monoclonal or polyclonal cell population. In some aspects, one or more CNNs are used to process images taken by the imaging system of the cells generated in (a) and cultured for a duration of time, thereby generating a time-lapse image set of the cells over time to identify and analyze morphological characteristics.

[0049] In yet another embodiment, the present invention provides an automated method for analyzing cells or cell populations. The method includes: a) culturing the cells or cell populations; b) analyzing the cells or cell populations using an imaging system including a non-transitory computer-readable medium having instructions for trained identification of morphological characteristics of the cells or identification of monoclonal or polyclonal cell populations; and, optionally, c) isolating a cell of interest from the cultured cells. In some aspects, one or more CNNs are used to process images taken by the imaging system of the cells generated in (a) and cultured for a period of time, thereby generating a time-lapse image set of the cells over time to identify a monoclonal or polyclonal cell population. In some aspects, one or more CNNs are used to process images taken by the imaging system of the cells generated in (a) and cultured for a period of time, thereby generating a time-lapse image set of the cells over time to identify and analyze morphological characteristics.

[0050] As is apparent from the present disclosure, the present invention is useful in the generation of iPSCs or differentiated cells, where identification and / or classification of monoclonal cell populations is desirable. Thus, in one embodiment, the present invention provides an automated system for generating iPSCs or generating differentiated cells from iPSCs or SCs. The system includes: a) an induction unit for automatically reprogramming iPSCs or differentiating SCs or iPSCs, the induction unit operable to contact cells with a reprogramming factor or differentiation factor; b) an imaging system operable to identify iPSCs or differentiated cells, the imaging system including a non-transitory computer-readable medium having instructions for identifying monoclonal or polyclonal cell populations; and, optionally, c) a sorting unit for isolating the identified cells.

[0051] In another aspect, the present invention provides an automated method for generating iPSCs or generating differentiated cells from iPSCs or SCs, the method comprising: a) generating iPSCs or generating SCs or differentiated cells from iPSCs; b) identifying iPSCs or differentiated cells using an imaging system comprising a non-transitory computer-readable medium having instructions for identifying a monoclonal or polyclonal cell population; and optionally c) isolating the monoclonal or polyclonal cells via a sorting unit.

[0052] In some aspects, one or more CNNs are used to process images taken by an imaging system of the cells generated in a) and cultured for a duration, thereby generating a set of images of the cells to identify monoclonal or polyclonal cell populations.

[0053] In some aspects, cell sorting is achieved by cell sorting or cell sorting techniques, which may optionally include flow cytometry. For example, cells may be sorted using single-cell sorting, fluorescence-activated cell sorting (FACS), and / or magnetic-activated cell sorting (MACS).

[0054] As used herein, "adult" refers to an organism from post-fetal, e.g., neonatal, through terminal stages, and includes, for example, cells obtained from delivered placental tissue, amniotic fluid, and / or umbilical cord blood.

[0055] As used herein, the term "adult differentiated cell" encompasses a wide range of differentiated cell types obtained from adult organisms that are amenable to iPSC generation using the automated system described below. Preferably, the adult differentiated cell is a "fibroblast." Fibroblasts, also known as "fibrocytes" in their less active form, are derived from the mesenchyme. Fibroblast functions include secreting precursors of extracellular matrix components, including collagen. Histologically, fibroblasts are highly branched cells, while fibrocytes are generally smaller and often described as spindle-shaped. Fibroblasts and fibrocytes derived from any tissue may be used as starting material for the automated workflow system of the present invention.

[0056] As used herein, the term "induced pluripotent stem cell" or iPSC means that the stem cell is generated from a differentiated adult cell that has been induced or altered, e.g., reprogrammed, to become a cell capable of differentiating into tissues of all three germ layers, mesoderm, endoderm, and ectoderm, or into the dermal layer. The generated iPSC does not refer to cells, as they are not found in nature.

[0057] As used herein, the term "stem cell" or "undifferentiated cell" refers to a cell in an undifferentiated or partially differentiated state that has the property of self-renewal and the developmental potential to differentiate into multiple cell types, without any specific meaning implied with respect to developmental potential (e.g., totipotent, pluripotent, multipotent, etc.). Stem cells can proliferate and give rise to more such stem cells while maintaining their developmental potential. In theory, self-renewal can occur by either of two main mechanisms. Stem cells can undergo asymmetric differentiation, known as obligatory asymmetrical differentiation; in this case, one daughter cell retains the developmental potential of the parent stem cell, and the other daughter cell expresses some other specific function, phenotype, and / or developmental potential that is somewhat different from the parent cell. The daughter cells themselves can be induced to proliferate and produce progeny; the progeny then differentiate into one or more mature cell types while retaining one or more cells with the developmental potential of the parent. Differentiated cells may be derived from multipotent cells, which themselves may be derived from multipotent cells, and so on. Each of these multipotent cells can be considered a stem cell, but the range of cell types, e.g., developmental potential, that can arise from each such stem cell can vary considerably. Alternatively, a proportion of stem cells within a population may undergo symmetric division into two stem cells, known as stochastic differentiation; thus, some stem cells within the population as a whole are maintained, while other cells within the population give rise only to differentiated progeny. Thus, the term "stem cell" refers to any subset of cells that, under certain circumstances, possess the developmental potential to differentiate into a more specialized or differentiated phenotype and, under certain circumstances, retain the ability to proliferate without substantial differentiation.In some embodiments, the term stem cell generally refers to a naturally occurring parent cell; its descendants (progeny cells) specialize through differentiation, acquiring entirely individual characteristics, for example, as occurs in the gradual diversification of cells and tissues of an embryo, many of which specialize in different directions. Some differentiated cells also have the ability to give rise to cells with greater developmental potential. Such ability can occur naturally or can be artificially induced by treatment with various factors. Cells that begin as stem cells can progress to a differentiated phenotype, but can then be induced to "reverse" and re-express the stem cell phenotype, which is often referred to by those skilled in the art by the terms "dedifferentiation," "reprogramming," or "retrodifferentiation."

[0058] The term "differentiated cell" encompasses any somatic cell that is not pluripotent in its native form; the term pluripotent is as defined herein. Thus, the term "differentiated cell" also encompasses partially differentiated cells, such as multipotent cells, partially reprogrammed stable non-pluripotent cells, or partially differentiated cells produced using any of the compositions and methods described herein. In some embodiments, a differentiated cell is a cell that is a stable intermediate cell, such as a partially reprogrammed non-pluripotent cell. Transition of differentiated cells (including partially reprogrammed stable non-pluripotent cell intermediates) to pluripotency requires a reprogramming stimulus that exceeds the stimulus that leads to partial loss of differentiation characteristics when placed in culture. Reprogrammed cells, and in some embodiments, partially reprogrammed cells, also have the property of being able to be passaged for extended periods without loss of growth potential, relative to parental cells, which have lower developmental potential and generally can only divide a limited number of times in culture. In some embodiments, the term "differentiated cell" also refers to a cell of a more specialized cell type (e.g., with less developmental potential) that has been derived from a cell of a less specialized cell type (e.g., with more developmental potential) (e.g., from an undifferentiated cell or a reprogrammed cell) through a cell differentiation process.

[0059] As used herein, the term "reprogramming" refers to a process that reverses the developmental potential of a cell or population of cells (e.g., somatic cells). In other words, reprogramming refers to a process that brings a cell to a state of greater developmental potential, e.g., back to a less differentiated state. Reprogrammed cells may be either partially differentiated or terminally differentiated prior to reprogramming. In some embodiments of the aspects described herein, reprogramming encompasses fully or partially reversing the differentiation state, e.g., increasing the developmental potential of the cell, to a differentiation state of the cell that has a pluripotent state. In some embodiments, reprogramming encompasses bringing a somatic cell to a pluripotent state, e.g., such that the cell has the developmental potential of an embryonic stem cell, e.g., an embryonic stem cell phenotype. In some embodiments, reprogramming also encompasses partially reversing the differentiation state or partially increasing the developmental potential of a cell, such as a somatic or unipotent cell, to a multipotent state. Reprogramming also encompasses partially reversing the differentiation state of a cell, such as a somatic cell or unipotent cell, to a state that is more susceptible to full reprogramming to a pluripotent state when subjected to additional manipulations as described herein. As a result of such manipulation, certain genes whose expression contributes to maintaining reprogramming may be endogenously expressed by the cell or its progeny. In certain embodiments, reprogramming a cell using synthetic modified RNA and methods described herein causes the cell to adopt a multipotent state (e.g., the cell becomes a multipotent cell). In some embodiments, reprogramming a cell (e.g., a somatic cell) using synthetic modified RNA and methods described herein causes the cell to adopt a pluripotent-like state or an embryonic stem cell phenotype. The resulting cells are referred to herein as "reprogrammed cells," "somatic pluripotent cells," and "RNA-induced somatic pluripotent cells."The term "partially reprogrammed somatic cell," as referred to herein, refers to a cell that has been reprogrammed from a cell with less developmental potential by the methods disclosed herein, such that the cell has been reprogrammed to a non-pluripotent and stable intermediate state rather than being fully reprogrammed to a pluripotent state. Such partially reprogrammed cells may have less developmental potential than pluripotent cells but more than multipotent cells; as those terms are defined herein. A partially reprogrammed cell can, for example, differentiate into one or two of the three germ layers, but not all three germ layers.

[0060] As used herein, the term "reprogramming factor" refers to a developmental potential modifier, such as a gene, protein, RNA, DNA, or small molecule, whose expression contributes to the reprogramming of cells, such as somatic cells, to a less differentiated or undifferentiated state, such as a cell in a pluripotent state or a partially pluripotent state; the term developmental potential modifier is as defined herein. Reprogramming factors can be transcription factors that can reprogram cells to a pluripotent state, such as SOX2, OCT3 / 4, KLF4, NANOG, LIN-28, and c-MYC, and also include any gene, protein, RNA, or small molecule that can be substituted for one or more of these in methods for reprogramming cells in vitro. In some embodiments, the exogenous expression of reprogramming factors using the synthetic modified RNA and methods described herein induces the endogenous expression of one or more reprogramming factors, such that the exogenous expression of one or more reprogramming factors is no longer required to stably maintain a cell in a reprogrammed or partially reprogrammed state.

[0061] As used herein, the term "differentiation factor" refers to a developmental potential modifier, such as a protein, RNA, or small molecule, that induces a cell to differentiate into a desired cell type; for example, a differentiation factor reduces the developmental potential of a cell; the term developmental potential modifier is as defined herein. In some embodiments, a differentiation factor may be, but is not necessarily, a cell type-specific polypeptide. Differentiation into a particular cell type may require the simultaneous and / or sequential expression of more than one differentiation factor. In some aspects described herein, the developmental potential of a cell or cell population is first increased through reprogramming or partial reprogramming with synthetic modified RNAs as described herein, and then the cells or progeny produced by such reprogramming are induced to differentiate to a lower developmental potential by contacting or introducing one or more synthetic modified RNAs encoding the differentiation factors into the cells or progeny.

[0062] In the context of cellular ontogeny, the terms "differentiate" or "differentiating" are relative terms that refer to the developmental process in which a cell has progressed further along the developmental pathway than its immediate precursor cell. Thus, in some embodiments, a reprogrammed cell, as that term is defined herein, may differentiate into a lineage-restricted precursor cell (such as a mesodermal stem cell), which may then differentiate further down the pathway into other types of precursor cells (such as tissue-specific precursors, e.g., cardiomyocyte precursors), and then into a terminally differentiated cell that plays a characteristic role in a given tissue type and may or may not retain the ability to further proliferate.

[0063] The present invention includes systems and processors for performing steps of the disclosed methods and is described in part in terms of functional components and various processing steps. Such functional components and processing steps may be realized by any number of components, operations, and techniques configured to perform the specified functions and achieve various results. For example, various biological samples, biomarkers, elements, materials, computers, data sources, storage systems and media, information collection techniques and processes, data processing standards, statistical analyses, and regression analyses may be utilized in the present invention, which may perform a variety of functions.

[0064] The image analysis method according to various aspects of the present invention may be implemented in any suitable manner, for example, by using a computer program running on a computer system. An exemplary analysis system according to various aspects of the present invention may be implemented in connection with a computer system, which may be a conventional computer system including a processor and random access memory, such as a remotely accessible application server, network server, personal computer, or workstation. The computer system also preferably includes an additional memory device or information storage system, such as a mass storage system, and a user interface, such as a conventional monitor, keyboard, and tracking device. However, the computer system may include any suitable computer system and associated devices and may be configured in any suitable manner. In one embodiment, the computer system includes a stand-alone system. In another embodiment, the computer system is part of a computer network including a server and a database.

[0065] The software necessary to receive, process, and analyze information may be implemented within a single device or multiple devices. The software may be accessible over a network, allowing information storage and processing to occur remotely from the user. Analysis systems and their various components according to various aspects of the present invention provide functions and operations to facilitate image analysis, such as data collection, processing, analysis, classification, and / or reporting. For example, in embodiments of the present disclosure, a computer system executes a computer program that may receive, search, analyze, classify, and / or report information related to images, cells, or cell populations. The computer program may include multiple modules that perform various functions or operations, such as a processing module for processing raw data and generating supplemental data, and an analysis module for analyzing the raw data and supplemental data to generate a quantitative assessment of a target object.

[0066] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. The terms "a" (or "an"), as well as "one or more," and "at least one," can be used interchangeably.

[0067] Furthermore, "and / or" should be taken as specifically disclosing each of the two specified features or components, with or without the other. Thus, the term "and / or" when used in a phrase such as "A and / or B" is intended to include A and B, A or B, A (only), and B (only). Similarly, the term "and / or" when used in a phrase such as "A, B, and / or C" is intended to include A, B, and C; A, B, or C; A or B; A or C; B or C; A and B; A and C; B and C; A (only); B (only); and C (only).

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art of the present invention.For example, The Dictionary of Cell and Molecular Biology (5th edition. JM Lackie ed., 2013), Oxford Dictionary of Biochemistry and Molecular Biology (2nd edition. R. Cammack et al. eds., 2008) and The Concise Dictionary of Biomedicine and Molecular Biology, PS. Juo, (2nd edition. 2002) etc. can provide those skilled in the art with the general definition of some terms used herein.

[0069] Units, prefixes, and symbols are expressed in the form accepted by the International System of Units (SI). Numerical ranges are inclusive of the numbers defining the range. The headings provided herein are not intended to limit the various aspects or embodiments of the invention, which can be grasped by reference to the specification as a whole. Accordingly, the terms defined immediately below are more fully defined by reference to the specification as a whole.

[0070] When embodiments are described with the word "comprising," otherwise similar embodiments described in terms of "consisting of" and / or "consisting essentially of" are also included.

[0071] The following examples are provided to further illustrate the advantages and features of the present invention, but are not intended to limit the scope of the invention. While the examples are typical of those that might be used, other procedures, methodologies, or techniques known to those skilled in the art may alternatively be used. [Example]

[0072] Example I Modular deep learning enables automated identification of monoclonal cell lines This example describes a system and computational method that leverages the temporal anisotropy inherent in cell culture processes. The computational workflow integrates multiple CNNs, each with its own unique "modular" functionality.

[0073] Our system and methodology provide a highly scalable framework that allowed us to analyze datasets spanning tens of thousands of images in less than an hour. This work demonstrates the first application of machine learning to the identification of monoclonal cell lines from brightfield microscopy through a combination of automated stem cell culture and deep learning.

[0074] method Single cloning of hiPSCs Destination plates (PerkinElmer #6005182) were precoated with 17µg of Geltrex™ LDEV-Free, hESC-Qualified, Reduced Growth Factor Basement Membrane Matrix™ (ThermoFisher #A1413302) diluted in 50µL of DMEM / F12 (ThermoFisher #A1413302) for 1 hour in a 37C incubator. After incubation, 150 μL of d0 medium (1x DMEM / F12, 1.5x PSC Freedom™ Supplement (ThermoFisher #A27336SA), 1.5x Antibiotic / Antimycotic (ThermoFisher #15240062), and 15% CloneR™ (Stemcell Technologies #05888) was added to the 50 μL of Geltrex™ + DMEM / F12 in the well and incubated for 1 hour at 37°C. hiPSC colonies maintained on Geltrex™ in Freedom™ PSC Medium (FRD1) (all ThermoFisher) were dissociated with Accutase™ (ThermoFisher #A1110501) for 5–10 minutes at 37°C. The Accutase™ was quenched with Sort™ Buffer (MACS Buffer Miltenyi containing 10% Clone®™), and the cell suspension was pelleted by centrifugation at 130 RCF. Cells were stained with SSEA4-647: 1:100; BD #560219, Tra-1-60-488: 1:100; BD #560173, CD56-V450: 1:100; BD #560360, and CD13-PE: 1:100; BD #555394, then rinsed with a second centrifugation and resuspended in Sort™ Buffer + Propidium Iodide (PI, 1:5000, ThermoFisher #P3566).Cells were then sorted into pre-prepared destination plates using a FACSARIA-IIu™ Cell Sorter (BD Biosciences) with a 100 μm ceramic nozzle at 23 psi sheath pressure. The flow cytometry gating strategy used is summarized in Figure 12. For samples sorted using the WOLFSorter™, SYTOX AADvanced™ Ready Flow™ Reagent (ThermoFisher #R37173) was added to the Sort™ Buffer instead of PI.

[0075] Image acquisition and labeling All images were obtained from a repository of historical data derived from the single cloning step utilized during the iPSC generation process of the NYSCF Global Stem Cell Array®. Previously used to manually verify clonality, these images were automatically generated once per 24 hours from seeding until plate disposal. All scans generated by the Nexcelom™ Celigo cytometer were bright-field images with a resolution of 1 μm per pixel, providing a 7544 x 7544 pixel image after stitching from 16 individual fields of view. We annotated a total of 3,139 images with bounding boxes and object classes. Additionally, 2,224 unannotated images of empty wells were included in the training set as background-only images. During preliminary studies, this was determined to be crucial for reducing the false discovery rate. All annotations were generated in Pascal VOC format using LabelImg™ software (Tzutalin, 2015). The dataset was augmented by applying random flip and rotation transformations to the images (e.g., Perez & Wang, 2017). The morphological criteria required to categorize each object class were specified by a PhD-level biologist specializing in iPSC culture. Annotations were performed by PhD-, MS-, and BS-level technicians, and all annotations were independently confirmed by an additional researcher.

[0076] Training a machine learning model We trained a RetinaNet™ detection model using a Keras RetinaNet™ implementation (github.com / fizyr / keras-retinanet) with a ResNet50™ convolutional backbone (He et al., In Proceedings of the IEEE conference on computer vision and pattern recognition; pp. 770-778 (2016)) without pre-trained weights. Pre-processing involved subtracting the ImageNet™ mean from the image and normalizing pixel intensity values ​​to a range of 0–1. We also implemented a homebrew algorithm to crop the thick black borders around the wells from the image; it removes the outermost lines on each edge of the image and repeats this process until the maximum raw pixel intensity value for that line exceeds 70. Each CNN model was trained for 60 epochs and the weights were saved after each epoch; this allowed the checkpoint with the smallest validation loss to be selected as the final model for use in the Monoqlo framework.

[0077] result Modularity of neural networks We modularized the task of automatically assigning clonality to four distinct functionalities enabled by deep learning (Figure 1). The decision to modularize was based on empirical reasoning made during preliminary investigations. Consistent with the principles of transfer learning, we initially suspected that the feature extraction capabilities of CNNs would be best optimized by combining all image types into a single training set. However, we found that networks trained in this manner performed poorly and frequently failed to distinguish between object classes. In particular, they often reported object types that could not feasibly occur in the images, such as detecting fully developed colonies in images generated immediately after seeding. This indicated that a single model would not perform well across the diverse image magnifications and object classes used during single cloning.

[0078] Instead, we stratified the training set based on chronological timestamps and zoom and crop levels, and then trained four separate neural networks, each with its own "modular" functionality. First, the term "global detection" is assigned to the task of detecting the presence or absence of colonies in full-well images. Second, the task of detecting colonies in cropped images of various well regions at various zoom magnifications is called "local detection." Third, the task of enumerating individual cells in fully zoomed cropped images is called "single-cell detection." We aimed to achieve all three of these tasks through the use of the RetinaNet™ detection architecture with focal loss (Lin et al., In Proceedings of the IEEE International Conference on Computer Vision; pp. 2980-2988 (2017)). Furthermore, for the only fully classification-based task in this work, we desired a model for categorizing images cropped around colony regions into morphological classes; this is referred to herein as "morphological classification" (summarized in Figures 5A-5B). Modularization in this manner allowed us to exploit the temporal directionality of the cell culture process, for example by limiting the detectable object classes to those that could realistically be present in the image based on its scan date.

[0079] Workflow Design Overview We designed a computational workflow that integrates each of the trained neural networks; we call it Monoqlo. The laboratory automation workflow that generates data for use with Monoqlo and the design of Monoqlo itself are outlined in Figures 2 and 3, respectively. The algorithm processes images on a well-by-well basis in a reverse-chronological fashion. That is, for each physical well, the algorithm begins by analyzing the most recently generated scan. In our case, this is an image that has been cropped only to remove the black edges of the image, preserving the entire field of view of the physical well. These images are passed to a global detection model; the output of the global detection model is a coordinate vector that bounds the bounding box of the detected colony.

[0080] The algorithm then expands these coordinates until each dimension of the bounding box is twice that of the predicted colony, loads the next most recent image of the same well, and crops the image to the resulting region. Because the plate orientation and physical positioning are preserved between scans, an earlier instantiation of the same colony is roughly centered within the newly cropped image. This image is then passed to a local detection model, which reports the bounding box of the earlier colony, indicating its position in the original, uncropped image when summed with the crop coordinates. The algorithm recursively repeats this process until the resulting most recent image is the earliest ("day 0") scan, generated within a few hours of sorting. This incremental, iterative aspect of the workflow, along with expanding the crop box dimensions, was found to be essential because non-radial growth and small positional shifts between scans consistently result in small deviations from precise center alignment from day to day. These deviations over a period of several days of imaging add up to a significant offset, so that a simple cut and zoom at the exact center of a late-stage colony rarely yields a field of view where the initiating cell is located.

[0081] Aside from counting individual starting cells, if two or more clearly distinct cell clusters are observed, they are assumed to have originated from two or more cells from the same FACS sorting run, and polyclonality can often be inferred. If either the global or local detection model reports a colony count of >1 at any point during the iterative, backward-looking process, the algorithm accordingly declares the well polyclonal and stops processing further images for that well. Alternatively, if the workflow continues to detect only one colony until the day 0 scan is reached, the resulting image can be zoomed and cropped precisely around the progenitor cell. This image can then be passed to a single-cell detection model, resulting in a count of the number of starting cells. Based on this criterion, the well can ultimately be declared monoclonal or polyclonal.

[0082] Temporal processing logic allows optimization In this case, a given monocloning "run" typically contains 300–900 plate wells, with typically 2–6 runs active at any given time. Scans per well occur daily over 12–30 days, so the average volume of each run at the time of processing by the algorithm is approximately 30,000 images. However, target labels in the monocloning case correspond to individual wells rather than to images. For this reason, a "well knockout" approach was used: if any one of several exclusion criteria is detected by the workflow, the algorithm removes the entire well from the workflow and ignores all subsequent scans for that well. For example, if no object is detected in the most recent scan, the well is reported as empty at the time of analysis, and its preceding characteristics are considered irrelevant. During testing, Monoqlo was run on eight 96-well plates. The average number of empty wells per plate at the time of processing was found to be 73, with a range of 41–92. Thus, for example, in the case where Monoqlo was applied to eight plates on day 15 of the monocloning process, the well knockout approach eliminated the need to process approximately 8,760 (76.8%) of the total 11,400 images based solely on emptyness. Wells found to be polyclonal at any stage of analysis were also excluded from further processing. In the same test run, an average of 11 polyclonal wells were found per plate, and polyclonality was declared after processing an average of 5.73 images. During real-time development, the exclusion criteria were further expanded to also exclude wells found to exhibit morphological markers of differentiation. Given the large datasets requiring daily analysis, our knockout approach offers significant improvements in computational time.

[0083] Neural networks learn to detect colonies and classify their morphology We began by evaluating the learning trajectory of each CNN in each task and benchmarking its prediction performance. In the case of the object detection network, our initial assessment metric was the change in the loss function value when tested on a held-out validation dataset; this dataset represented 20% of our total image set. Accurate accuracy metrics are not automatically generated by the learning algorithm during training of such networks; the model may correctly detect an object even when the coordinates of the labeled and predicted bounding boxes do not match exactly. Instead, we manually assessed the performance of the networks by visually comparing the labels and predictions on validation images with their bounding boxes. From these comparisons, we quantified detection performance based on two metrics: 1) the percentage of labeled objects that were correctly predicted and classified, and 2) the number of false positives, in which the model detected an object when none was present, as a ratio to the total number of images analyzed. Our model validation results are summarized in Figure 4. Furthermore, the true colony width, which the biologists measured using the image scale bar, was highly predicted by the X dimension of the bounding box (Pearson's r(266) = 0.917, p < 2.2e-16) (Figure 6).

[0084] Modular deep learning workflow identifies clonality We benchmarked the effectiveness of Monoqlo as an integrated, modular workflow by first testing its accuracy on a manually curated, class-balanced validation set, followed by post-hoc evaluation of clonality identification performance (independent of morphology) on a raw, unfiltered dataset derived from a real-world monoclonalization run. The curated test set contained 100 wells from each of three classes: empty, monoclonal, and polyclonal. These wells were randomly selected from a historical record of manually classified wells. The imaging date at which processing began for each well was randomly generated from days 8 to 18. We validated the real-world scenario for a monoclonalization run (DMR0001) containing a total of 768 wells over a 19-day time frame, resulting in a data volume of 18,240 images. Manual image review found 561 of these wells to be empty; that is, they contained no indicators of live cells, regardless of remnants of dead colonies, non-biological debris, and other artifacts. Monoqlo correctly rejected 556 of these wells (99.1%). The remaining five empty wells were reported as monoclonal; this likely resulted in a false positive by the global detection model due to unidentified abiotic artifacts that superficially appeared similar to cell colonies (Figure 7). Consequently, Monoqlo identified 194 non-empty wells. This included 115 declared monoclonal wells, of which two and five wells were found to have ground truth classifications of "polyclonal" and "empty," respectively; the remaining 108 wells (93.9%) were consistent with ground truth. Additionally, 61 wells were reported as polyclonal; 57 (93.4%) of these were confirmed by ground truthing, and four were found to be monoclonal. The results of both validation runs are summarized in Figure 5.

[0085] Homegrown programmatic solutions improve deep learning workflows Even when shortcomings of our post-trained CNN may lead to erroneous results, we identified several situations where they could be robustly corrected using simple programmatic logic. Perhaps most notably, we found that the detection CNN often tended to report multiple overlapping colonies in an image region when only a single colony was present in the ground truth (Figure 8). This could be partially mitigated by adjusting the size and distribution of anchor boxes. However, doing so is laborious, can only be done before training the model, and provides an incomplete solution. Instead, our algorithm combines overlapping boxes and considers the resulting box as a single object. In the case of colony detection, this never results in the loss of polyclonal identification. To illustrate, consider the concept of "colony splitting," which arises due to Monoqlo's reverse-chronological approach. Colonies that overlap each other on day N are spatially separated on day N K and grow into a combined mass on day N + K, where K is a variable amount of time that depends on the growth rate and the original separation distance (Figures 9-11). Thus, overlapping object detections can be safely considered as a single object by our algorithm; if they represent multiple colonies, they will later be detected as completely separated from each other in earlier images and therefore declared polyclonal.

[0086] Consideration This research represents the first successful attempt to automate clonality identification using a deep learning object detection approach. It is expected to have the potential to remove significant limitations on scalability in several cell culture domains. This includes the case of iPSC derivation, where single-cell cloning is considered essential for two reasons. First, in the case of viral reprogramming, the residual payload of Sendai virus vectors used to deliver transcription factors to internal cells during reprogramming varies greatly between cells. Second, reprogramming processes often lead to severe chromosomal abnormalities, likely due to stress-induced mitotic disruption. Both of these factors cause severe phenotypic variation, resulting in unpredictable and highly heterogeneous cell lines; this gives rise to the need for single-cell cloning, which has historically been a bottleneck in iPSC generation. Although it has been suggested that the physical single-cell cloning process may impose additional physiological stress on cells, single-cell cloning remains crucial for several use cases. Given the extent to which cohort size determines the feasibility of population studies, removing this bottleneck, as shown in this study, is a major step in fully unlocking the vast research potential of iPSCs.

[0087] Perhaps more importantly, beyond initial derivation, significant efforts are being made to optimize the editing efficiency of CRISPR-Cas9 and other forms of genome engineering in iPSCs, which hold enormous potential for functional annotation of gene variants, disease modeling, and validation of identified polymorphisms in gene association studies. Due to the genomic heterogeneity introduced by the editing process, newly edited populations must be monocloned to ensure that all cells possess the same genotype. While we focused on iPSCs in this study, the same is true for gene editing in all cell types. Thus, genome engineering pipelines represent another critical case where the Monoqlo framework can address a major bottleneck in disease research and therapeutic development.

[0088] It was anticipated that this algorithm could be adapted to any cell type, as long as the cells are imageable and form discrete clonal clusters. As a key example, antibody development is one of the most common cases in which monoclonalization is used due to the epitope specificity of monoclonal antibodies. Many of the cell types most frequently used in antibody development have been successfully detected in microscopic imaging with CNN. Because monoclonal antibodies have become a central component in many drug discovery efforts, the Monoqlo framework may have the potential to provide a valuable tool for the pharmaceutical industry as a whole.

[0089] This work further expands the application of deep learning in iPSC process automation. In efforts to eliminate the need for immunostaining and fluorescence microscopy imaging, which require much greater financial investment and research time, there is significant interest in optimizing CNNs for use with brightfield microscopy. For example, previous attempts have successfully trained deep learning models to predict fluorescent labels from brightfield images alone. This work further demonstrates the predictive power of deep learning for a variety of analytical tasks using simple microscopy images that do not require fluorescent labeling.

[0090] Standard CNN architectures, such as Resnet50™, have previously been shown to be able to train and differentiate differentiated from undifferentiated stem cells, even at early stages in culture. Our classification CNN differs from previous studies in that it further stratifies the training classes rather than adopting a binary "differentiated versus undifferentiated" approach. This approach helped to enhance the robustness of our algorithm when applied to real-world cell culture scenarios, where iPSC colony morphology varies significantly due to factors other than pluripotent state. Additionally, we trained the network on images cropped around distinct, single colonies, rather than on field images containing numerous randomly seeded cell aggregates. In this sense, our training data resembles that utilized when vector-based CNNs are used to distinguish between "healthy" and "unhealthy" colonies. However, this approach requires significant custom preprocessing steps and, crucially, the need to manually crop precise colony regions, limiting its usefulness in real-world automated scenarios. We automated the segmentation step by using a classification network in conjunction with a colony detection model; this allows for fully autonomous deployment in laboratory automation scenarios.

[0091] Shortcomings of this approach have also been recognized. For example, in cases where two or more starting cells appear precisely adjacent to one another in the earliest available scan, the clonal status of the well must be considered ambiguous. This is because it is not possible to determine whether the cells were independently sorted from the source plate or whether a single cell was successfully sorted and subsequently divided in isolation. However, it should be noted that there is a time lag between the time the cells are seeded and the time they attach to the substrate, during which they cannot be imaged. For this reason, the timing window of the first scan is critical. Certain other efforts have attempted to address this ambiguity through the application of fluorescence microscopy to nuclear-stained images, which allow for nuclei segmentation and help resolve the spatial distribution of individual cells. However, this does not completely eliminate the ambiguity; physically adjacent cells, even if apparently distinct, may still have polyclonal origins. A limited number of feasible approaches to address this ambiguity have been suggested. Researchers may wish to simply assume that wells containing multiple cells during the earliest scan are polyclonal. Otherwise, ambiguity could be resolved only by generating images within minutes of seeding. However, optical focusing problems are unavoidable due to the time lag required for cells to attach. Therefore, initiating cells are likely to be invisible at times, making it impossible to reliably verify monoclonality.

[0092] In this study, the 100% detection rate of colonies large enough for passaging suggested the suitability of Monoqlo for deployment as a reliable, fully autonomous system.

[0093] Monoqlo promises to aid researchers in answering several important unanswered questions about the predictive potential of deep neural networks in iPSC research. Studies have shown that deep learning approaches can sometimes distinguish biological groups within images, even when morphological phenotypes were previously unknown or suspected to exist but were not visible even to trained researchers. For example, CNNs have been shown to predict factors such as cardiovascular disease risk, gender, and smoking status from individual retinal images—none of which were previously thought to manifest morphologically in the retina. Furthermore, in the case of iPSCs, deep neural networks have been successfully trained to predict donor identity from images of clinical-grade iPSC-derived retinal pigment epithelium. These findings suggest that previously unidentified predictive markers likely exist in iPSC colony morphology. For example, they may be able to predict, at early stages, whether currently undifferentiated colonies will spontaneously differentiate with better-than-random accuracy. Given the significant costs associated with maintaining cells in culture that may eventually become unusable, successful training of such models would be of great benefit to iPSC derivation.Other potential targets for CNN classification or regression-based prediction include Sendai virus load, future QC pass / fail status, and relative differentiation affinity to specific germ layers.

[0094] Training such models always requires large training volumes. The Monoqlo framework allows for algorithmic segmentation and extraction of colonies from raw datasets, in addition to automatically filtering out images of empty wells, which typically represent the majority of images. In many cases, researchers may also be able to batch-label images based on classifications they assigned to recent images of a given colony or well. Applying a differentiation-identifying classification network allows Monoqlo to retrospectively assign labels such as "will differentiate" or "won't differentiate" to earlier instantiations of colonies. This potentially alleviates the need for the laborious task of manually reviewing and labeling unfiltered image sets, allowing for the partial or complete autonomous generation of large training volumes for future models. As such, our algorithm provides a valuable tool for generating custom datasets for future investigations into the use of deep learning in iPSC research.

[0095] In summary, we have demonstrated a framework for automating the verification of monoclonality under brightfield microscopy using deep learning algorithms with a modular design, requiring relatively little labeling. We further extended the workflow's functionality to include colony morphology classification, demonstrating its potential for autonomous monitoring of monoclonal cell line development and clonal selection in an automated workflow. Monoqlo represents a crucial step toward enabling widespread adoption of high-throughput cell line generation and editing workflows. It eliminates a critical bottleneck in the specific case of iPSC derivation and genome editing, potentially advancing current technology toward the goal of unlimited upscaling and dissemination of pluripotent stem cells for biomedical research applications. Furthermore, as opposed to relying solely on machine learning models to tackle all aspects of a given task, our work serves as a useful example highlighting the benefits of combining the now-well-recognized vast capabilities of convolutional neural networks with human-designed algorithmic solutions.

[0096] Although the invention has been described with reference to the above examples, it will be understood that modifications and variations are encompassed within the spirit and scope of the invention. Accordingly, the invention is limited only by the appended claims.

Claims

1. An imaging system, including: (a) an imaging device; and (b) a controller operatively connected to the imaging device, the controller is operable to generate an image via the imaging device and to analyze the generated image via a processor to determine clonality of the target object; The processor: (i) generating a plurality of time-lapse images of an image area via the imaging device; (ii) identifying a target object within the image area of ​​a most recent image of the plurality of time-lapse images; (iii) generating a target object image area within the image area of ​​the most recent image that includes the identified target object, the target object image area having a perimeter within the image area of ​​the most recent image; (iv) using a prior image of the image area, cropping the prior image to generate a cropped image area that matches the size of the perimeter of the target object image area; (v) generating a location region of the extracted image area within the image area of ​​the most recent image; (vi) analyzing the location area of ​​the most recent image; and (vii) recursively repeating (iv) through (vi) for the plurality of time-lapse images to determine the clonality of the target object. including functionality to controller.

2. 2. The system of claim 1, wherein (i) through (vi) are repeated for each successive image in the plurality of time-lapse images.

3. 3. The system of claim 2, wherein the plurality of time-lapse images comprises more than 10, 100, 1,000, 10,000, 100,000 or more images.

4. 3. The system of claim 2, wherein (i) through (vi) are repeated when only one target object is identified within the image area.

5. 10. The system of claim 1, further comprising identifying a target object within a location region of the most recent image.

6. 6. The system of claim 5, further comprising analyzing the target object.

7. 7. The system of claim 6, wherein analyzing the target object comprises classifying the target object based on attributes of the target object.

8. 8. The system of claim 7, wherein the attribute is a physical characteristic of the target object.

9. 9. The system of claim 8, wherein the physical characteristic is a size or a shape.

10. 10. The system of claim 1, wherein (i) through (vi) are performed via one or more convolutional neural networks (CNNs).

11. 10. The system of claim 1, wherein the target object is a cell or a cell colony.

12. 12. The system of claim 11, wherein the cells of the cell colonies are monoclonal.

13. 13. A computer-implemented method comprising performing image analysis, comprising identifying and optionally analyzing a target object in an image using the system of any one of claims 1 to 12.

14. 14. The method of claim 13, wherein analyzing the target object comprises classifying the target object based on attributes of the target object.

15. 15. The method of claim 14, wherein the attribute is a physical characteristic of the target object.

16. 16. The method of claim 15, wherein the physical characteristic is size or shape.

17. 16. The method of claim 15, wherein the target object is a cell or a cell colony, and the physical attribute is a morphological characteristic of the cell.

18. 1. An automated method for generating iPSCs or for generating differentiated cells from iPSCs or SCs, comprising the steps of: (a) generating iPSCs or generating differentiated cells from SCs or iPSCs; (b) identifying the iPSCs or differentiated cells using the imaging system of any one of claims 1 to 12, wherein the controller identifies a monoclonal or polyclonal cell population; and optionally, (c) isolating said monoclonal or polyclonal cells through a sorting unit.

19. A method for determining the clonality of a cell population, comprising the steps of: (a) culturing the cells for a duration to generate a cell population; and (b) analyzing said cell population over said duration utilizing the imaging system of any one of claims 1 to 12, wherein a controller identifies monoclonal or polyclonal cell populations, thereby determining whether said cell population is monoclonal or polyclonal.

20. (a) a cell culture unit for culturing cells or cell populations; (b) the imaging system of any one of claims 1 to 12, wherein the controller is operable to analyze the cell or cell population by identifying morphological characteristics of the cell or identifying a monoclonal or polyclonal cell population; and optionally, (c) a sorting unit for isolating cells of interest from said cell culture unit; 1. An automated system for analyzing a cell or cell population, comprising:

21. 1. An automated method for analyzing a cell or a population of cells, comprising the steps of: (a) culturing a cell or cell population; (b) analyzing the cell or cell population using the imaging system of any one of claims 1 to 12, wherein the controller is operable to analyze the cell or cell population by identifying morphological characteristics of the cell or identifying a monoclonal or polyclonal cell population; and optionally, (c) isolating a cell of interest from the cultured cells.

22. 13. The system of claim 12, further comprising a cell isolation module for isolating monoclonal cells.

23. 23. The system of claim 22, further comprising a protein isolation module for isolating proteins produced by the monoclonal cells.

24. (a) culturing cells in sample wells; and (b) analyzing the cell using the imaging system of any one of claims 1 to 12, wherein the target object is the cell. A computer-implemented method comprising:

25. 25. The method of claim 24, wherein the analyzing step comprises using one or more CNNs to process images taken by an imaging system of the cultured cells over a duration to thereby generate a time-lapse image set of the cells over time to determine characteristics of the identified cells.

26. 26. The method of any one of claims 24 or 25, wherein the characteristic is a morphological characteristic, clonality, karyotype, phenotype, abnormality, and / or disease state.

Images