CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit under 35 U.S.C. § 119(e) of U.S. Provisional Application No. 61/610,397, filed March 13, 2012, the content of which is incorporated by reference in its entirety into the present disclosure.

BACKGROUND

[0002] There has been an increasing attention on the applications of human embryonic stem (hES) cells, especially in regenerate medicine and disease model for drug developments. With the improving understanding and developing researches, various types of functional tissues, such as skin, nervous tissue, cornea, and hear muscle, have been expected to developed from human stem cells. For instance, neurons and oligodendrocytes can be promoted under the laminin and all-trans-retinoic acid simulation. Differentiated

cardiomocytes have been transplanted and functioned in animal model. In addition, hES cells have been demonstrated as a practical model for anticancer drugs development. The hES cells present almost unlimited applications and opportunities for future researching and biotechnology, however, the labor-intense and delicate hES incubation process mainly limits its developments.

[0003] In a typical process (according to the protocols from National stem cell bank), after hES cell growth for 7—10 days, undifferentiated hES cell colonies are selected manually and are passed to new plates. During this subculture process, determining the quality of hES cell colonies is based on human-experience with bright-field light microscopy. Comparing to the rest steps of incubation, manual selection process plays a critical role in determining the quality of hES cell passages. However, this time-consuming and labor -intense manual selection seems hinder the massive productivity. Furthermore, lack of standard criteria for cell quality may restrict medical applications of hES cells, especially in biotechnology industry.

[0004] In order to overcome this obstruction, certain technologies have been introduced to study cell properties. For instance, laser flow cytometry has been broadly used and can provide comprehensive information. However, the hES cell colonies are squeezed within sheath flow and can break into small pieces in the cytometry pipeline. Additionally, this method is limited due to the the dispersion of the cell colonies that obstruct the subculture process. Sophisticated image analysis systems have also been developed to study cell colonies. In these systems, based on acquired bright field images, methylcellulose labeled colonies can be scored with specific designed software. However, this method is limited by the availability, cost, and time associated with the labeling process.

[0005] Thus, there is a need in the art for non-disruptive methods of determining the viability or quality of cells without the use of a label or a labeling step.

SUMMARY

[0006] The instant application describes methods useful for creating a label-free, noninvasive, non-disruptive platform to monitor the quality of cell colonies. Aspects and methods disclosed include methods for determining the viability of a cell colony comprising obtaining image data of a scattergram of the cell colony; analyzing the image data; and determining the viability status of a cell colony based on said analysis.

[0007] In another aspect, there is a method for determining the differentiation status of a cell colony comprising obtaining image data of a scattergram of a cell colony; analyzing the image data; and determining the differentiation status of a cell colony based on said analysis.

[0008] A colony of cells may be illuminated with coherent light, and the light scattered in the forward direction may be detected and analyzed. The forward-scattered light patterns (scattergrams) may be characterized and associated with previously determined features associated with known organisms. The process may be performed in an automated manner.

[0009] Substantially coherent light from a source such as a semiconductor laser or other laser type may be used to create a scattergram arising from the interaction of the incident light with a colony of cells. An image of the forward-scattered light pattern may be recorded by a camera or similar device, and the image may be characterized by pattern recognition and computer-vision techniques so as to the classify scattergrams arising from the illumination of the colony with laser light.

[0010] In an aspect, a plurality of images of a pre-determined viable, pre-determined nonviable, pre-determined differentiated or pre-determined undifferentiated cell colony may be analyzed so as to determine the identifying characteristics or features extracted from the image by numerical analysis methods, so as to train a characterizing algorithm to identify the viability status or differentiation status of a cell colony for which the viability or

differentiation status is unknown.

[0011] A variety of feature extraction algorithms may be used, either individually or in combination so as to characterize an image. Forward scattergrams of the cell colony may exhibit generally circular symmetry, and algorithms such as Zernike and Chebyshev

(alternatively spelled Tchebichef) moments may be used. As the scattergrams may also exhibit textures, another set of characterizing data may be obtained using Harelick texture features. Accordingly, in one embodiment, the feature vector is extracted with at least one numerical algorithm selected from the group consisting of Zernike, Chebyshev, and Harelick.

[0012] The values of the moment invariants (features) may be represented by a vector of features (scalar values) extracted from the image and then compared with criteria established by training a recognition algorithm with known data sets so as to identify the viability or differentiation status of a cell colony.

BRIEF DESCRIPTION OF THE DRAWINGS

[0013] FIG. 1 depicts a sketch of the optical system.

[0014] FIG. 2 shows the forward-scattering patents of hESCs colonies. A single colony was illuminated with a laser beam and the scattering patent was projected on the detectors. The light intensity of scattering patents created by good colonies showed more homogeneous distribution (a), comparing to the one from bad colonies (b).

[0015] FIG. 3 shows fluorescent images of hESCs colonies. An immunohistological staining assay was applied to determine the qualities of hESCs colonies. Anti-Oct-4 antibody was used to label the self-renewing pluripotent stem cells in the colonies. Results indicated that more Oct-4 marked cells in colonies which were categorized as good hESCs colonies (a) ,than bad hESCs colonies (b).

[0016] FIG. 4 depicts the expression levels of Oct-4/DAPI in hESCs colonies. Colonies categorized into good/bad colonies showed different Oct-4/DAPI expression ratio. Higher Oct-4/DAPI expression ratio indicated a higher percentage of pluripotent stem cells in determined good colonies. [0017] FIG. 5 shows 3-D confocal images of hESCs colonies. The morphology of determined good/bad colonies (a, b) were investigated with laser scanning confocal microscopy.

DETAILED DESCRIPTION

[0018] The practice of the present invention will employ, unless otherwise indicated, conventional techniques of tissue culture, immunology, molecular biology, microbiology, cell biology and recombinant DNA, which are within the skill of the art. See, e.g., Sambrook et al, (1989) Molecular Cloning: A Laboratory Manual, 2nd edition; Ausubel et al, eds. (1987) Current Protocols In Molecular Biology; MacPherson, B.D. Hames and G.R. Taylor eds., (1995) PCR 2: A Practical Approach; Harlow and Lane, eds. (1988) Antibodies, A

Laboratory Manual; Harlow and Lane, eds. (1999) Using Antibodies, a Laboratory Manual; and R.I. Freshney, ed. (1987) Animal Cell Culture.

[0019] All numerical designations, e.g., pH, temperature, time, concentration, and molecular weight, including ranges, are approximations which are varied ( + ) or ( - ) by increments of 1.0 or 0.1, as appropriate. It is to be understood, although not always explicitly stated that all numerical designations are preceded by the term "about". It also is to be understood, although not always explicitly stated, that the reagents described herein are merely exemplary and that equivalents of such are known in the art.

[0020] As used in the specification and claims, the singular form "a," "an" and "the" include plural references unless the context clearly dictates otherwise.

[0021] As used herein, the term "comprising" is intended to mean that the compositions and methods include the recited elements, but do not exclude others. "Consisting essentially of when used to define compositions and methods, shall mean excluding other elements of any essential significance to the combination when used for the intended purpose. Thus, a composition consisting essentially of the elements as defined herein would not exclude trace contaminants or inert carriers. "Consisting of shall mean excluding more than trace elements of other ingredients and substantial method steps. Embodiments defined by each of these transition terms are within the scope of this invention.

[0022] The term "viability status" refers to whether the cell is classified as living or dead. [0023] The term "colony" refers to a collection of cells, usually visible, growing on the surface or within the surface of a substrate. Typically, all the cells within the colony descend from a single cell and are, therefore, genetically identical (except for any spontaneous mutations).

[0024] The term "differentiation status" refers to whether the cell is classified as differentiated or undifferentiated. In developmental biology, cellular differentiation is the process by which a less specialized cell becomes a more specialized cell type. Differentiation occurs numerous times during the development of a multicellular organism as the organism changes from a simple zygote to a complex system of tissues and cell types. An

undifferentiated cell can refer to a pluripotent stem cell (a cell with the potential to form any type of cell found in the mammalian body) or multipotent progenitor cells, which are further specialized. Examples of multipotent progenitor cells include hematopoietic, mesenchymal, epithelial, or muscle stem cells. A pluripotent cell can differentiate into a progenitor cell, and a progenitor cell can undergo further differentiation into specific cell types. For example, hamatopoietic stem cells can differentiate into multiple cell types such as, for example, red blood cells, white blood cells, and platelets. Mesenchymal stem cells can differentiate into stromal cells, fat cells, and types of bone cells. Therefore, the term "differentiation status" refers to the point at which the cell is at on the differentiation pathway.

[0025] As used herein, a "pluripotent cell" broadly refers to stem cells with similar functional and phenotypic properties to embryonic stem cells with respect to the ability for self-renewal and pluripotency (i.e., the ability to differentiate into cells of multiple lineages). Pluripotent cells refer to cells both of embryonic and non-embryonic origin. For example, pluripotent cells includes Induced Pluripotent Stem Cells (iPSCs).

[0026] An "induced pluripotent stem cell" or "iPSC" or "iPS cell" refers to an artificially derived stem cell from a non-pluripotent cell, typically an adult somatic cell, produced by inducing expression of one or more reprogramming genes or corresponding proteins or R As. Such stem cell specific genes include, but are not limited to, the family of octamer transcription factors, i.e. Oct-3/4; the family of Sox genes, i.e. Soxl, Sox2, Sox3, Sox 15 and Sox 18; the family of Klf genes, i.e. Klfl, Klf2, Klf4 and Klf5; the family of Myc genes, i.e. c-myc and L-myc; the family of Nanog genes, i.e. OCT4, NANOG and REX1; or LIN28. Examples of iPSCs and methods of preparing them are described in Takahashi et al, (2007) Cell. 131(5):861-72; Takahashi & Yamanaka (2006) Cell 126:663-76; Okita et al, (2007) Nature 448:260-262; Yu et al, (2007) Science 318(5858): 1917-20; and Nakagawa et al, (2008) Nat. Biotechnol. 26(l): 101-6.

[0027] In pattern recognition and machine learning, a "feature vector" is an n-dimensional vector of numerical features that represent some object. Many algorithms in machine learning require a numerical representation of objects, since such representations facilitate processing and statistical analysis. When representing images, the feature values might correspond to the pixels of an image, when representing texts perhaps to term occurrence frequencies.

[0028] One method disclosed herein relates to a method for determining the viability of a cell colony comprising obtaining image data of a scattergram of a cell colony; analyzing the image data; and determining the viability status of a cell colony based on said analysis.

[0029] In another aspect, there is a method for determining the differentiation status of a cell colony comprising obtaining image data of a scattergram of a cell colony; analyzing the image data; and determining the differentiation status of a cell colony based on said analysis.

[0030] In one embodiment, the methods disclosed herein further comprise determining a centroid position of the colony to be analyzed prior to obtaining the image data. The step of normalizing and centering the image may further include the steps of imaging the cultured area so as to identify colonies suitable to produce scattergram data. Such colonies may be identified on the basis of size and optical properties in either transmission or reflection. The coordinates of the centroid of each suitable colony are measured such that selected colonies may be translated by the apparatus such that the centroid of the selected colony is disposed so as to coincide with the centroid of the illuminating beam. In a related embodiment, the positioning the colony further comprises scanning a region of the substrate to identify a position of one or more colonies and adjusting the relative orientation of the coherent light source and the colony such that a centroid of a coherent light beam substantially coincides the the centroid of the position of the colony. A cell colony may be disposed so as to be illuminated by an optical radiation source, where the optical wavelength regime may be one or more wavelengths in the range of about 300 nm to about 800 nm. The optical radiation may be generated by a laser source, which may be a semiconductor laser, a gas laser, or the like. Laser light sources are known to have a coherent radiation characteristic. The coherence of each type of laser may be different, and may vary with such parameters as the laser current. For this reason, it should be understood that the term "coherent" light source encompasses a substantially coherent light source. In a related embodiment, adjusting the relative orientation is performed by a two-axis positioner. A two-axis positioner allows a cell colony to be positioned in the XY plane.

[0031] The laser beam may be positioned with respect to the colony manually or motorized stages may be employed. Colonies can be located using a laser equipped with a single line projector (Lasiris 501L-635-5 mW, for example) and a line scanner (Hamamatsu 512 pixel, 25 μιη pitch, 2.5 mm length, from Hamamatsu, Bridgewater, N.J., for example).

Alternatively a camera system such as an Alta U260 (Apogee Instruments, Auburn Calif, for example) using reflected light may be used. In another aspect, non-monochromatic light, or monochromatic light at a wavelength different from the scattergram wavelength may be used for cell colony sizing and location purposes.

[0032] Further, the diameter of the light beam may be changed so as to intercept a greater or lesser portion of a cell colony, and the beam intensity may be varied in accordance with the optical transmission losses and the sensitivity of the camera apparatus. Other optical elements including, but not limited to, lenses, polarizing filters, quarter-wave plates, wavelength-dependent filters and the like may be placed in the optical path.

[0033] The system may be automated such that, for example, a user places a sample in sample holder and the system moves the sample (e.g., with an automated x-y stage), illuminates the sample using the optical source, analyzes the scattergram, and tabulates, displays, or otherwise provides the results to the user without the need for manual intervention.

[0034] In another embodiment, the method further comprises culturing a sample of material on a substrate to form a colony; positioning the colony such that light from a coherent light source illuminates the colony; and recording the image of a scattergram formed by the illuminated colony. In one embodiment, the cell colony is comprised of Eukaryotic cells. Eukaryotic cells include plant cells, animal cells and fungal cells, for example. In one embodiment, the eukaryotic cell is an animal cell. In a related embodiment, the cell is a stem cell or an iPS cell. In a further embodiment, the cell is a human cell or a human stem cell. In yet a further embodiment, the human stem cell is a human embryonic stem cell. The growth conditions and the substrate may vary, depending on the specific type of the bacteria or other microorganism.

[0035] The cells may be cultured on a substrate appropriate to the specific cell type being employed. The substrate may be transparent or translucent. In one embodiment, the cell colony is in cell culture. Culturing of stem cells or iPS cells can be done through the use of various mediums and techniques developed to culture primate pluripotent stem cells, more specially, embryonic stem cells, as described in U.S. patent publication 2007/0238170 and U.S. patent publication 2003/0211603. For example, hES and iPS cells can be maintained in 80% DMEM (Gibco #10829-018 or #11965-092), 20% defined fetal bovine serum (FBS), and antibiotics. Other factors may be added to culturing media such as non-essential amino acids, L-glutamine, and Pmercaptoethanol. Alternatively, ES cells can be maintained in serum-free medium, made with 80% Knock-Out DMEM (Gibco #10829-018) and 20% serum replacement (Gibco #10828-028). Additionally or alternatively, cells may be cultured in feeder-based systems, utilizing irradiated fibroblasts as feeder cells. In one embodiment, the substrate is one or more of the group consisting of a cell culture dish, matrigel, gelatin, collagen I, collagen IV, fibronectin, laminin, poly-D-lysine, and poly-L-lysine. In one embodiment, the substrate is a cell culture dish. In a related embodiment, the cell culture dish further comprises a feeder layer.

[0036] In one embodiment, the image data is recorded on a photographic film or by an array of photosensitive electronic elements. Forward-scattered light may be recorded by a camera or similar device. The camera may use film, or electronic means such as a charge coupled device (CCD) or complementary metal oxide semiconductor (CMOS), or the like, to record the forward-scattered image on film, in a memory, or similar device. The pattern and intensity of the forward-scattered light (a scattergram) may be analyzed to provide a characterization of the cell colony being illuminated.

[0037] A detection screen may be useful for visualizing the data where the forward- scattered light is caused to impinge on the surface of the screen and is viewed by a camera or equivalent sensing device; however, the forward propagated and forward-scattered light may be caused to impinge directly on a camera or a lens and image plane such that it may be imaged and recorded. [0038] Methods disclosed herein include analyzing the image data by extracting a feature vector from the image. In a related embodiment, extracting a feature vector comprises characterizing the image data using a function having an invariant magnitude in at least one dimension. In one embodiment, the image data is normalized prior to feature vector extraction.

[0039] In an aspect, the scattergram images may be characterized by applying a azimuthally invariant orthogonal moment technique, such as that known as a Zernike moment invariant, to obtain a vector characteristic of the sample. Generally, lower-order Zernike moments quantify low-frequency components (which may be considered as "global characteristics") of an image and higher-order moments represent the high-frequency contents (which may be considered as "fine details"). Therefore, there is always a tradeoff between the desired level of image details that can be analyzed, and the order of the moments to be used. Images may be translated so that the center of the scatter pattern is at the center of the image. To compute the Zernike moments of a given image, the center of the image is taken as the origin and pixel coordinates are mapped to the range of the unit circle.

[0040] The magnitude of a Zernike moment is azimuthally invariant, so that the affect of azimuthal variations of the image are minimized. Other similar analytical techniques, such as discrete Krawtchouk or radial Chebyshev polynomials or continuous pseudo-Zernike polynomials may be used, and may be adapted to similar analytical use.

[0041] As with all digital data processing, the resolution of the image, the granularity of the calculations and the accuracy of the numerical analysis algorithms are chosen as a balance between accuracy, noise generation, memory capacity, computation speed and the like, and differing parametric values and specific analytic techniques may be chosen by persons skilled in the art to perform the functions of the method and system disclosed herein.

[0042] In addition to the substantial azimuthal symmetry, the scattergrams may also exhibit a texture. One method of texture analysis is a so-called Haralick feature analysis. This is a grey-scale co-occurrence matrix (GLCM). Such GLCM analysis may be used to quantify the number of occurrences at various distances and angles of pixel intensity values with respect to each other. Using such analysis, image features as angular second moment, contrast, sum average, sum variance, inverse difference moment, sum of squares (variance), entropy, sum entropy, difference entropy, difference variance, information measure of correlation, and maximum correlation coefficient may be calculated. In one embodiment, extracting a feature vector comprises characterizing the image by computing a gray-level co-occurrence matrix (GLCM). In another embodiment, the feature vector is extracted using a function having an invariant magnitude in at least one dimension and by computing a gray-level co-occurrence matrix (GLCM).

[0043] Selection among extracted scattergram features encompasses tradeoffs between desired properties. For example, a higher order of moment invariant provides more sensitivity but also makes the features more susceptible to noise. Therefore, feature reduction may be performed to select the most distinctive features. Feature reduction may be divided into categories: feature selection, in which features carrying the most information are picked out through some selection scheme, and feature recombination, in which some features are combined (e.g., with different weights) into a new (independent) feature.

[0044] The dimensionality of the feature vector of the Zernike moments obtained may be reduced by techniques such as principal component analysis (PCA), non-linear iterative partial least squares (NIPALS), stepwise discriminant analysis (SDA) or other similar methods in order to plot the data in a two or three dimensional form and to visualize data clusters representing different cell colonies.

[0045] The feature vectors may be clustered by unsupervised machine learning methods such as K-Mean clustering, Ward's hierarchical clustering, Kohonen's self-organizing maps or similar methods. The feature vectors may be also classified by supervised learning methods such as linear or quadratic discriminant analysis (LDA, QDA), neural networks (NNs), or support vector machines (SVM). In one embodiment, the training and analysis is performed by a (SVM).

[0046] SVMs are based on the concept of decision hyperplanes that define decision boundaries. An optimal decision hyperplane may be defined as a decision function with maximal margin between the vectors of two classes. SVMs are a set of related supervised learning methods used for classification and regression. They belong to a family of generalized linear classifiers.

[0047] A property of SVMs is that they simultaneously minimize the empirical

classification error and maximize the geometric margin; hence they are also known as maximum margin classifiers. [0048] Support vector machines map input vectors to a higher dimensional space where a maximal separating hyperplane is constructed. Two locally parallel hyperplanes are constructed on each side of the hyperplane that separates the data. The separating hyperplane is the hyperplane that maximizes the distance between the two locally parallel hyperplanes.

[0049] Using a machine learning technique, which may be a support vector machine (SVM) classifier, decision tree, maximum likelihood classifier, neural networks, or the like, appropriate decision criteria may be developed with respect to the observed features. These features may be embodied in a classification algorithm. The learning process may be either supervised or unsupervised.

[0050] In one embodiment, determining the viability or differentiation status of a cell colony further comprises comparing a feature vector extracted from the image data with a pre-determined data or training set characterizing either a viable or non- viable cell colony. Such methods are similar to supervised classification. In supervised classification, training data containing examples of pre-determined categories are presented to a learning

mechanism, which learns one or more sets of relationships that define each of the known classes. In one embodiment, the pre-determined data set can be established using a viable cell, a non-viable cell, a differentiated cell or an undifferentiated cell. In certain

embodiments, the pre-determined data set is trained using feature vectors extracted from a plurality of viable or non- viable cell colonies or a plurality of differentiated or

undifferentiated cells.

[0051] The training set can be classified as viable, non-viable, differentiated, or

undifferentiated by using techniques well known in the art. There are many well known experimental techniques or commercially available kits useful for classifying a cell as viable or non viable. For example, a LIVE/DEAD® Viability Kit (available commercially from Invirogen), BrdU detection assay, or CellTrace assay (available commercially from invitrogen) can be used to classify cells as viable or non viable. In a LIVE/DEAD ®

Viability Kit, membrane-permeant calcein AM is cleaved by esterases in live cells to yield cytoplasmic green fluorescence, and membrane-impermeant ethidium homodimer-1 labels nucleic acids of membrane-compromised cells with red fluorescence. In a BrdU assay, cells are pulsed with BrdU, and it is incorporated into newly synthesized DNA strands of actively proliferating cells. The incorporation of BrdU into cellular DNA may then be detected using anti-BrdU antibodies, allowing assessment of the population of cells which are synthesizing DNA. CeliTrace calceiii blue AM is a proven short-term, blue-fluorescent tracer for labeling of live cells. The probe possesses acetoxymethyi (AM) esters that allow its passive diffusion across cell membranes. Upon cleavage of the AM esters by intracellular esterases, the molecule thai remains (absorption/emission maxima -360/449 nm) is relatively polar and is retained by cells for several hours. In one embodiment, viability or non-viability is determined by a method selected from the group consisting of Membrane-permeant calcein AM cleavage assay, BrdU detection assay, and viable cell labeling assay. In another embodiment, the ratio of cells positive for viablility to number of cells in the colony is greater than 0.6. Alternatively, the ratio is greater than about 0.7, or greater than about 0.8, or greater than about 0.85.

[0052] Methods for determining whether a cell is differentiated or undifferentiated can be done according to methods known in the art. For example, in some instances, the expression of certain proteins can indicate a differentiation state when the expression of that protein is specific to that differentiation state. Markers useful for identifying stem cells or iPS cells include, for example, Oct-4, alkaline phosphates, SSEA-1, Nanog, Sox2, and STAT3. In one embodiment, the undifferentiated cell marker is selected from the group consisting of Oct-4, alkaline phosphates, SSEA-1, Nanog, Sox2, and STAT3. In another embodiment, the ratio of cells positive for an undifferentiated cell marker to number of cells in the colony is greater than 0.5. Alternatively, the ratio is greater than about 0.6, or greater than about 0.7, or greater than about 0.8, or greater than about 0.9. In a specific embodiment, the ratio of cells positive for an undifferentiated cell marker to number of cells in the colony is about 0.9.

[0053] New data may then be applied to the learning mechanism, which then classifies the new data using the learned relationships. Examples of supervised classification processes include linear regression processes (e.g., multiple linear regression (LR), partial least squares (PLS) regression and principal components regression (PCR)), binary decision trees (e.g., recursive partitioning processes such as CART— classification and regression trees), artificial neural networks such as back propagation networks, discriminant analyses (e.g., Bayesian classifier or Fischer analysis), logistic classifiers, and support vector classifiers (support vector machines).

[0054] One supervised classification method is a recursive partitioning process. Recursive partitioning processes use recursive partitioning trees to classify data derived from unknown samples. Further details about recursive partitioning processes are provided in U.S. Patent Application No. 2002/0138208 Al to Paulse et al, "Method for Analyzing Mass Spectra."

[0055] In other embodiments, the classification models that are created can be formed using unsupervised learning methods. Unsupervised classification attempts to learn classifications based on similarities in the training data set, without pre-classifying the spectra from which the training data set was derived. Unsupervised learning methods include cluster analyses. A cluster analysis attempts to divide the data into "clusters" or groups that ideally should have members that are very similar to each other, and very dissimilar to members of other clusters. Similarity is then measured using some distance metric, which measures the distance between data items, and clusters together data items that are closer to each other. Clustering techniques include the MacQueen's K-means algorithm and the Kohonen's Self-Organizing Map algorithm.

[0056] Learning algorithms asserted for use in classifying biological information are described, for example, in PCT International Publication No. WO 01/31580 (Barnhill et al, "Methods and Devices for Identifying Patterns in Biological Systems and Methods of Use Thereof), U.S. Patent Application No. 2002 0193950 Al (Gavin et al, "Method or analyzing mass spectra"), U.S. Patent Application No. 2003 0004402 Al (Hitt et al, "Process for Discriminating Between Biological States Based on Hidden Patterns from Biological Data"), and U.S. Patent Application No. 2003 0055615 Al (Zhang et al, "Systems and Methods for Processing Biological Expression Data").

[0057] The classification models can be formed on and used on any suitable digital computer. Suitable digital computers include micro, mini, or large computers using any standard or specialized operating system, such as a Unix, Windows. TM. or Linux.TM. based operating system. The digital computer that is used may be physically separate from the device that is used to create the data of interest, or it may be coupled to such device.

[0058] The training data set and the classification models according to embodiments of the invention can be embodied by computer code that is executed or used by a digital computer. The computer code can be stored on any suitable computer readable media including optical or magnetic disks, sticks, tapes, etc., and can be written in any suitable computer

programming language including C, C ⁺⁺ , visual basic, etc. [0059] The minimum diameter of the colony that can be measured is dependent on multiple variables such as, by way of example only, the diameter of the laser, the background noise, and the detection device used. In one embodiment, cell colony is at least about 0.6 mm in diameter. Alternatively, the cell colony is at least about 0.5 mm in diameter, or at least about 0.7 mm in diameter, or at least about 0.8 mm in diameter, or at least about 0.9 mm in diameter, or at least about 1 mm in diameter. In certain embodiments, the minimum diameter of a cell colony that is being analyzed is about 0.5 mm or about 0.6 mm, or about 0.7 mm, or about 0.8 mm, or about 0.9 mm, or about 1 mm.

[0060] In one embodiment, the analysis of the image data comprises determining the height of the colony. In a related embodiment, the cell colony is determined to be viable when the height of the colony is from about 30 μιη to about 70 μιη. In a further embodiment, the cell colony is determined to be viable when the height of the colony is about 50 μιη.

Examples

Example 1: Culturing Human Stem Cells

[0061] A hES cell line H9 from WiCell (Madison, WI, passage 32-60) was cultured followed the protocols provided by national stem cell bank. Undifferentiated hES cells were maintained on feeder layers of mitomycin C treated mouse embryonic fibroblasts (MEFs). Medium composed of DMEM-F-12, Knockout Serum Replacer, basic fibroblast growth factor L-glutamine, and MEM nonessential amino acid. Medium was changed daily and hES cells were manually selected to passage every 7 days.

Example 2: Laser forward-scattering system and image analysis

[0062] In this study, an automated BARDOT system was modified for hES cell colony scanning. This system was composed of a laser diode (0.95mW, 632nm), monochromatic CMOS images sensor, and x-y scanning stage. The laser beam was steered to illuminate on single colony and the created forwardscatter signal was collected and analyzed. To extract the features from scatter patterns, Pseudo-Zernike moments (PZMs) were applied in the current models. Support vector machine (SVM)-based algorithm was employed for images recognition and classification. Example 3: Immunohistological staining

[0063] After cultured in medium for 7 days, hES cells were fixed with 4% wt/vol paraformaldehyde for 20 min at room temperature. Paraformaldehyde was removed with 3 subsequence PBS rinses. Triton-X (0.1%, Sigma) was used to penetrate the cell membranes. Next, 2% BSA was added and the cells were incubated for 40 min to block non-specific binding. Primary antibody anti-OCT-4 (available commercially from Millipore; diluted 1 : 100 from the commercial stock) was used to indentify pluripotent stem cells on colonies. Alexafluor-488 was used for secondary antibody staining. All samples were stained with 4', 6-diamdino-2-phenylindole (DAPI, available commercially from Invitrogen , diluted 1 :3000 from the commercial stock) for quantifying cell numbers.

Example 4: Fluorescent imaging of hES cell colonies

[0064] Fluorescent microscopy was performed to collect images for Oct-4 quantities analysis. Alexafluor-488 was excited at 488 nm and the emission was collected between 500 nm and 560 nm. DAPI staining signals were collected to indentify the cell population size. In this study, images of colonies categorized into two separated groups by BARDOT were collected (N=18, respectively.). Laser scanning confocal microscopy was applied to investigate the surface morphology of good/bad colonies.

Example 5: Determining cell viability and differentiation status from scatter profiles

[0065] In order to eliminate the variations from cell culturing conditions, scatter images collected in this study came from hES cell cultured for 7 days and Forward scattering images of hES cell colonies were displayed in FIG. 2. Comparing to the scatter patterns of bad hES cell colonies, the light intensity of good hES cell colonies distributed more symmetrically. Irregular scatter patterns were observed in most of bad hES cell colonies. For the training of recognition system, 290 scatter patterns were collected to establish the colony database. Each image was represented by 161 selected features with Fisher's criterion. The classification was performed with support vector machine (SVM) - based algorithm. In order to explore the method's capacity to identify the good/bad colonies, 100 pre-scored colonies were classified with pretrained system. The classification accuracy for a good colony was 87%, and for a bad colony was 83%. [0066] In addition to comparing the scoring of human judgment, the classification results were investigated with standard immunohitological staining. Eighteen colonies from each class of hES cells were selected and stained with anti- Oct-4 for identifying pluripotent cells within colonies. DAPI staining was applied to quantify the cell number within a single colony. Both Expression intensities of both Oct-4 and DAPI were acquired with epi- fluorescent microscopy (FIG. 3). The expression ratio of Oct-4 to DAPI indicated higher percentage of pluripotent cells presented in good colonies (FIG. 4).

[0067] The correlation of scattering pattern and colony morphology was investigated with laser scanning confocal microscopy. The profile of good hES colony is about 20 μιη in diameter with 50 μιη height. For bad colonies, the height varied from 30 μιη to 150 μιη

(FIG. 5). The data indicated hES cells arranged as uniform layers within good colonies resulted in homogeneous scatter patterns. Cells stacked into various heights and formed nonuniform folds on bad colony surfaces. The formation of non-uniform scatter patterns might be explained by the non-even spatial cells distribution.

[0068] This study demonstrates the feasibility for determining the quality of hES cell colonies with optical forward-scattering technology. Integrating with SVM machine learning, this technology provides essential information for automatic hES cell

categorization. Without any biochemistry labeling process, the grouping results showed highly correlation dependence to the results of a standard immunoflurescent assay. In addition, the label- free process highly decreases the cost and shortens the time for specimen preparation. Though only two fundamental categories of hES cells database were established, the non-invasive optical system can be applied to various types of stem cell colonies such as cells grown on Matrigel, iPS cells, colony-forming cells, etc. Due to the system simplicity, low-cost and efficient sample preparation process, the hES cell colonies scoring method developed in this study may provide a reference calibrator of cell quality and accelerate the massive productivity in regenerate medicine fields.

[0069] It should be understood that although the present invention has been specifically disclosed by preferred embodiments and optional features, modification, improvement and variation of the inventions embodied therein herein disclosed may be resorted to by those skilled in the art, and that such modifications, improvements and variations are considered to be within the scope of this invention. The materials, methods, and examples provided here are representative of preferred embodiments, are exemplary, and are not intended as limitations on the scope of the invention.

[0070] The invention has been described broadly and generically herein. Each of the narrower species and subgeneric groupings falling within the generic disclosure also form part of the invention. This includes the generic description of the invention with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.

[0071] In addition, where features or aspects of the invention are described in terms of Markush groups, those skilled in the art will recognize that the invention is also thereby described in terms of any individual member or subgroup of members of the Markush group.

[0072] All publications, patent applications, patents, and other references mentioned herein are expressly incorporated by reference in their entirety, to the same extent as if each were incorporated by reference individually. In case of conflict, the present specification, including definitions, will control.