Method and system for dimension reduction

Abstract

A computer-implemented method for dimensionality reduction is provided. The method includes receiving a secondary dataset including data points collected from a secondary sample and calculating a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. The calculated second dimensionality is compatible with the dimensionality-reduced reference dataset. Systems and non-transitory computer-readable storage media configured to perform the methods of the present invention are also provided.

Description

[Background technology]

[0001] Introduction Flow cytometry is a technique used to characterize and frequently sort biological materials, such as cells in a blood sample or particles of interest contained in another type of biological or chemical sample. A flow cytometer typically includes a sample reservoir for receiving a fluid sample, such as a blood sample, and a sheath reservoir containing a sheath fluid. The flow cytometer directs the sheath fluid toward the flow cell while transporting particles (including cells) in the fluid sample as a stream of cells toward the flow cell. To characterize components of the flow stream, light is irradiated onto the flow stream. Variations in the material within the flow stream, such as the form or presence of fluorescent labels, can cause variations in the observed light, enabling characterization and separation. For example, particles, such as molecules in fluid suspension, analyte-bound beads, or individual cells, pass through a detection region where the particles are exposed to excitation light, typically from one or more lasers, and the particles' light scattering and fluorescence properties are measured. Particles or their components are typically labeled with fluorescent dyes for easy detection. By labeling different particles or components with spectrally distinct fluorescent dyes, multiple different particles or components can be detected simultaneously. In some implementations, the analyzer includes multiple detectors, one for each scattering parameter being measured and one or more for each distinct dye being detected. For example, some embodiments include a spectral configuration in which two or more sensors or detectors are used per dye. The resulting data includes the measured signal for each of the light scattering detectors and the fluorescence emission.

[0002] A flow cytometer may further include means for recording measured data and analyzing the data. For example, data storage and analysis may be performed using a computer connected to the detection electronics. For example, data may be stored in a table format, with each row corresponding to the data of one particle and each column corresponding to each measured feature. The use of a standard file format, such as the "FCS" file format, for storing data from a particle analyzer facilitates analysis of the data using separate programs and / or machines. Using current analysis methods, data is typically displayed as a one-dimensional histogram or two-dimensional (2D) plot for ease of visualization, although other methods may be used to visualize multidimensional data.

[0003] For example, parameters measured using a flow cytometer typically include light at the excitation wavelength scattered by particles primarily at narrow angles along the forward direction, called forward scatter (FSC), excitation light scattered by particles in a direction orthogonal to the excitation laser, called side scatter (SSC), and light emitted by fluorescent molecules in one or more detectors that measure signals across a range of spectral wavelengths, or by fluorescent dyes that are primarily detected by that particular detector or detector array. Different cell types can be identified by their light scattering characteristics and fluorescence emissions resulting from labeling various cellular proteins or other components with fluorochrome-conjugated antibodies or other fluorescent probes.

[0004] A typical high-dimensional analysis workflow involves creating or deriving a parameter set that represents cells in a low-dimensional graph by creating smaller parameter sets (most commonly two) that attempt to summarize information from all other parameters. This is known as dimensionality reduction (dim redux). A typical dimensionality reduction workflow assumes that the user concatenates all input samples and then performs dimensionality reduction on the single concatenated dataset. Traditionally, flow cytometry users performing dimensionality reduction, such as tSNE, UMAP, EmbedSOM, TriMap, and PacMAP, are forced to forward concatenate input samples because there is no general solution that allows them to compute embeddings for samples based on previously computed embeddings for different samples. For example, Figure 1 illustrates a typical dimensionality reduction workflow in flow cytometry. As shown in Figure 1, datasets 101a–101c (e.g., in the form of .fcs files) from different samples are concatenated into a single file 102, based on which new parameters 103 are derived. A common follow-up step is to use a clustering algorithm to partition the data into similar and dissimilar cell groups that generally coincide with phenotype, and store the cluster membership number as a derived parameter.

[0005] These aforementioned techniques are generally non-deterministic and produce different results when run multiple times. Cells of the same or similar phenotype are grouped together when the algorithm is run, but they do not always orient in the same way, making direct comparisons difficult, and the same phenotypes are not always assigned the same cluster number. Traditionally, all of the data to be compared must be included in a parameter creation step, which means everything must be processed at once. If additional datasets are generated, the entire process must be restarted. Summary of the Invention

[0006] The inventors have found that conventional methods of dimensionality reduction are impractical because they require collecting all data upfront and also require computing the dimensionality reduction for large input files. This leads to challenges with computational resources and computation time. Furthermore, adding new data to the analysis requires rerunning the entire analysis, thereby exacerbating inefficiencies. The concatenation step can also lose connection to metadata, which then needs to be tracked manually. Furthermore, it has been discovered that there is no general solution that allows users to compute embeddings of samples based on previously computed embeddings of different samples, as opposed to embedding new samples in existing dimensionality reductions. Therefore, an algorithmic technique in which the dimensionality reduction of a new sample is computed based on the existing dimensionality reduction of a different (i.e., reference) sample is desirable. The methods, systems, and computer-readable media of the present invention fulfill this need.

[0007] Aspects of the present invention include computer-implemented methods of dimensionality reduction. The methods include receiving a secondary dataset including data points collected from a secondary sample and calculating a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. The second dimensionality reduction calculated by the subject methods is compatible with the dimensionality-reduced reference dataset. In some cases, the methods include determining data points collected from the reference sample that are closest to each data point collected from the secondary sample with respect to one or more given data parameters. For example, calculating the second dimensionality reduction may include calculating k nearest neighbors in the reference dataset for each data point collected from the secondary sample. The value of k may vary and may range from 1 to m, where m is half the number of data points in the reference sample. In some cases, k ranges from 2 to 5. In some embodiments, the method includes calculating the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, a locality-sensitive hash, a hierarchical navigable small-world tree, an approximate nearest neighbor with random projection trees, a GPU-based KNN search, or a brute-force KNN search. In some embodiments, the method includes calculating a distance (e.g., Manhattan distance, Euclidean distance, Chebyshev distance, Minkowski distance, cosine distance) between each data point collected from the secondary sample and each of the k nearest neighbors. In certain versions, the method includes calculating a second dimensionality reduction based on a weighted average of the positions of the k nearest neighbors. In other cases, such as when categorical parameters are derived, the method includes calculating the second dimensionality reduction based on a weighted majority voting approach. Some embodiments of the method include calculating a weighted average using weights obtained for each of the k nearest neighbors based on distance. In some cases, the secondary dataset includes data points associated with multiple parameters consistent with the reference dataset.The method according to some embodiments includes generating derived parameters of the secondary dataset based on parameters (e.g., fluorescence parameters, scattering parameters, imaging parameters, or categorical parameters) of the dimensionality-reduced reference dataset. In some cases, the derived parameters are fluorescence parameters. In selected versions, generating the derived parameters of the secondary dataset includes linear interpolation.

[0008] Methods according to some embodiments include calculating a quality score designed to indicate whether the data is sufficiently stable over time. In some such cases, the method includes calculating an input quality score indicating the degree to which each data point collected from the secondary sample is related to a data point in the reference dataset. Calculating the input quality score may include, for example, obtaining a normalized average distance to k nearest neighbors of each data point collected from the secondary sample. In additional cases, the method includes calculating an output quality score that measures the separation of the k nearest neighbors in the dimensionality-reduced first dataset. Calculating the output quality score may include, for example, obtaining a normalized average distance of each of the k nearest neighbors in the dimensionality-reduced reference dataset.

[0009] In some cases, the reference dataset and the secondary dataset are composed of flow cytometer data. In selected versions, the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled reference dataset. In certain implementations, the method includes receiving a reference dataset and performing a first dimensionality reduction. Methods according to some embodiments include transforming the reference dataset and the secondary dataset with, for example, a linear function, a logarithmic function, a hyperbolic arcsine function, or a biexponential function. In some cases, the method includes calculating a dimensionality reduction of multiple secondary datasets including data points collected from multiple secondary samples.

[0010]

[0010] Embodiments of the present invention further include systems and non-transitory computer-readable storage media configured to perform the subject methods (e.g., described above and herein). For example, embodiments of the present invention include a processor including a memory operably coupled to the processor, the memory having stored thereon instructions that, when executed by the processor, cause the processor to receive a secondary dataset including data points collected from a secondary sample and calculate a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. The second dimensionality reduction calculated by the processor is compatible with the dimensionality-reduced reference dataset. Similarly, embodiments of the present invention include a non-transitory computer-readable storage medium having stored thereon instructions for dimensionality reduction by a method including receiving a secondary dataset including data points collected from a secondary sample and calculating a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. [Brief explanation of the drawings]

[0011] The invention may be best understood from the following detailed description when read in conjunction with the accompanying drawings, in which:

[0012] [Figure 1] 1 illustrates a conventional dimensionality reduction workflow in flow cytometry. [Figure 2] 1 illustrates a dimensionality reduction workflow according to certain embodiments of the present invention. [Figure 3] 1 shows a conceptual diagram of a dimension reduction method according to an embodiment of the present invention; [Figure 4] 1 illustrates a dimensionality reduction workflow according to certain embodiments of the present invention. [Figure 5] 1 illustrates an exemplary graphical user interface for use during dimensionality reduction, in accordance with certain embodiments of the present invention. [Figure 6] FIG. 1 shows a functional block diagram of a flow cytometer in accordance with certain embodiments. [Figure 7] 1 illustrates a control system in accordance with certain embodiments of the present invention. [Figure 8A] 1 illustrates a schematic diagram of a particle sorter system in accordance with certain embodiments. [Figure 8B] 1 illustrates a schematic diagram of a particle sorter system in accordance with certain embodiments. [Figure 9] 1 illustrates an image-enabled particle sorter in accordance with certain embodiments. [Figure 10] 1 illustrates a block diagram of a computing system in accordance with certain embodiments. [Figure 11A] 10 illustrates an exemplary dimensionality reduction of a first sample. [Figure 11B] 10 illustrates an exemplary dimensionality reduction of a second sample. [Figure 12A] 10 shows an example dimensionality reduction illustrating embedding using overlay. [Figure 12B] 10 shows an example dimensionality reduction illustrating embedding using overlay. [Figure 13A] 10 shows an example dimensionality reduction illustrating embedding using overlay. [Figure 13B] 10 shows an example dimensionality reduction illustrating embedding using overlay. [Figure 14A] We show how input and output quality scores can be utilized to gate events with high quality embedding results. [Figure 14B] We show how input and output quality scores can be utilized to gate events with high quality embedding results. DETAILED DESCRIPTION OF THE INVENTION

[0013] A computer-implemented method for dimensionality reduction is provided. The method includes receiving a secondary dataset including data points collected from a secondary sample and calculating a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. The calculated second dimensionality is compatible with the dimensionality-reduced reference dataset. Systems and non-transitory computer-readable storage media configured to perform the methods of the present invention are also provided.

[0014] Before describing the present invention in more detail, it is to be understood that this invention is not limited to particular embodiments described, as such may, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting, since the scope of the present invention will be limited only by the appended claims.

[0015] Where a range of values ​​is presented, unless the context clearly dictates otherwise, it is understood that each intervening value, to the tenth of the unit of the lower limit, between the upper and lower limits of that range, and any other stated or intervening value in that stated range, is encompassed within the invention. The upper and lower limits of these smaller ranges may independently be included in the smaller ranges and are also encompassed within the invention, subject to any specifically excluded limit in the stated range. Where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the invention.

[0016] Certain ranges are described herein by numerical values ​​preceded by the term "about." The term "about" is used herein to literally support the exact number it precedes, as well as a number that is near or approximately the number preceded by the term. In determining whether a number is near or approximately a specifically stated number, the unstated near or approximately number may be a number that, in the context in which it is presented, represents a substantial equivalent to the specifically stated number.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although any methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present invention, representative exemplary methods and materials are now described.

[0018] All publications and patents cited herein are incorporated by reference to disclose and describe the methods and / or materials for which the publications are cited, as if each individual publication or patent was specifically and individually indicated to be incorporated by reference. The citation of any publication is for its disclosure prior to the filing date and should not be construed as an admission that the present invention is not entitled to antedate such publication by virtue of prior invention. Further, the publication dates provided may be different from the actual publication dates, which may need to be independently confirmed.

[0019] It should be noted that, 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. It should be further noted that the claims may be drafted to exclude any optional element. Accordingly, this statement is intended to serve as a predicate for use of exclusive terminology, such as "solely," "only," and the like, in connection with the recitation of claim elements or the use of a "negative" limitation.

[0020] As will be apparent to those skilled in the art upon reading this disclosure, each of the individual embodiments described and illustrated herein has individual components and features which may be readily separated or combined with the features of any of the other several embodiments without departing from the scope or spirit of the invention. Any recited method can be carried out in the order of events recited or in any other order which is logically possible.

[0021] While the systems and methods are described for grammatical fluidity with functional descriptions, it is to be clearly understood that the claims should not be construed as necessarily limited by "means" or "step" limitation constructions unless expressly formulated under 35 U.S.C. § 112, but rather should be given the full scope of the meaning and equivalents of the definitions provided by the claims under the doctrine of equivalents, and that if a claim is expressly formulated under 35 U.S.C. § 112, then the full statutory equivalents under 35 U.S.C. § 112 should be given.

[0022] Computer implementation of dimensionality reduction As noted above, aspects of the present invention include computer-implemented methods of dimensionality reduction. The term "dimensionality reduction" is referred to in its conventional sense and describes the transformation of data from a high-dimensional space to a relatively low-dimensional space. Typically, and / or ideally, the dimensionality-reduced data (i.e., data in a low-dimensional space) preserves the characteristics of the data in the high-dimensional space such that the characteristics are somehow represented in the dimensionality-reduced data. Methods of the present invention enhance the transformation of data from a high-dimensional space to a relatively low-dimensional space, for example, by improving data quality and providing a means that may notify a user if data quality is insufficient. Furthermore, methods according to some embodiments of the present invention improve computational efficiency by 5% or more, e.g., 10% or more, e.g., 15% or more, e.g., 20% or more, e.g., 25% or more, including 30% or more.

[0023] The methods of the present invention include calculating a second dimensionality reduction of a secondary dataset based on a first dimensionality reduction of a reference dataset containing data points collected from a reference sample. The second dimensionality reduction calculated by the subject method is compatible with the dimensionality-reduced reference dataset. By "compatible," we mean that the second dimensionality reduction allows for meaningful comparison between the secondary dataset and the dimensionality-reduced reference dataset. As discussed above in the introduction section, previous techniques of dimensionality reduction are non-deterministic, i.e., produce different results when performed multiple times. In contrast, the compatibility of the second dimensionality reduction with the dimensionality-reduced reference dataset means that data clusters in the dimensionality-reduced dataset generated by the second dimensionality reduction calculated by the subject method have identifiable characteristics that allow them to be compared with corresponding clusters in the dimensionality-reduced reference dataset, even when the second dimensionality reduction is performed multiple times. For example, cells of the same or similar phenotype grouped together in clusters in the dimensionality-reduced dataset generated by the second dimensionality reduction calculated by the subject method may have the same or similar orientation relative to the corresponding clusters in the dimensionality-reduced reference dataset. Furthermore, cells of the same or similar phenotype grouped together in clusters in the dimensionality-reduced dataset produced by the second dimensionality reduction calculated by the subject method may have the same or similar cluster numbers relative to corresponding clusters in the dimensionality-reduced reference dataset. In some cases, multiple iterations of the second dimensionality reduction of the present invention result in dimensionality-reduced data that differ in cluster-identifying characteristics (e.g., cluster number, orientation, etc.) by 10% or less between iterations, such as 9% or less, for example 8% or less, for example 7% or less, for example 6% or less, for example 5% or less, for example 4% or less, for example 3% or less, for example 2% or less (including 1% or less).

[0024] The dimensionality reduction of the present invention can take any suitable form. In some embodiments, the dimensionality reduction is performed by the t-distribution stochastic neighbor embedding (t-SNE) algorithm. The t-SNE algorithm is described in Laurens van der Maaten & Geoffrey Hinton. Learning Research. (2008), which is incorporated herein by reference. In some embodiments, the dimensionality reduction is performed by the Uniform Manifold Approximation and Projection (UMAP) algorithm. The UMAP algorithm is described in McInnes et al. ARXIV. (2018), which is incorporated herein by reference. In some embodiments, the dimensionality reduction is performed by the TriMap algorithm. The TriMap algorithm is described in Ehsan Amid & Manfred K. Warmuth. ARXIV. (2019), which is incorporated herein by reference. In some embodiments, the dimensionality reduction is performed by the EmbedSOM algorithm. The EmbedSOM algorithm is described, for example, in Kratochvil et al. bioRxiv. (2018), which is incorporated herein by reference. In some cases, the dimensionality reduction is performed by the PacMAP algorithm. The PacMAP algorithm is described, for example, in Wang et al. The Journal of Machine Learning Research. (2021) 22(1):9129-9201, the disclosure of which is incorporated herein by reference. In some cases, the first and second dimensionality reductions are selected from tSNE, UMAP, EmbedSOM, TriMap, and PacMAP, although any dimensionality reduction currently existing or yet to be invented may be adapted for use herein.

[0025] The methods of the present invention include receiving a secondary dataset comprising data points collected from a secondary sample. The secondary dataset may be received from any convenient source. In some embodiments, the secondary dataset is received from a flow cytometer. In such embodiments, the secondary dataset is considered flow cytometer data and may be received from any suitable flow cytometer, including, but not limited to, those described herein.Exemplary flow cytometers include the BD Biosciences FACSCanto™ flow cytometer, BD Biosciences FACSCanto™ II flow cytometer, BD Accuri™ flow cytometer, BD Accuri™ C6 Plus flow cytometer, BD Biosciences FACSCelesta™ flow cytometer, BD Biosciences FACSLyric™ flow cytometer, BD Biosciences FACSVerse™ flow cytometer, BD Biosciences FACSymphony™ flow cytometer, BD Biosciences LSRFortessa™ flow cytometer, BD Biosciences LSRFortessa™ X-20 flow cytometer, BD Biosciences FACSPresto™ flow cytometer, BD Biosciences FACSVia™ flow cytometer, and BD Biosciences FACSCalibur™ cell sorter, BD Biosciences FACSCount™ cell sorter, BD Biosciences FACSLyric™ cell sorter, BD Examples of suitable cell sorters include the BD Biosciences Via™ cell sorter, the BD Biosciences Influx™ cell sorter, the BD Biosciences Jazz™ cell sorter, the BD Biosciences Aria™ cell sorter, the BD Biosciences FACSAria™ II cell sorter, the BD Biosciences FACSAria™ III cell sorter, the BD Biosciences FACSAria™ Fusion cell sorter, and the BD Biosciences FACSMelody™ cell sorter, the BD Biosciences FACSymphony™ S6 cell sorter, and the like. In some cases, the secondary dataset is received from a database. In such cases, the database may be hosted locally, for example, on any of a variety of known or future memory storage devices.Examples include any commonly available random access memory (RAM), magnetic media such as a resident hard disk or tape, optical media such as a read-and-write compact disk, flash memory device, or other memory storage device. Alternatively, the database may be an internet database. Exemplary internet databases include, but are not limited to, the FlowRepository database (flowrepository(dot)org). In some cases where the secondary dataset includes flow cytometer data, a data point within the dataset may be considered an "event." In some embodiments, the secondary dataset is provided as a .fcs file.

[0026] In some cases, the method includes receiving a plurality of secondary datasets. For example, in some cases, the method includes receiving a number of secondary datasets in the range of 2 to 50 (e.g., 2 to 25, including 2 to 10). In some examples, the method includes receiving two or more secondary datasets, e.g., three or more secondary datasets, e.g., four or more secondary datasets, e.g., five or more secondary datasets, e.g., six or more secondary datasets, e.g., seven or more secondary datasets, e.g., eight or more secondary datasets, e.g., nine or more secondary datasets, including ten or more secondary datasets. In some such versions, the method includes computing a dimensionality reduction for the plurality of secondary datasets, which includes data points collected from a plurality of secondary samples.

[0027] Further, the method may include receiving a reference dataset and performing a first dimensionality reduction. In such embodiments, the first dimensionality reduction may be performed according to any suitable method of dimensionality reduction. For example, in some embodiments, the first dimensionality reduction is selected from tSNE, UMAP, EmbedSOM, TriMap, and PacMAP. Generally, the same type of dimensionality reduction is used for the secondary dataset and is adjusted as described below. Alternatively, the reference dataset and the first dimensionality reduction may be stored in a memory operably coupled to a processor configured to perform the subject method. In some cases, the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled reference dataset. The reference dataset may be received from any convenient source, including, but not limited to, those described above with respect to the secondary dataset. In some examples, the method includes receiving multiple reference datasets. For example, in some cases, the method includes receiving several reference datasets in the range of 2 to 50 (e.g., 2 to 25, including 2 to 10).

[0028] In some embodiments, calculating the second dimensionality reduction involves determining the data point collected from the reference sample that is closest to each data point collected from the secondary sample with respect to a given data parameter. "Parameter" refers to the characteristic of the analyte being measured. For example, in embodiments in which the reference dataset and the secondary dataset include flow cytometer data, the parameter may include, for example, forward scatter (FSC), side scatter (SSC), CD3, CD4, CD8, CD25, etc. In some cases, the parameter is a fluorescence parameter, i.e., the characteristic of interest is measured by association with a fluorescent molecule (e.g., a fluorescent dye). The parameters described herein may also be embedding parameters, e.g., parameters related to the type of dimensionality reduction being performed (e.g., t-SNE, UMAP, EmbedSOM, PacMAP, etc.). In select cases, determining the data point collected from the reference sample that is closest to each data point collected from the secondary sample is done with respect to a parameter known to be present in both the reference dataset and the secondary dataset. In some cases, the secondary dataset includes data points associated with multiple parameters that match the reference dataset. In other words, the parameter is common to the two datasets. As an example, if the dimensionality-reduced reference dataset has a parameter set of {CD3, CD4, CD8, CD25, tSNE1, tSNE2} and the secondary dataset has a parameter set of {CD3, CD4, CD8 and CD25}, the four common parameters {CD3, CD4, CD8 and CD25} may be used in the calculation.

[0029] In some cases, computing the second dimensionality reduction includes a proximity search configured to find one or more points in the set that are closest to one or more data points collected from the secondary sample. For example, in some cases, the method includes computing nearest neighbors for each data point collected from the secondary sample. The nearest neighbor problem can be defined as finding the nearest point from S to q, where S is a set of points in space M and q∈M. In select examples, the method includes computing k nearest neighbors in the reference dataset for each data point collected from the secondary sample, where k is a positive integer. The value of k can vary, but in some embodiments, k ranges from 1 to m, where m is equal to half the number of data points in the reference sample. In some cases, k ranges from 2 to 5. In particular cases, k is 1 or more, e.g., 2 or more, e.g., 3 or more, e.g., 4 or more, e.g., 5 or more, e.g., 6 or more, e.g., 7 or more, e.g., 8 or more, e.g., 9 or more, including 10 or more. The manner in which the k nearest neighbors are computed can vary. In some cases, the method includes calculating the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, locality-sensitive hashing, a hierarchical navigable small-world, approximate nearest neighbors with random projection trees, a GPU-based KNN search, or a brute-force KNN search. In some cases, the method includes calculating the k nearest neighbors using a vantage point tree. Vantage point trees are described, for example, in Yianilos, Peter N. Soda. (1993) 93(194):311-21, which is incorporated herein by reference. In select examples, the method includes calculating the k nearest neighbors using a k-dimensional tree (kd-tree). Kd-dimensional trees are described, for example, in Bentley, JL Comm. ACM. (1975) 18(9):509 517, which is incorporated herein by reference. In select examples, the method includes calculating the k nearest neighbors using a ball tree (metric tree).Ball trees are described, for example, in Omohundro, SMFive balltree construction algorithms. (1989), which is incorporated herein by reference. In some cases, the method includes calculating the k nearest neighbors using locality-sensitive hashing. Locality-sensitive hashing is described, for example, in Pauleve et al. Pattern recognition letters. (2010) 31(11):1348-1358, which is incorporated herein by reference. In some cases, the method includes calculating the k nearest neighbors using hierarchical navigable small world (HNSW). Hierarchical navigable small worlds are described, for example, in Malkov et al. IEEE transactions on pattern analysis and machine intelligence. (2018) 42(4):824-836, which is incorporated herein by reference. In some cases, the method includes calculating the k nearest neighbors using approximate nearest neighbors with randomly projected trees. Such trees are described, for example, in Hyvonen et al., In 2016 IEEE International Conference on Big Data (Big Data), pp. 881-888, which is incorporated herein by reference. In some cases, the method includes calculating the k nearest neighbors using a GPU-based KNN search. GPU-based KNN search is described, for example, in Garcia et al., In 2010 IEEE International Conference on Image Processing, pp. 3757-3760, which is incorporated herein by reference. In some cases, the method includes performing a brute-force KNN search.

[0030] The method according to some embodiments further includes calculating a distance between each data point collected from the secondary sample and each of the k nearest neighbors. The calculated distances may vary. In some cases, the distance is the Euclidean distance, i.e., the length of the line segment between two points. In other cases, the distance is the Manhattan distance, where the distance between two points is the sum of the absolute differences of their Cartesian coordinates. In still other cases, the distance is the Chebyshev distance, i.e., the maximum difference between two vectors along any coordinate dimension. In still other cases, the distance is the Minkowski distance, i.e., a generalization of the Euclidean distance and the Manhattan distance. In still other cases, the distance is the cosine distance, i.e., the complement of cosine similarity. In particular examples, the distance is selected from the group consisting of the Euclidean distance, the Manhattan distance, the Chebyshev distance, the Minkowski distance, and the cosine distance.

[0031] In an embodiment, the method includes calculating a second dimensionality reduction based on a weighted average of the positions of the k nearest neighbors. In other words, the method includes taking a weighted average of derived parameters in the reference dataset. In other words, the position of each data point in the secondary dataset in the dimensionally reduced space is defined by the weighted positions of its nearest neighbors from the high-dimensional space in the dimensionally reduced space. The weighted average can be calculated using weights obtained for each of the k nearest neighbors based on distance (e.g., Euclidean distance, Manhattan distance, Chebyshev distance, Minkowski distance, and cosine distance, as desired). In other words, the nearest neighbors are weighted differently in the average depending on their distance. Closer neighbors have a larger weight. The weights used in the weighted average can vary. In some cases, the method includes obtaining the weights as follows:

[0032]

number

[0033] In an additional case, the method includes obtaining the weights as follows:

[0034]

number

[0035] In still a further case, the method includes obtaining the weights as follows:

[0036]

number

[0037] As shown in the formula above, the weight increases as the distance decreases.

[0038] In some embodiments, the method includes creating derived parameters for the secondary dataset based on parameters of the dimensionality-reduced reference dataset. In such cases, the parameters are selected from fluorescence parameters, scattering parameters, imaging parameters, or categorical parameters. In certain implementations, the derived parameters are fluorescence parameters. For example, the computer-implemented methods of the present invention can be used to create a "virtual tube." In other words, assuming a shared parameter backbone exists between the reference dataset / reference sample and the secondary dataset / secondary sample, values ​​of fluorescence parameters not present in the secondary dataset can be estimated based on those parameters in the reference dataset. This allows users to create a "virtual tube" for visualizing the co-expression pattern of markers not present together in a single tube (i.e., dataset). In additional implementations, the derived parameters are scattering parameters. In still further implementations, the derived parameters are imaging parameters. If the parameters are fluorescence parameters, scattering parameters, or imaging parameters, the parameters can be calculated by the weighted average described above. In some embodiments, the parameters are categorical parameters, such as clustering parameters. In such embodiments, the calculation is performed by weighted majority voting. For example, if a data point's seven nearest neighbors belong to cluster {1,1,1,1,2,1,3}, the data point would be assigned to cluster 1 because the majority of its neighbors belong to this cluster. For data points with many nearest neighbors missing from a cluster, the data point may be assigned to cluster -1, an outlier cluster. Alternatively, the method may include placing all data points into a cluster, for example, if the experiment requires classifying all cells. In this case, the method includes assigning the data point to the same cluster as its single nearest neighbor.

[0039] In some cases, creating derived parameters for the secondary dataset includes interpolation. In other words, parameters for data points in the secondary dataset are interpolated based on known data points in the reference dataset and the dimension-reduced reference dataset. In some examples, the interpolation is nearest neighbor interpolation. In other examples, the interpolation is linear interpolation. This process may be used for continuous variables, such as dimension-reduced parameters, for example.

[0040] Some embodiments of the method include calculating a quality score designed to indicate whether the data are sufficiently stable over time. In some cases, a user may consider one or more quality scores to determine the quality of the second dimensionality reduction. For example, the quality score may indicate the degree to which a reference dataset and a secondary dataset can be meaningfully compared. In some cases, the quality score may indicate the precision with which parameters are derived for the secondary dataset. In some embodiments, the method includes calculating an input quality score that indicates the degree to which each data point collected from the secondary sample is related to a data point in the reference dataset. In some cases, the input quality of a given cell is the average distance of the nearest neighbors to the cell, normalized by the number of dimensions used. Simply put, the smaller this number, the better the nearest neighbors represent the cell of interest. Normalization may vary depending on the distance metric used. In embodiments where the distance is Manhattan distance, normalization is performed by dividing by the number of dimensions used in the calculation. In embodiments where the distance is Euclidean distance, normalization may be performed by dividing by the square root of the number of dimensions used in the calculation. In embodiments where the distance is the Chebyshev distance, normalization is not required because the scores are already normalized, and if normalization is used, it may derive scores that are reasonably independent of the number of dimensions and distance metrics used.

[0041] In additional embodiments, the method includes calculating an output quality score that measures the separation of k nearest neighbors in the dimensionality-reduced first dataset. In some cases, calculating the output quality score includes obtaining a normalized average distance for each of the k nearest neighbors in the dimensionality-reduced reference dataset. In other words, the output quality of a given cell is the average normalized distance between nearest neighbor cells in the embedding parameter space. The lower the number, the better. A large number can indicate that the nearest neighbors in the measured parameter space are significantly separated in the dimensionality-reduced space (e.g., tSNE dividing a homogeneous population into separate islands). Normalization can be performed as described above for the input quality score.

[0042] The method may further include transforming the reference dataset and the secondary dataset. Specifically, methods according to some embodiments include transforming parameters in the reference dataset and parameters in the secondary dataset. In some embodiments, the method includes transforming parameters of the reference dataset and parameters of the secondary dataset with a linear function, a logarithmic function, a hyperbolic arcsine function, or a bi-exponential function. In some embodiments, the method includes transforming parameters of the reference dataset and parameters of the secondary dataset with a linear function. In some embodiments, the method includes transforming parameters of the reference dataset and parameters of the secondary dataset with a logarithmic function. In some embodiments, the method includes transforming parameters of the reference dataset and parameters of the secondary dataset with a hyperbolic arcsine function. In some such embodiments, the hyperbolic arcsine function is performed with a modulus factor of 150. In some embodiments, the method includes transforming parameters of the reference dataset and parameters of the secondary dataset with a bi-exponential function.

[0043] 2 illustrates an approach to dimensionality reduction according to an embodiment of the present invention. A first dimensionality reduction is performed on a reference data set 201a containing data points collected from a reference sample. This results in a dimensionally reduced reference data set containing derived parameters 202a. The reference data set 201a and the derived parameters 202a are used to calculate a second dimensionality reduction of a secondary data set 201b, which results in parameters 202b that are compatible with the derived parameters 202a.

[0044] FIG. 3 shows a conceptual diagram of a dimension reduction method according to an embodiment of the present invention. First, a dimension reduction R from dimension D is performed on a reference dataset from a reference sample. The method then involves calculating a dimension reduction R of a secondary dataset that is compatible with the dimension-reduced reference dataset. This involves calculating, for each data point (event e) in the secondary dataset, a NN as the k nearest neighbors of e in the reference dataset using dimension D. In the example of FIG. 3, nearest neighbors nn1, nn2, and nn3 are depicted. The method also calculates the nearest neighbors nn for each of the NN. i About,nn i Distance d as the distance from to e i In the example of FIG. 3, the distance d1 between nn1 and e is 1, the distance d2 between nn2 and e is 2, and the distance d3 between nn3 and e is 4.

[0045] Then, the method calculates the d i , and calculating the input quality score of e as the average of . In other words, the input quality is the normalized average distance of e to its nearest neighbors in D, with closer neighbors being better, i.e., smaller numbers being better. In this example, the input quality score is calculated by ((1 + 2 + 4) / 3) / ND, where ND varies depending on the type of distance metric used. If the distance is Manhattan distance, then ND = ||D||. If the distance is Euclidean distance, then ND = sqrt(||D||). If the distance is Chebyshev distance, then ND = 1. The method then calculates the input quality score of nn i andnn jnn1-nn2|=5.5, |nn2-nn3|=4, and |nn1-nn3|=5.7. The output quality score is calculated as ((5.5+4+5.7) / 3) / NR, where NR varies depending on the type of distance metric used. If the distance is Manhattan distance, NR=||R||. If the distance is Euclidean distance, NR=sqrt(||R||). If the distance is Chebyshev distance, NR=1.

[0046] The data points in the reference dataset identified as nearest neighbors are then used to create derived parameters for the data points in the secondary dataset, for example, using linear interpolation of continuous variables, such as dimensionality reduction parameters. This involves taking a weighted average of the derived parameters in the reference dataset. The weighting is performed by the proximity of a given cell to each of its nearest neighbors. In other words, each d i About w i =1 / (1+d i ) as weight w i In the example of Figure 3, the weight w1 for nn1 is 1 / 2, the weight w2 for nn2 is 1 / 3, and the weight w3 for nn3 is 1 / 5. i For each reduced dimension r of R using as weights, the method involves computing the r value of e as the weighted average of the r values ​​of the nearest neighbors NN.

[0047] 4 illustrates a dimensionality reduction workflow involving derived fluorescence parameters according to certain embodiments of the present invention. As shown in FIG. 4, a reference dataset 401a and a secondary dataset 401b share a shared backbone of parameters / markers D. The reference dataset 401a has parameters / markers R1 specific to the dataset of reference sample S1, and the secondary dataset 401b has parameters / markers R2 specific to the dataset of secondary sample S2. Using methods of the present invention (e.g., described above), a "virtual tube" 402 for the secondary sample can be created, and tube-specific markers R1 are estimated within the secondary dataset.

[0048] In some cases, the method includes receiving input from a user. The input may relate, for example, to which parameters are derived for the secondary dataset and / or the quality score to be calculated. In some cases, the user selects parameters known to be common to the two datasets, i.e., parameters in the secondary dataset that match the reference dataset. In particular cases, the method includes inputting a value of k for the k-nearest neighbor search described above. In selection embodiments, the method includes selecting a distance metric (e.g., Manhattan distance, Euclidean distance, Chebyshev distance, Minkowski distance, or cosine distance) to be used in the calculation. In embodiments, the method includes selecting a number of data points in the reference dataset to be used in the calculation. In additional cases, the user may select whether an input quality score is calculated. In yet additional cases, the user may select whether an output quality score is calculated. These selections may be achieved, for example, via an input manager. In some cases, the input manager is operably coupled to a graphical user interface through which input is entered. In particular examples, input is entered (e.g., remotely) on an internet website menu interface and communicated to the input manager via the internet or a local area network. In some embodiments, the input manager is operably coupled to one or more searchable databases (e.g., catalogs) of parameters. In particular examples, the input manager includes a database of parameters. All or a portion of each database parameter may be displayed on a graphical user interface, such as a list, drop-down menu, or other configuration (e.g., tile). For example, the graphical user interface may display a list of parameters simultaneously (i.e., on a single screen) or may contain one or more drop-down menus.In some cases, the graphical user interface includes multiple drop-down menus: a first menu for selecting parameters in the reference dataset and a second menu for selecting target parameters in the secondary dataset to be derived and embedded. In other embodiments, information can be entered into appropriate text fields by selecting checkboxes, selecting one or more items from drop-down menus, or using a combination thereof. The methods of the invention can include using the selected parameters for the methods of the invention.

[0049] FIG. 5 shows an exemplary graphical user interface 500 for receiving input from a user. The graphical user interface 500 includes a menu 502 for selecting common parameters between the reference dataset and the secondary dataset and a menu 503 for selecting parameters to be derived / embedded in the secondary dataset during dimensionality reduction. In the example of FIG. 5, check boxes 503 and 504 are provided to indicate whether to calculate input and output quality scores, respectively. The graphical user interface 500 also includes a text field 505 for indicating the maximum size of the reference dataset. The maximum training size limits the search area for nearest neighbor detection, essentially downsampling to the entered number before finding nearest neighbors. The user may also indicate in text field 506 the number of nearest neighbors (i.e., k) to be used to calculate k nearest neighbors in the reference dataset for each data point collected from the secondary sample. The graphical user interface 500 also includes a drop-down menu 507 for selecting a distance metric.

[0050] As described above, the reference data set and the secondary data set may include flow cytometer data in some embodiments. In particular examples, the flow cytometer data is fluorescence flow cytometer data. "Fluorescence flow cytometer data" refers to information about parameters of a sample (e.g., cells, particles) in a flow cell collected by any number of fluorescence detectors in a particle analyzer. In embodiments, the fluorescence flow cytometer data includes signals from a plurality of different fluorochromes, for example, in the range of 2 to 40 different fluorochromes (including 3 to 30 different fluorochromes, e.g., 3 to 20 different fluorochromes), and in some examples, 3 to 5 different fluorochromes. In some embodiments, the plurality of different fluorochromes includes two or more different fluorochromes, such as 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 11 or more, 12 or more, 13 or more, 14 or more, 15 or more, 20 or more, 25 or more, and 30 or more different fluorochromes. The fluorescence flow cytometer data may be acquired by any convenient protocol, including those described below.

[0051] In some embodiments, the method includes generating one or more population clusters based on determined parameters of analytes (e.g., cells, particles) in the sample. As used herein, a "population" or "subpopulation" of analytes, such as cells or other particles, generally refers to a group of analytes that share characteristics (e.g., optical, impedance, or temporal characteristics) related to one or more measured fluorescence parameters, such that the measured parameter data form a cluster in data space. Thus, populations are recognized as clusters in the data. Conversely, each data cluster is generally interpreted as corresponding to a population of a particular type of cell or analyte, although clusters corresponding to noise or background are also commonly observed. Clusters can be defined in a subset of dimensions, e.g., with respect to a subset of measured fluorescence parameters (i.e., fluorescent dyes), which correspond to populations that differ only in a subset of the measured parameters or features extracted from the sample measurements.

[0052] The method for generating flow cytometer data used in the present invention can vary as needed. For example, a sample containing particles (e.g., in a flow stream of a flow cytometer) can be illuminated with light from a light source. In some embodiments, the light source is a broadband light source, emitting light having a wide range of wavelengths, e.g., spanning 50 nm or more, e.g., 100 nm or more, e.g., 150 nm or more, e.g., 200 nm or more, e.g., 250 nm or more, e.g., 300 nm or more, e.g., 350 nm or more, e.g., 400 nm or more (including 500 nm or more). For example, one suitable broadband light source emits light having a wavelength between 200 nm and 1500 nm. Another example of a suitable broadband light source includes a light source that emits light having a wavelength between 400 nm and 1000 nm. Where the method includes irradiating with a broadband light source, the broadband light source protocol of interest can include, but is not limited to, a halogen lamp, a deuterium arc lamp, a xenon arc lamp, a stabilized fiber-coupled broadband light source, a broadband LED with a continuous spectrum, a superluminescent light emitting diode, a semiconductor light emitting diode, a broadband LED white light source, a multi-LED integrated white light source, or any combination thereof, among other broadband light sources.

[0053] In other embodiments, the method comprises irradiating with a narrowband light source that emits a specific wavelength or narrow range of wavelengths, for example, a light source that emits light in a narrow range, such as a range of 50 nm or less, for example, 40 nm or less, for example, 30 nm or less, for example, 25 nm or less, for example, 20 nm or less, for example, 15 nm or less, for example, 10 nm or less, for example, 5 nm or less, for example, 2 nm or less (including light sources that emit specific wavelengths of light (i.e., monochromatic light)). When the method comprises irradiating with a narrowband light source, the narrowband light source protocol of interest may include, but is not limited to, a narrow wavelength LED, laser diode, or broadband light source coupled to one or more optical bandpass filters, diffraction gratings, monochromators, or any combination thereof.

[0054] Aspects of the invention include collecting fluorescence with a fluorescence detector. The fluorescence detector, in some examples, may be configured to detect fluorescent emission from fluorescent molecules associated with particles in the flow cell, e.g., labeled specific binding members (e.g., labeled antibodies that specifically bind to markers of interest). In certain embodiments, the method includes detecting fluorescence from the sample with one or more fluorescence detectors (e.g., two or more, e.g., three or more, e.g., four or more, e.g., five or more, e.g., six or more, e.g., seven or more, e.g., eight or more, e.g., nine or more, e.g., ten or more, e.g., fifteen or more, including 25 or more fluorescence detectors). In embodiments, each of the fluorescence detectors is configured to generate a fluorescence data signal. Fluorescence from the sample may be detected by each fluorescence detector independently over one or more wavelength ranges from 200 nm to 1200 nm. In some examples, the method includes detecting fluorescence from the sample over a wavelength range, e.g., 200 nm to 1200 nm, e.g., 300 nm to 1100 nm, e.g., 400 nm to 1000 nm, e.g., 500 nm to 900 nm (including 600 nm to 800 nm). In other examples, the method includes detecting fluorescence with each fluorescence detector at one or more specific wavelengths. For example, fluorescence may be detected at one or more of 450 nm, 518 nm, 519 nm, 561 nm, 578 nm, 605 nm, 607 nm, 625 nm, 650 nm, 660 nm, 667 nm, 670 nm, 668 nm, 695 nm, 710 nm, 723 nm, 780 nm, 785 nm, 647 nm, 617 nm, and any combination thereof, depending on the number of different fluorescence detectors in the optical detection system of interest. In certain embodiments, the method includes detecting wavelengths of light corresponding to the fluorescence peak wavelengths of specific fluorescent dyes present in the sample. In an embodiment, the fluorescence flow cytometer data is received from one or more fluorescence detectors (e.g., one or more detection channels), such as two or more, for example three or more, for example four or more, for example five or more, for example six or more, including eight or more fluorescence detectors (e.g., eight or more detection channels).

[0055] In some embodiments, the datasets described herein may be subjected to further analysis and / or processing. Additional processing methods that may be used in conjunction with the methods of the present disclosure are described, for example, in U.S. Patent Nos. 11,506,593 and 11,674,879, and U.S. Patent Application Publication Nos. 2021 / 0358566 and 2022 / 0390349, the disclosures of which are incorporated herein by reference in their entireties.

[0056] The reference sample and the secondary sample may vary. In embodiments, the reference sample and the secondary sample are collected from the same organism or organism type (e.g., species). In some embodiments, the reference sample and the secondary sample are produced or prepared in the same manner or using the same protocol. In some examples, the sample analyzed in the present method is a biological sample. The term "biological sample" is used in its conventional sense, and in particular refers to a whole organism, a whole plant, a whole fungus, or a subset of animal tissues, cells, or component parts, such as may be found in blood, mucus, lymph, synovial fluid, cerebrospinal fluid, saliva, bronchoalveolar lavage fluid, amniotic fluid, amniotic fluid, umbilical cord blood, urine, vaginal fluid, and semen. Thus, "biological sample" refers to both an intact organism or a subset of its tissues, as well as homogenates made from an organism or a subset of its tissues, including, but not limited to, lysates or extracts, such as plasma, serum, cerebrospinal fluid, lymph, skin, respiratory, gastrointestinal, cardiovascular, and urogenital tract sections, tears, saliva, milk, blood cells, tumors, and organs. The biological sample can be any type of biological tissue, including both healthy and diseased tissue (e.g., cancerous, malignant, necrotic, etc.). In certain embodiments, the biological sample is a liquid sample such as blood or a derivative thereof, e.g., plasma, tears, urine, semen, etc., and in some instances, the sample is a blood sample, including whole blood, such as blood obtained from venipuncture or a finger stick (which may or may not be combined with any reagents, such as preservatives, anticoagulants, etc., prior to assay).

[0057] In certain embodiments, the source of the sample is a "mammal," a term used broadly to describe organisms belonging to the class Mammalia, including the orders Carnivora (e.g., dogs and cats), Rodents (e.g., mice, guinea pigs, and rats), and Primates (e.g., humans, chimpanzees, and monkeys). In some examples, the subject is a human. The methods may be applied to samples obtained from human subjects of both genders and at any developmental stage (i.e., newborn, infant, juvenile, adolescent, adult), and in certain embodiments, the human subject is a juvenile, adolescent, or adult. While the present invention may be applied to samples from human subjects, it should be understood that the methods may also be performed on samples from other animal subjects (i.e., "non-human subjects"), such as, but not limited to, birds, mice, rats, dogs, cats, livestock, and horses.

[0058] Cells of interest can be targeted for characterization according to various parameters, such as phenotypic characteristics identified through the attachment of specific fluorescent labels to the cells of interest. In some embodiments, the system is configured to deflect analyzed droplets determined to contain target cells. A variety of cells can be characterized using the subject methods. Target cells of interest include, but are not limited to, stem cells, T cells, dendritic cells, B cells, granulocytes, leukemia cells, lymphoma cells, viral cells (e.g., HIV cells), NK cells, macrophages, monocytes, fibroblasts, epithelial cells, endothelial cells, and erythroid cells. Target cells of interest include cells bearing favorable cell surface markers or antigens that can be captured or labeled by favorable affinity factors or conjugates thereof. For example, target cells may comprise cell surface antigens such as CD11b, CD123, CD14, CD15, CD16, CD19, CD193, CD2, CD25, CD27, CD3, CD335, CD36, CD4, CD43, CD45RO, CD56, CD61, CD7, CD8, CD34, CD1c, CD23, CD304, CD235a, T cell receptor alpha / beta, T cell receptor gamma / delta, CD253, CD95, CD20, CD105, CD117, CD120b, Notch 4, Lgr5 (N-terminus), SSEA-3, TRA-1-60 antigen, disialoganglioside GD2, and CD71. In some embodiments, the target cells are selected from HIV-containing cells, Treg cells, antigen-specific T cell populations, tumor cells or hematopoietic progenitor cells (CD34+) obtained from whole blood, bone marrow, or umbilical cord blood.

[0059] The subject methods may further include using the particles in research, laboratory testing, or therapy. In some embodiments, the subject methods include obtaining individual cells prepared from a biological sample of a target fluid or tissue. For example, the subject methods include obtaining cells from a fluid or tissue sample used as a research or diagnostic specimen for a disease such as cancer. Similarly, the subject methods include obtaining cells from a fluid or tissue sample used for therapy. Cell therapy protocols are protocols in which viable cellular material, including, for example, cells and tissue, can be prepared and introduced into a subject as a therapeutic treatment. Conditions that can be treated by administering flow cytometry-sorted samples include, but are not limited to, blood disorders, immune system disorders, organ damage, and the like.

[0060] A typical cell therapy protocol may include the following steps: sample collection, cell isolation, genetic modification, culture and in vitro expansion, cell harvesting, sample volume reduction and washing, biopreservation, storage, and cell introduction into a subject. The protocol may begin with collecting viable cells and tissues from a subject's source tissue to produce a cell and / or tissue sample. The sample may be collected by any suitable procedure, including, for example, administering a cell mobilizing agent to the subject, withdrawing blood from the subject, removing bone marrow from the subject, etc. After collecting the sample, cell enrichment may be performed by several methods, including, for example, centrifugation-based methods, filter-based methods, elution, magnetic separation, fluorescence-activated cell sorting (FACS), etc. In some cases, the enriched cells may be genetically modified by any convenient method, such as nuclease-mediated gene editing. The genetically modified cells may be cultured, activated, and expanded in vitro. In some cases, the cells may be preserved, e.g., cryopreserved, and stored for future use, whereupon the cells may be thawed and then administered to a patient, e.g., the cells may be infused into a patient.

[0061] system Aspects of the present invention further include systems. The systems of interest include a processor including a memory operably coupled to the processor, the memory storing instructions that, when executed by the processor, cause the processor to perform a method of the present invention (e.g., the method described above). In particular, the processor is configured to receive a secondary dataset including data points collected from a secondary sample and to calculate a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. As described above, the second dimensionality reduction is compatible with the dimensionality-reduced reference dataset.

[0062] In some cases, the subject processor operates with programmable logic, which may be implemented in hardware, software, firmware, or any combination thereof, to calculate a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from the reference sample, for example. For example, if the programmable logic is implemented in software, the method may be realized at least in part by a computer-readable data storage medium including program code including instructions configured, when executed, to calculate a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from the reference sample. For example, the program code may include instructions for determining the data points collected from the reference sample that are closest to each data point collected from the secondary sample with respect to a given data parameter, for example, by calculating k nearest neighbors in the reference dataset for each data point collected from the secondary sample. The processor may be further configured to perform any of the other method steps described above.

[0063] The subject programmable logic may be implemented in any of a variety of devices, such as a specifically programmed event processing computer, a wireless communication device, an integrated circuit device, or the like. In some embodiments, the programmable logic may be executed by a specially programmed processor, which may include one or more processors, such as one or more digital signal processors (DSPs), configurable microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other equivalent integrated or discrete logic circuits. Combinations of computing devices, such as a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration in at least partial data connection, may implement one or more of the features described herein. In certain cases, the system further includes a display configured to output results of the subject methods (e.g., a dimension-reduced plot of the data, etc.). Subject displays may include, but are not limited to, monitors, tablet computers, smartphones, or other electronic devices configured to present graphical interfaces.

[0064] In embodiments, the system is configured to analyze the data within software or analysis tools for analyzing flow cytometer data, such as FlowJo™ (Ashland, OR). FlowJo™ is a software package developed by FlowJo LLC (a subsidiary of Becton Dickinson) for analyzing flow cytometer data. The software is configured to manage flow cytometer data and generate graphical reports thereon (https: / / www(dot)flowjo(dot)com / learn / flowjo-university / flowjo). The initial data can be analyzed within the data analysis software or tool (e.g., FlowJo™) by appropriate means, such as manual gating, cluster analysis, or other computational techniques. The system, or portions thereof, can be implemented as a software component of software for analyzing data, such as FlowJo™. In these embodiments, a computerized system according to the present disclosure can function as a software “plug-in” to an existing software package, such as FlowJo™.

[0065] In some embodiments, the subject processor is used as part of or in combination with a flow cytometer. The subject flow cytometer includes a flow cell. The term "flow cell" is used in its conventional sense to refer to a component, such as a cuvette, containing a flow channel with a liquid flow stream for transporting particles in sheath fluid. The subject cuvette includes a container with a passageway therethrough. The flow stream may contain a liquid sample injected from a sample tube. The subject flow cell includes an optically accessible flow channel. In some examples, the flow cell includes a transparent material (e.g., quartz) that allows light to pass through. In some embodiments, the flow cell is a stream-in-air flow cell in which light interrogation of particles occurs in free space. In some cases, the flow stream is configured to be illuminated with light from one or more light sources at the interrogation point. As discussed herein, "interrogation point" refers to a region in the flow stream where particles are illuminated by light from a light source, e.g., for analysis. The size of the interrogation point can be varied as needed. For example, if 0 μm represents the axis of light emitted by the light source associated with the light, the interrogation zone may range from -100 μm to 100 μm, such as from -50 μm to 50 μm, such as from -25 μm to 40 μm, including from -15 μm to 30 μm. The flow stream comprising the flow channel may contain a liquid sample injected from a sample tube. In certain embodiments, the flow stream may comprise a narrow, rapidly flowing liquid stream arranged such that linearly separated particles transported therein are separated from one another in a single-file fashion. After the particles are illuminated within the flow cell, particle-modulated light may be observed.

[0066] Any convenient flow cell that propagates the fluid sample to the sample interrogation region may be used, and in some embodiments the flow cell includes a cylindrical flow cell, a frusto-conical flow cell, or a flow cell that includes a proximal cylindrical portion defining a longitudinal axis and a distal frusto-conical portion terminating in a flat surface having an orifice transverse to the longitudinal axis.

[0067] Aspects of the subject systems also include a light source for illuminating the flow stream at the interrogation point. Any convenient light source, such as a laser, may be used. In embodiments, the laser may be any convenient laser, such as a continuous wave laser. For example, the laser may be a diode laser, such as an ultraviolet diode laser, a visible diode laser, or a near-infrared diode laser. In other embodiments, the laser may be a helium-neon (HeNe) laser. In some examples, the laser is a gas laser, such as a helium-neon laser, an argon laser, a krypton laser, a xenon laser, a nitrogen laser, a CO2 laser, a CO2 laser, an argon-fluorine (ArF) excimer laser, a krypton-fluorine (KrF) excimer laser, a xenon-chlorine (XeCL) excimer laser, or a xenon-fluorine (XeF) excimer laser, or a combination thereof. In other examples, the subject flow cytometer includes a dye laser, such as a stilbene, coumarin, or rhodamine laser. In yet another example, lasers of interest include metal vapor lasers such as helium-cadmium (HeCd) lasers, helium-mercury (HeHg) lasers, helium-selenium (HeSe) lasers, helium-silver (HeAg) lasers, strontium lasers, neon-copper (NeCu) lasers, copper lasers, or gold lasers, and combinations thereof. In yet another example, flow cytometers of interest include solid-state lasers such as ruby ​​lasers, Nd:YAG lasers, NdCrYAG lasers, Er:YAG lasers, Nd:YLF lasers, Nd:YVO4 lasers, Nd:YCa4O(BO3)3 lasers, Nd:YCOB lasers, titanium sapphire lasers, thulium YAG lasers, ytterbium YAG lasers, ytterbium 2O3 lasers, or cerium-doped lasers, and combinations thereof.

[0068] The laser light source according to certain embodiments can also include one or more optical adjustment components. In certain embodiments, the optical adjustment component may include any device disposed between the light source and the flow cell that can change the spatial width of the illumination or some other characteristic of the illumination from the light source, such as the illumination direction, wavelength, beam width, beam intensity, and focus. The optical adjustment protocol can include any convenient device that adjusts one or more characteristics of the light source, including, but not limited to, lenses, mirrors, filters, optical fibers, wavelength separators, pinholes, slits, collimation protocols, and combinations thereof. In certain embodiments, the target flow cytometer includes one or more focusing lenses. In one example, the focusing lens may be a reduction lens. In yet other embodiments, the target flow cytometer includes optical fibers.

[0069] Where the optical adjustment component is configured to move, it may be configured to move continuously or at discrete intervals, for example in increments of 0.01 μm or more, for example 0.05 μm or more, for example 0.1 μm or more, for example 0.5 μm or more, such as 1 μm or more, for example 10 μm or more, for example 100 μm or more, such as 500 μm or more, for example 1 mm or more, for example 5 mm or more, for example 10 mm or more (including increments of 25 mm or more).

[0070] Any displacement protocol can be used to move the optical adjustment component structure, such as those coupled to a movable support stage or using a motorized translation stage, a lead screw translation assembly, a geared translation device, e.g., a stepper motor, a servo motor, a brushless electric motor, a brushed DC motor, a microstep drive motor, a high resolution stepper motor, among other types of motors.

[0071] The light source may be positioned at any suitable distance from the flow cell, for example, the light source and the flow cell are separated by 0.005 mm or more, for example, 0.01 mm or more, for example, 0.05 mm or more, for example, 0.1 mm or more, for example, 0.5 mm or more, for example, 1 mm or more, for example, 5 mm or more, for example, 10 mm or more, for example, 25 mm or more (including a distance of 100 mm or more). Furthermore, the light source may be positioned at any suitable angle relative to the flow cell, for example, an angle ranging from 10 degrees to 90 degrees, for example, from 15 degrees to 85 degrees, for example, from 20 degrees to 80 degrees, for example, from 25 degrees to 75 degrees (including an angle of 30 degrees to 60 degrees, for example, 90 degrees).

[0072] Systems according to certain embodiments include multiple light sources. In some cases, the multiple light sources include multiple lasers. For example, two or more lasers, e.g., three or more lasers, e.g., four or more lasers, e.g., five or more lasers, e.g., ten or more lasers, etc., including fifteen or more lasers configured to provide laser light for discrete illumination of the flowstream. Depending on the desired wavelength of light for illuminating the flowstream, each laser may have a specific wavelength that varies from 200 nm to 1500 nm, e.g., 250 nm to 1250 nm, e.g., 300 nm to 1000 nm, e.g., 350 nm to 900 nm, including 400 nm to 800 nm. In certain embodiments, the lasers of interest may include one or more of a 405 nm laser, a 488 nm laser, a 561 nm laser, and a 635 nm laser.

[0073] The subject systems may include one or more detectors for detecting particle-modulated light intensity data. In some embodiments, the particle-modulated light detector includes one or more forward-scattered light detectors configured to detect forward-scattered light. For example, the subject particle analyzer may include one forward-scattered light detector or multiple forward-scattered light detectors, e.g., two or more, e.g., three or more, e.g., four or more (including five or more). In certain embodiments, the particle analyzer includes one forward-scattered light detector. In other embodiments, the particle analyzer includes two forward-scattered light detectors.

[0074] Any convenient detector for detecting collected light can be used in the forward scattered light detector described herein. Detectors of interest can include, but are not limited to, optical sensors or detectors such as active pixel sensors (APS), avalanche photodiodes, image sensors, charge-coupled devices (CCDs), intensified charge-coupled devices (ICCDs), light-emitting diodes, photon counters, bolometers, pyroelectric detectors, photoresistors, photocells, photodiodes, photomultiplier tubes (PMTs), phototransistors, quantum dot photoconductors, or photodiodes, and combinations thereof, among other detectors. In certain embodiments, the collected light is measured with a charge-coupled device (CCD), a semiconductor charge-coupled device (CCD), an active pixel sensor (APS), a complementary metal-oxide semiconductor (CMOS) image sensor, or an N-type metal-oxide semiconductor (NMOS) image sensor. In certain embodiments, the detector has a resolution of 0.01 cm. 2 ~10cm 2 , e.g. 0.05cm 2 ~9cm 2 , e.g. 0.1cm 2 ~8cm 2 , e.g. 0.5cm 2 ~7cm 2 Range (1cm 2 ~5cm 2 and a photomultiplier tube such as a photomultiplier tube having an active detection surface area of ​​each region that is

[0075] In embodiments, the forward scattered light detector is configured to measure light continuously or at discrete intervals. In some examples, the detector is configured to continuously obtain measurements of the collected light. In other examples, the detector is configured to make measurements at discrete intervals, such as measuring light every 0.001 milliseconds, 0.01 milliseconds, 0.1 milliseconds, 1 millisecond, 10 milliseconds, 100 milliseconds (including every 1000 milliseconds), or some other interval.

[0076] In some examples, the system includes one or more side scattered light detectors for detecting side scattered wavelengths of light (i.e., light refracted and reflected from the surface and internal structures of the particle). In some embodiments, the particle analyzer includes a single side scattered light detector. In other embodiments, the particle analyzer includes a plurality of side scattered light detectors, e.g., two or more, e.g., three or more, e.g., four or more (including five or more).

[0077] Any convenient detector for detecting collected light may be used in the side-scattered light detector described herein. Detectors of interest may include, but are not limited to, optical sensors or detectors such as active pixel sensors (APS), avalanche photodiodes, image sensors, charge-coupled devices (CCDs), intensified charge-coupled devices (ICCDs), light-emitting diodes, photon counters, bolometers, pyroelectric detectors, photoresistors, photocells, photodiodes, photomultiplier tubes (PMTs), phototransistors, quantum dot photoconductors, or photodiodes, and combinations thereof, among other detectors. In certain embodiments, the collected light is measured with a charge-coupled device (CCD), a semiconductor charge-coupled device (CCD), an active pixel sensor (APS), a complementary metal-oxide semiconductor (CMOS) image sensor, or an N-type metal-oxide semiconductor (NMOS) image sensor. In certain embodiments, the detector has a resolution of 0.01 cm. 2 ~10cm 2 , e.g. 0.05cm 2 ~9cm 2 , e.g. 0.1cm 2 ~8cm 2 , e.g. 0.5cm 2 ~7cm 2 Range (1cm 2 ~5cm 2 and a photomultiplier tube such as a photomultiplier tube having an active detection surface area of ​​each region that is

[0078] In embodiments, the subject systems also include a fluorescence detector configured to detect one or more fluorescent wavelengths of light, hi other embodiments, the particle analyzer includes multiple fluorescence detectors, e.g., 2 or more, e.g., 3 or more, e.g., 4 or more, 5 or more, 10 or more, 15 or more, including 20 or more.

[0079] Any convenient detector for detecting collected light may be used in the fluorescence detectors described herein. Detectors of interest may include, but are not limited to, optical sensors or detectors such as active pixel sensors (APS), avalanche photodiodes, image sensors, charge-coupled devices (CCDs), intensified charge-coupled devices (ICCDs), light-emitting diodes, photon counters, bolometers, pyroelectric detectors, photoresistors, photocells, photodiodes, photomultiplier tubes (PMTs), phototransistors, quantum dot photoconductors, or photodiodes, and combinations thereof, among other detectors. In certain embodiments, the collected light is measured with a charge-coupled device (CCD), a semiconductor charge-coupled device (CCD), an active pixel sensor (APS), a complementary metal-oxide semiconductor (CMOS) image sensor, or an N-type metal-oxide semiconductor (NMOS) image sensor. In certain embodiments, the detector has a resolution of 0.01 cm. 2 ~10cm 2 , e.g. 0.05cm 2 ~9cm 2 , e.g. 0.1cm 2 ~8cm 2 , e.g. 0.5cm 2 ~7cm 2 Range (1cm 2 ~5cm 2 and a photomultiplier tube such as a photomultiplier tube having an active detection surface area of ​​each region that is

[0080] When a subject particle analyzer includes multiple fluorescence detectors, each fluorescence detector may be the same, or the collection of fluorescence detectors may be a combination of different types of detectors. For example, when a subject particle analyzer includes two fluorescence detectors, in some embodiments, the first fluorescence detector is a CCD-type device and the second fluorescence detector (or imaging sensor) is a CMOS-type device. In other embodiments, both the first and second fluorescence detectors are CCD-type devices. In still other embodiments, both the first and second fluorescence detectors are CMOS-type devices. In still other embodiments, the first fluorescence detector is a CCD-type device and the second fluorescence detector is a photomultiplier tube (PMT). In still other embodiments, the first fluorescence detector is a CMOS-type device and the second fluorescence detector is a photomultiplier tube. In still other embodiments, both the first and second fluorescence detectors are photomultiplier tubes.

[0081] In embodiments of the present disclosure, the subject fluorescence detector is configured to measure the collected light at one or more wavelengths, e.g., two or more wavelengths, e.g., five or more different wavelengths, e.g., ten or more different wavelengths, e.g., twenty-five or more different wavelengths, e.g., fifty or more different wavelengths, e.g., one hundred or more different wavelengths, e.g., two ... three hundred or more different wavelengths, including measuring light emitted by the sample in the flow stream at four hundred or more different wavelengths. In some embodiments, two or more detectors of the particle analyzers described herein are configured to measure the same or overlapping wavelengths of the collected light.

[0082] In some embodiments, the target fluorescence detector is configured to measure light collected over a wavelength range (e.g., 200 nm to 1000 nm). In certain embodiments, the target detector is configured to collect a spectrum of light over a wavelength range. For example, a particle analyzer may include one or more detectors configured to collect a spectrum of light over one or more wavelength ranges from 200 nm to 1000 nm. In still other embodiments, the target detector is configured to measure light emitted by a sample in the flow stream at one or more specific wavelengths. For example, a particle analyzer may include one or more detectors configured to measure light at one or more of the following wavelengths: 450 nm, 518 nm, 519 nm, 561 nm, 578 nm, 605 nm, 607 nm, 625 nm, 650 nm, 660 nm, 667 nm, 670 nm, 668 nm, 695 nm, 710 nm, 723 nm, 780 nm, 785 nm, 647 nm, 617 nm, and any combination thereof. In certain embodiments, one or more detectors may be configured to pair with a particular fluorophore, such as one used with a sample in a fluorescence assay.

[0083] In some embodiments, one or more of the particle-modulation light detectors include one or more detector arrays, such as an array of photodiodes. In these embodiments, each detector array may include four or more detectors, e.g., 10 or more detectors, e.g., 25 or more detectors, e.g., 50 or more detectors, e.g., 100 or more detectors, e.g., 250 or more detectors, e.g., 500 or more detectors, e.g., 750 or more detectors (including 1000 or more detectors). For example, a detector may be a photodiode array having four or more photodiodes, e.g., 10 or more photodiodes, e.g., 25 or more photodiodes, e.g., 50 or more photodiodes, e.g., 100 or more photodiodes, e.g., 250 or more photodiodes, e.g., 500 or more photodiodes, e.g., 750 or more photodiodes, including 1000 or more photodiodes.

[0084] The detectors can be arranged in any geometric configuration as desired, including, but not limited to, square, rectangular, trapezoidal, triangular, hexagonal, heptagonal, octagonal, non-angular, decagonal, dodecagonal, circular, elliptical, and irregularly patterned configurations. The detectors within a detector array may be oriented relative to one another (referenced to the XZ plane) at angles ranging from 10° to 180°, e.g., 15° to 170°, e.g., 20° to 160°, e.g., 25° to 150°, e.g., 30° to 120° (including 45° to 90°). The detector array may be of any suitable shape, including rectilinear shapes such as square, rectangular, trapezoidal, triangular, and hexagonal; curvilinear shapes such as circular and elliptical; and irregular shapes such as a parabolic base coupled to a flat top. In certain embodiments, the detector array has an active surface that is rectangular in shape.

[0085] In some embodiments, the particle analyzer includes one or more wavelength separators disposed between the flow cell and the particle-modulated light detector. The term "wavelength separator" is used herein in its conventional sense to refer to an optical component configured to separate light collected from a sample into predetermined spectral ranges. In some embodiments, the particle analyzer includes a single wavelength separator. In other embodiments, the particle analyzer includes multiple wavelength separators, e.g., two or more wavelength separators, e.g., three or more, e.g., four or more, e.g., five or more, e.g., six or more, e.g., seven or more, e.g., eight or more, e.g., nine or more, e.g., ten or more, e.g., fifteen or more, e.g., twenty-five or more, e.g., fifty or more, e.g., seventy-five or more, e.g., seventy-five or more, including one hundred or more wavelength separators. In some embodiments, the wavelength separator is configured to separate light collected from a sample into predetermined spectral ranges by passing light having a predetermined spectral range and reflecting one or more remaining spectral ranges of light. In other embodiments, the wavelength separator is configured to separate light collected from a sample into predetermined spectral ranges by passing light having a predetermined spectral range and absorbing one or more remaining spectral ranges of light. In yet other embodiments, the wavelength separators are configured to spatially diffract the light collected from the sample into predetermined spectral ranges. Each wavelength separator may be any convenient light separation protocol, such as one or more dichroic mirrors, bandpass filters, diffraction gratings, beam splitters, or prisms. In some embodiments, the wavelength separator is a prism. In other embodiments, the wavelength separator is a diffraction grating. In certain embodiments, the wavelength separator of the subject optical detection system is a dichroic mirror.

[0086] In certain cases, one or more detectors in the system may be considered a trigger sensor (i.e., a sensor that observes the presence of a particle and generates a trigger signal). In some embodiments, the trigger sensor is a forward-scattered light detector (e.g., such as those described above). In other cases, the trigger sensor is an axial light loss (ALL) channel sensor. In such cases, the processor may be configured to calculate a trigger window based on the trigger signal, the trigger window providing a time period during which a particle is expected to pass through the detection zone of the detector, and to obtain a baseline noise level for time periods outside the trigger window.

[0087] Suitable flow cytometry systems include those described in Ormerod (ed.), Flow Cytometry: A Practical Approach, Oxford University Press (1997), Jaroszeski et al. (eds.), Flow Cytometry Protocols, Methods in Molecular Biology No. 91, Humana Press (1997), Practical Flow Cytometry, 3rd ed., Wiley-Liss (1995), Virgo, et al. (2012) Ann Clin Biochem. Jan;49(pt1):17-28, Linden, et al., Semin Thromb Hemost. 2004 Oct;30(5):502-11, Alison, et al. J Pathol, 2010 Dec;222(4):335-344, and Herbig, et al. (2007) Crit Rev Ther Drug Carrier Syst. 24(3):203-255, the disclosures of which are incorporated herein by reference.In particular examples, flow cytometry systems of interest include a BD Biosciences FACSCanto™ flow cytometer, a BD Biosciences FACSCanto™ II flow cytometer, a BD Accuri™ flow cytometer, a BD Accuri™ C6 Plus flow cytometer, a BD Biosciences FACSCelesta™ flow cytometer, a BD Biosciences FACSLyric™ flow cytometer, a BD Biosciences FACSVerse™ flow cytometer, a BD Biosciences FACSymphony™ flow cytometer, a BD Biosciences LSRFortessa™ flow cytometer, a BD Biosciences LSRFortessa™ X-20 flow cytometer, a BD Biosciences FACSPresto™ flow cytometer, a BD Biosciences FACSVia™ flow cytometer, and a BD Biosciences FACSCalibur™ cell sorter, a BD Biosciences FACSCount™ cell sorter, a BD Biosciences These include the FACSLyric™ cell sorter, BD Biosciences Via™ cell sorter, BD Biosciences Influx™ cell sorter, BD Biosciences Jazz™ cell sorter, BD Biosciences Aria™ cell sorter, BD Biosciences FACSAria™ II cell sorter, BD Biosciences FACSAria™ III cell sorter, BD Biosciences FACSAria™ Fusion cell sorter, and BD Biosciences FACSMelody™ cell sorter, BD Biosciences FACSymphony™ S6 cell sorter, etc.

[0088] In some embodiments, the subject system is a hybrid system as described in U.S. Pat. Nos. 10,663,476, 10,620,111, 10,613,017, 10,605,713, 10,585,031, 10,578,542, 10,578,469, 10,481,074, 10,302 ,545 specification, 10,145,793 specification, 10,113,967 specification, 10,006,852 specification, 9,952,076 specification, 9,933, Specification No. 341, Specification No. 9,726,527, Specification No. 9,453,789, Specification No. 9,200,334, Specification No. 9,097,640, Specification No. 9,095,494 Specification, Specification No. 9,092,034, Specification No. 8,975,595, Specification No. 8,753,573, Specification No. 8,233,146, Specification No. 8,140,300, Specification No. 7,544,326, Specification No. 7,201,875, Specification No. 7,129,505, Specification No. 6,821,740, Specification No. 6,813,017, Specification No. 6, The flow cytometry system may be a flow cytometry system such as those described in U.S. Patent Nos. 809,804, 6,372,506, 5,700,692, 5,643,796, 5,627,040, 5,620,842, 5,602,039, 4,987,086, and 4,498,766, the disclosures of which are incorporated herein by reference in their entirety. In some cases, the flow cytometry system of the present invention is a clustered wavelength division (CWD) system. CWD systems are described, for example, in U.S. Patent Application Publication No. 2021 / 0247293, the disclosure of which is incorporated herein by reference in its entirety.

[0089] In certain examples, the flow cytometry system of the present invention may be implemented in accordance with the principles of Diebold, et al. Nature Photonics Vol. 7(10); 806-810 (2013) and U.S. Patent Nos. 9,423,353, 9,784,661, 9,983,132, 10,006,852, 10,078,045, 10,036,699, 10,222,316, 10,288,546, 10,324,019, 10,408,758, 10,451,538, 10,620,111, and the like. and configured for imaging particles in a flowstream by fluorescence imaging using radio frequency tagged emission (FIRE), as described in U.S. Patent Application Publication Nos. 2017 / 0133857, 2017 / 0328826, 2017 / 0350803, 2018 / 0275042, 2019 / 0376895, and 2019 / 0376894, the disclosures of which are incorporated herein by reference.

[0090] 6 shows a system 600 for flow cytometry according to an exemplary embodiment of the invention. System 600 includes a flow cytometer 610, a controller / processor 690, and a memory 695. Flow cytometer 610 includes one or more excitation lasers 615a-615c, a focusing lens 620, a flow chamber 625, a forward scatter detector 630, a side scatter detector 635, a fluorescence collection lens 640, one or more beam splitters 645a-645g, one or more bandpass filters 650a-650e, one or more longpass ("LP") filters 655a-655b, and one or more fluorescence detectors 660a-660f.

[0091] Excitation lasers 615a-615c emit light in the form of laser beams. In the exemplary system of FIG. 6, the wavelengths of the laser beams emitted from excitation lasers 615a-615c are 488 nm, 633 nm, and 325 nm, respectively. The laser beams are first directed through one or more of beam splitters 645a and 645b. Beam splitter 645a transmits 488 nm light and reflects 633 nm light. Beam splitter 645b transmits ultraviolet light (light with wavelengths ranging from 10 nm to 400 nm) and reflects 488 nm and 633 nm light.

[0092] The laser beam is then directed to a focusing lens 620, which focuses the beam onto a portion of the flow stream where the sample particles are located, within a flow chamber 625. The flow chamber is the part of a fluidic system that directs particles, typically one at a time, in a stream toward the focused laser beam for interrogation. A flow chamber can include a flow cell in a benchtop cytometer or a nozzle tip in a stream-in air cytometer.

[0093] Light from the laser beam interacts with the sample particles by diffraction, refraction, reflection, scattering, and absorption by re-emission at a variety of different wavelengths, depending on particle characteristics such as particle size, internal structure, and the presence of one or more fluorescent molecules attached to or naturally present on or within the particle. The fluorescent emission and diffracted, refracted, reflected, and scattered light can be sent via one or more of beam splitters 645c-645g, bandpass filters 650a-650e, longpass filters 655a-655b, and fluorescence collection lens 640 to one or more of forward scatter detector 630, side scatter detector 635, and one or more fluorescence detectors 660a-660f.

[0094] The fluorescence collection lens 640 collects light emitted from particle-laser beam interactions and routes it toward one or more beam splitters and filters. Bandpass filters, such as bandpass filters 650a-650e, allow a narrow range of wavelengths to pass through the filter. For example, bandpass filter 650a is a 510 / 20 filter. The first number represents the center of the spectral band. The second number provides the range of the spectral band. Thus, a 510 / 20 filter extends 10 nm on either side of the center of the spectral band, from 500 nm to 520 nm. Shortpass filters transmit light equal to or shorter than a specific wavelength. Longpass filters, such as longpass filters 655a-655b, transmit wavelengths equal to or longer than a specific wavelength of light. For example, longpass filter 655b, a 670 nm longpass filter, transmits light above 670 nm. Filters are often selected to optimize the detector's specificity for a particular fluorescent dye. The filter can be configured so that the spectral band of light transmitted to the detector is close to the emission peak of the fluorescent dye.

[0095] The forward scatter detector 630 is positioned slightly off-axis from the direct beam through the flow cell and is configured to detect diffracted light, or excitation light traveling primarily forward through or around the particle. The intensity of light detected by the forward scatter detector depends on the overall size of the particle. The forward scatter detector may include a photodiode. The side scatter detector 635 is configured to detect refracted and reflected light from the particle's surface and internal structure, which tends to increase as the particle's structure becomes more complex. Fluorescence emission from fluorescent molecules associated with the particle can be detected by one or more fluorescence detectors 660a-660f. The side scatter detector 635 and the fluorescence detector may include photomultiplier tubes. The signals detected by the forward scatter detector 630, side scatter detector 635, and fluorescence detector can be converted to electronic signals (voltage) by the detectors. This data can provide information about the sample.

[0096] Those skilled in the art will recognize that flow cytometers according to embodiments of the present invention are not limited to the flow cytometer shown in Figure 6, but may include any flow cytometer known in the art. For example, a flow cytometer may have any number of lasers, beam splitters, filters, and detectors at various wavelengths and in a variety of different configurations.

[0097] During operation, the operation of the cytometer is controlled by controller / processor 690, and measurement data from the detectors may be stored in memory 695 and processed by controller / processor 690. Although not explicitly shown, controller / processor 690 may be coupled to detectors to receive output signals from the detectors and may also be coupled to electrical and electromechanical components of flow cytometer 610 to control lasers, fluid flow parameters, etc. Input / output (I / O) functionality 697 may also be provided in the system. Memory 695, controller / processor 690, and I / O 697 may be provided entirely as an integral part of flow cytometer 610. In such embodiments, a display may also form part of I / O functionality 697 for presenting experimental data to a user of cytometer 610. Alternatively, some or all of memory 695 and controller / processor 690 and I / O functionality may be part of one or more external devices, such as a general-purpose computer. In some embodiments, some or all of memory 695 and controller / processor 690 may be in wireless or wired communication with cytometer 610. Controller / processor 690, together with memory 695 and I / O 697, can be configured to perform a variety of functions associated with the preparation and analysis of flow cytometer experiments.

[0098] The system shown in FIG. 6 includes six different detectors that detect fluorescence in six different wavelength bands (sometimes referred to herein as the "filter windows" of a given detector) as defined by the configuration of filters and / or splitters in the beam paths from the flow cell 625 to each detector. Different fluorescent molecules used in a flow cytometer experiment emit light in their own characteristic wavelength bands. The particular fluorescent labels used in the experiment and their associated fluorescence emission bands can be selected to closely match the filter windows of the detectors. The I / O 697 can be configured to receive data regarding a flow cytometer experiment involving a panel of fluorescent labels and multiple cell populations with multiple markers, each cell population having a subset of multiple markers. The I / O 697 can also be configured to receive biological data assigning one or more markers to one or more cell populations, marker density data, emission spectrum data, data assigning labels to one or more markers, and cytometer configuration data. Flow cytometer experiment data, such as label spectral characteristics and flow cytometer configuration data, can also be stored in memory 695. The controller / processor 690 can be configured to evaluate one or more assignments of labels to markers.

[0099] In some embodiments, the subject system is a particle sorting system configured to sort particles using an enclosed particle sorting module, such as that described in U.S. Patent Application Publication No. 2017 / 0299493, filed March 28, 2017, the disclosure of which is incorporated herein by reference. In certain embodiments, particles (e.g., cells) of a sample are sorted using a sorting determination module having multiple sorting determination units, such as that described in U.S. Patent Application Publication No. 2020 / 0256781, filed December 23, 2019, the disclosure of which is incorporated herein by reference. In some embodiments, a system for sorting components of a sample includes a particle sorting module with deflection plates, such as that described in U.S. Patent Application Publication No. 2017 / 0299493, filed March 28, 2017, the disclosure of which is incorporated herein by reference.

[0100] FIG. 7 shows a functional block diagram of an example system having a processor 700 for analyzing and displaying biological events. The processor 700 can be configured to implement various processes for controlling the graphical display of the biological events. A flow cytometer or sorting system 702 can be configured to acquire biological event data. For example, a flow cytometer can generate flow cytometry event data (e.g., particle-modulated light data). The flow cytometer 702 can be configured to provide the biological event data to the processor 700. A data communication channel can be included between the flow cytometer 702 and the processor 700. The biological event data can be provided to the processor 700 via the data communication channel.

[0101] Processor 700 can be configured to receive biological event data from flow cytometer 702. The biological event data received from flow cytometer 702 can include flow cytometry event data. Processor 700 is configured to evaluate the data received from flow cytometer 702, e.g., as described above. Processor 700 can be configured to perform the subject methods, e.g., to receive a secondary dataset including data points collected from a secondary sample (optionally by the flow cytometer or sorting system 702) and calculate a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample.

[0102] The processor 700 can be configured to provide a graphical display including the biological event data to the display device 706. For example, the processor 700 may provide the calculated ideal gains for each of the detectors in the system (e.g., fluorescence detectors 360a-360e) to the graphical display device 706. These ideal gains may be output, for example, in the form of tube target values ​​(TTVs). The processor 700 may further be configured to render a region of interest, for example, as a gate around a population of biological event data shown by the display device 706, overlaid on a first plot. In some embodiments, the gate can be a logical combination of one or more graphical regions of interest depicted on a histogram or bivariate plot of a single parameter. In some embodiments, the display can be used to display particle parameters or saturation detector data.

[0103] The processor 700 may be further configured to display the biological event data within the gate on the display device 706 differently from other events within the biological event data outside the gate. For example, the processor 700 may be configured to render the color of the biological event data contained within the gate differently from the color of the biological event data outside the gate. The display device 706 may be implemented as a monitor, tablet computer, smartphone, or other electronic device configured to present a graphical interface.

[0104] The processor 700 may be configured to receive a gate selection signal identifying a gate from a first input device. For example, the first input device may be implemented as a mouse 710. The mouse 710 may initiate a gate selection signal to the processor 700 identifying a gate to be displayed on or operated via the display device 706 (e.g., by clicking the desired gate when a cursor is positioned there). In some implementations, the first device may be implemented as a keyboard 708 or other means for conveying input signals to the processor 700, such as a touchscreen, a stylus, a photodetector, or a voice recognition system. Some input devices may include multiple input functions. In such implementations, each input function can be considered an input device. For example, as shown in FIG. 7, the mouse 710 may include a right mouse button and a left mouse button, each capable of generating a trigger event.

[0105] The trigger event can cause the processor 700 to change how the data is displayed, what portions of the data are actually displayed on the display device 706, and / or provide input for further processing, such as selecting a population for particle sorting purposes.

[0106] In some embodiments, the processor 700 can be configured to detect when a gate selection is initiated by the mouse 710. The processor 700 can be further configured to automatically modify the visualization of the plot to facilitate the gating process. The modification can be based on a particular distribution of the biological event data received by the processor 700. In some embodiments, the processor 700 extends the first gate such that a second gate is generated (e.g., as described above).

[0107] The processor 700 may be connected to a storage device 704. The storage device 704 may be configured to receive and store biological event data from the processor 700. The storage device 704 may also be configured to receive and store flow cytometry event data from the processor 700. The storage device 704 may be further configured to enable retrieval of biological event data, such as flow cytometry event data, by the processor 700.

[0108] The display device 706 can be configured to receive display data from the processor 700. The display data can include plots of biological event data and gates that delineate sections of the plot. The display device 706 can be further configured to modify the information presented according to input received from the processor 700, along with input from the flow cytometer 702, the storage device 704, the keyboard 708, and / or the mouse 710.

[0109] In some implementations, the processor 700 can generate a user interface for receiving example events for sorting. For example, the user interface can include a mechanism for receiving example events or example images. The example events or images or example gates can be provided prior to collection of event data for the sample or based on an initial set of events for a portion of the sample.

[0110] FIG. 8A is a schematic diagram of a particle sorter system 800 (e.g., flow cytometer 702) according to one embodiment presented herein. In some embodiments, particle sorter system 800 is a cell sorter system. As shown in FIG. 8A, a droplet-forming transducer 802 (e.g., a piezoelectric oscillator) is coupled to a fluid conduit 801, which may be coupled to, include, or be a nozzle 803. Within fluid conduit 801, sheath fluid 804 hydrodynamically focuses sample fluid 806 containing particles 809 into a moving fluid column 808 (e.g., a stream). Within moving fluid column 808, particles 809 (e.g., cells) are aligned in single file across a monitoring area 811 (e.g., where laser streams intersect) illuminated by an illumination source 812 (e.g., a laser). Vibration of droplet-forming transducer 802 causes moving fluid column 808 to break up into multiple droplets 810 , some of which contain particles 809 .

[0111] During operation, the detection station 814 (e.g., an event detector) identifies when a particle (or cell) of interest crosses the monitoring area 811. The detection station 814 is fed to a timing circuit 828, which in turn feeds a flash charge circuit 830. At a drop breakoff point, signaled by a timed drop delay (Δt), a flash charge can be applied to the moving fluid column 808 so that the droplets of interest carry a charge. The droplets of interest may contain one or more particles or cells to be sorted. The injected droplets can then be sorted by activating a deflection plate (not shown) to deflect the droplets into a collection tube or a container, such as a multi-well or microwell sample plate, and a well or microwell can be associated with the particular droplet of interest. As shown in FIG. 8A, the droplets can be collected in a waste receptacle 838.

[0112] Detection system 816 (e.g., a droplet boundary detector) helps automatically determine the phase of the droplet drive signal when a particle of interest passes through monitoring area 811. An exemplary droplet boundary detector is described in U.S. Patent No. 7,679,039, which is incorporated herein by reference in its entirety. Detection system 816 allows the instrument to accurately calculate the location of each detected particle in the droplet. Detection system 816 can provide amplitude signal 820 and / or phase 818 signals, which then (via amplifier 822) provide to amplitude control circuit 826 and / or frequency control circuit 824. Amplitude control circuit 826 and / or frequency control circuit 824 then control droplet forming transducer 802. Amplitude control circuit 826 and / or frequency control circuit 824 can be included in a control system.

[0113] In some implementations, the sorting electronics (e.g., detection system 816, detection station 814, and processor 840) can be coupled with a memory configured to store the detected events and sorting decisions based thereon. The sorting decisions can be included in the particle's event data. In some implementations, the detection system 816 and detection station 814 can be implemented as a single detection unit or can be communicatively coupled such that event measurements can be collected by either the detection system 816 or the detection station 814 and provided to a non-collecting element.

[0114] FIG. 8B is a schematic diagram of a particle sorter system according to one embodiment presented herein. The particle sorter system 800 shown in FIG. 8B includes deflection plates 852 and 854. An electric charge can be applied via stream charging wires within the barbs. This creates a stream of droplets 810 containing particles 809 for analysis. The particles can be illuminated with one or more light sources (e.g., lasers) to generate light scattering and fluorescence information. The information about the particles is analyzed, such as by sorting electronics or other detection systems (not shown in FIG. 8B). Deflection plates 852 and 854 can be independently controlled to attract or repel charged droplets, directing them toward a destination collection vessel (e.g., any of 872, 874, 876, or 878). As shown in FIG. 8B, deflection plates 852 and 854 can be controlled to direct particles along a first path 862 toward vessel 874 or along a second path 868 toward vessel 878. If the particle is not of interest (e.g., does not exhibit scattering or illumination information within the specified sorting range), the deflector may allow the particle to continue along flow path 864. Such uncharged droplets may enter a waste receptacle, such as via aspirator 870.

[0115] Sorting electronics can be included to initiate measurement collection, receive particle fluorescent signals, and determine how to adjust the deflection plates to cause particle sorting. An exemplary implementation of the embodiment shown in Figure 8B includes the BD FACSAria™ line of flow cytometers commercially available from Becton, Dickinson and Company (Franklin Lakes, NJ).

[0116] In certain embodiments, the system is a fluorescence imaging using radio frequency tagged luminescence imaging enabled particle sorter as shown in FIG. 9 . Particle sorter 900 includes a light illumination component 900a including a light source 901 (e.g., a 488 nm laser) generating an output beam of light 901a that is split into beams 902a and 902b by a beam splitter 902. Light beam 902a propagates through an acousto-optic device (e.g., an acousto-optic deflector, AOD) 903 to generate an output beam 903a having one or more angularly deflected light beams. In some examples, output beam 903a generated from acousto-optic device 903 includes a local oscillator beam and multiple radio frequency comb beams. Light beam 902b propagates through an acousto-optic device (e.g., an acousto-optic deflector, AOD) 904 to generate an output beam 904a having one or more angularly deflected light beams. In some examples, output beam 904a generated from acousto-optic device 904 includes a local oscillator beam and multiple radio frequency comb beams. Output beams 903a and 904a generated from acousto-optical devices 903 and 904, respectively, are combined with beam splitter 905 to generate output beam 905a, which is conveyed through optical components 906 (e.g., an objective lens) to illuminate particles in flow cell 907. In certain embodiments, acousto-optical device 903 (AOD) splits a single laser beam into an array of beamlets, each having a different optical frequency and angle. A second AOD 904 adjusts the optical frequency of a reference beam, which is then overlapped with the array of beamlets at beam combiner 905. In certain embodiments, the light illumination system having a light source and acousto-optical device can also include those described in Schraivogel, et al. (“High-speed fluorescence image-enabled cell sorting,” Science (2022), 375(6578):315-320) and U.S. Patent Application Publication No. 2021 / 0404943, the disclosures of which are incorporated herein by reference.

[0117] Output beam 905a illuminates sample particles 908 propagating through flow cell 907 (e.g., with sheath fluid 909) in illumination region 910. As shown in illumination region 910, multiple beams (e.g., angularly deflected, high-frequency shifted optical beams shown as dots across illumination region 910) overlap with a reference local oscillator beam (shown as hatched across illumination region 910). Due to their different optical frequencies, the overlapping beams exhibit beat behavior, whereby each beamlet emits at a distinct frequency f 1-n carries a sinusoidal modulation.

[0118] Light from the illuminated sample is conveyed to a light detection system 900b, which includes multiple photodetectors. The light detection system 900b includes a forward-scattered light photodetector 911 for generating a forward-scattered image 911a and a side-scattered light photodetector 912 for generating a side-scattered image 912a. The light detection system 900b also includes a bright-field photodetector 913 for generating a light loss image 913a. In some embodiments, the forward-scattered detector 911 and the side-scattered detector 912 are photodiodes (e.g., avalanche photodiodes, APDs). In some examples, the bright-field photodetector 913 is a photomultiplier tube (PMT). Fluorescence from the illuminated sample is also detected by fluorescence photodetectors 914-917. In some examples, the photodetectors 914-917 are photomultiplier tubes. The light from the illuminated sample is directed to the side-scattered detection channel 912 and the fluorescence detection channels 914-917 via a beam splitter 920. The optical detection system 900b includes bandpass optical components 921, 922, 923, and 924 (e.g., dichroic mirrors) for transmitting light of predetermined wavelengths to the optical detectors 914-917. In some examples, the optical component 921 is a 534 nm / 40 nm bandpass. In some examples, the optical component 922 is a 586 nm / 42 nm bandpass. In some examples, the optical component 923 is a 700 nm / 54 nm bandpass. In some examples, the optical component 924 is a 783 nm / 56 nm bandpass. The first number represents the center of the spectral band. The second number provides the range of the spectral band. Thus, a 510 / 20 filter extends 10 nm on either side of the center of the spectral band, i.e., from 500 nm to 520 nm.

[0119] Data signals generated in response to light detected in scattered light detection channels 911 and 912, bright-field light detection channel 913, and fluorescence detection channels 914-917 are processed by real-time digital processing by processors 950 and 951. Images 911a-917a can be generated in each light detection channel based on the data signals generated by processors 950 and 951. Image-enabled sorting is performed in response to a sorting signal generated by a sorting trigger 952. Sorting component 900c includes deflection plates 931 for deflecting particles into a sample container 932 or to a waste stream 933. In some examples, sorting component 900c is configured to sort particles using an enclosed particle sorting module, such as that described in U.S. Patent Application Publication No. 2017 / 0299493 (filed March 28, 2017, the disclosure of which is incorporated herein by reference). In certain embodiments, the sorting component 900c includes a sorting determination module having multiple sorting determination units, such as those described in U.S. Patent Application Publication No. 2020 / 0256781, the disclosure of which is incorporated herein by reference.

[0120] Non-transitory computer-readable storage medium Aspects of the present disclosure further include non-transitory computer-readable storage media having instructions for implementing the subject methods. The computer-readable storage media may be used on one or more computers for fully or partially automating systems for implementing the methods described herein. In certain embodiments, instructions according to the methods described herein may be encoded on a computer-readable medium in the form of "programming," and the term "computer-readable medium" as used herein refers to any non-transitory storage medium involved in providing instructions and data to a computer for execution and processing. Examples of suitable non-transitory storage media include floppy disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, DVD-ROMs, Blu-ray disks, solid-state disks, flash drives, and network-attached storage (NAS), whether such devices are internal or external to the computer. Files containing information may be "stored" on a computer-readable medium, where "storing" refers to recording information so that it can be accessed and retrieved at a later date by a computer. The computer-implemented methods described herein may be implemented using programming that can be written in one or more of any number of computer programming languages. Such languages ​​include, for example, Java, Python, Visual Basic, and C++, as well as many others.

[0121] In some embodiments, a target computer-readable storage medium includes a computer program stored thereon, the computer program including instructions that, when loaded into a computer, perform the methods of the present invention described herein, i.e., receive a secondary dataset comprising data points collected from a secondary sample, and calculate a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset comprising data points collected from a reference sample, the second dimensionality reduction calculated by the target computer-readable storage medium being compatible with the dimensionality-reduced reference dataset.

[0122] Computer Control System Aspects of the present disclosure further include a computer control system, the system including one or more computers for full or partial automation. In some embodiments, the system includes a computer having a non-transitory computer-readable storage medium storing a computer program, which, when loaded into the computer, performs the method of the present invention described herein, i.e., receives a secondary dataset including data points collected from a secondary sample, and calculates a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample. The calculated second dimensionality reduction is compatible with the dimensionality-reduced reference dataset.

[0123] The system may include a display and an operator input device. The operator input device may be, for example, a keyboard, a mouse, etc. The processing module includes a processor that accesses a memory in which instructions for executing the steps of the subject method are stored. The processing module may include an operating system, a graphical user interface (GUI) controller, a system memory, a memory storage device, an input / output controller, a cache memory, a data backup unit, and many other devices. The processor may be a commercially available processor or one of other processors that are or become available. The processor executes an operating system, which interfaces with firmware and hardware in well-known ways and facilitates the processor's coordination and execution of the functions of various computer programs, which may be written in various programming languages, such as Java, Perl, C++, Python, other high-level or low-level languages, and combinations thereof, as known in the art. The operating system typically cooperates with the processor to coordinate and execute the functions of the other components of the computer. The operating system also provides scheduling, input / output control, file and data management, memory management, and communication control and related services, all in accordance with known techniques. In some embodiments, the processor includes analog electronics that provide feedback control, such as negative feedback control.

[0124] System memory may be any of a variety of known or future memory storage devices. Examples include any commonly available random access memory (RAM), magnetic media such as a resident hard disk or tape, optical media such as a read-and-write compact disc, flash memory devices, or other memory storage devices. The memory storage device may be any of a variety of known or future devices, including a compact disc drive, tape drive, or diskette drive. Such types of memory storage devices typically read from and / or write to a program storage medium (not shown), such as a compact disc. Any of these program storage media, or others now in use or that may later be developed, may be considered a computer program product. As will be appreciated, these program storage media typically store computer software programs and / or data. Computer software programs, also known as computer control logic, are typically stored in system memory and / or program storage devices used in conjunction with the memory storage devices.

[0125] In some embodiments, a computer program product is described that includes a computer-usable medium having stored thereon control logic (a computer software program including program code). The control logic, when executed by a processor of a computer, causes the processor to perform the functions described herein. In other embodiments, some functions are implemented primarily in hardware, for example, using hardware state machines. Implementing a hardware state machine to perform the functions described herein will be apparent to one skilled in the art.

[0126] The memory may be any suitable device from which the processor can store and retrieve data, such as a magnetic, optical, or solid-state storage device (including a magnetic or optical disk, or tape, or RAM, or any other suitable device, fixed or portable). The processor may include a general-purpose digital microprocessor that is appropriately programmed from a computer-readable medium carrying the necessary program code. The programming may be provided remotely to the processor via a communications channel or may be pre-stored in a computer program product, such as memory or some other portable or fixed computer-readable storage medium using any of these devices in conjunction with the memory. For example, a magnetic or optical disk may carry the program and be read by a disk writer / reader. The system of the present invention also includes programming, e.g., in the form of a computer program product, algorithms for use in implementing the above-described methods. The programming according to the present invention may be recorded on a computer-readable medium, e.g., any medium that can be directly read and accessed by a computer. Such media include, but are not limited to, magnetic storage media such as floppy disks, hard disk storage media, and magnetic tape, optical storage media such as CD-ROM, storage media such as RAM, ROM, portable flash drives, and hybrids of these categories such as magnetic / optical storage media.

[0127] The processor may also have access to a communication channel for communicating with a user in a remote location, where remote location means that the user is not in direct contact with the system but relays input information to the input manager from an external device, such as a computer connected to a wide area network ("WAN"), a telephone network, a satellite network, or any other suitable communication channel, including a mobile phone (i.e., a smartphone).

[0128] In some embodiments, a system according to the present disclosure may be configured to include a communications interface. In some embodiments, the communications interface includes a receiver and / or a transmitter for communicating with a network and / or another device. The communications interface may be configured for wired or wireless communications, including, but not limited to, radio frequency (RF) communications (e.g., radio frequency identification (RFID), Zigbee communications protocol, Wi-Fi, infrared, wireless universal serial bus (USB), ultra-wideband (UWB), Bluetooth® communications protocol, and cellular communications such as code division multiple access (CDMA) or global system for mobile communications (GSM).

[0129] In one embodiment, the communication interface is configured to include one or more communication ports, e.g., a physical port or interface such as a USB port, a USB-C port, an RS-232 port, or any other suitable electrical connection port that enables data communication between the system of interest and other external devices, such as a computer terminal (e.g., in a doctor's office or hospital environment) configured for similar complementary data communication.

[0130] In one embodiment, the communication interface is configured for infrared communication, Bluetooth® communication, or any other suitable wireless communication protocol to enable the target system to communicate with computer terminals and / or other devices such as networks, communication-enabled mobile phones, personal digital assistants, or any other communication device that a user can integrate and use.

[0131] In one embodiment, the communication interface is configured to provide connectivity for data transfer using Internet Protocol (IP) over a cellular network, Short Message Service (SMS), a wireless connection to a personal computer (PC) in a local area network (LAN) connected to the Internet, or a Wi-Fi connection to the Internet at a Wi-Fi hotspot.

[0132] In one embodiment, the target system is configured to wirelessly communicate with a server device via a communications interface using common standards such as, for example, 802.11 or Bluetooth® RF protocols, or the IrDA infrared protocol. The server device may be another portable device, such as a smartphone, personal digital assistant (PDA), or notebook computer, or a larger device, such as a desktop computer, appliance, etc. In some embodiments, the server device has a display, such as a liquid crystal display (LCD), and input devices, such as buttons, a keyboard, a mouse, or a touchscreen.

[0133] In some embodiments, the communication interface is configured to automatically or semi-automatically communicate data stored in the target system, e.g., the optional data storage unit, with a network or server device using one or more of the communication protocols and / or mechanisms described above.

[0134] The output controller may include a controller for any of a variety of known display devices for presenting information to a user, whether human or machine, local or remote. When one of the display devices provides visual information, this information may typically be logically and / or physically organized as an array of graphical elements. A graphical user interface (GUI) controller provides a graphical input / output interface between the system and the user and may include any of a variety of known or future software programs for processing the user's input. The functional elements of the computer may communicate with each other via a system bus. Some of these communications may be achieved in alternative embodiments using a network or other type of remote communication. The output manager may also communicate information generated by the processing modules to a user at a remote location, for example, via the Internet, telephone, or satellite network, according to known techniques. Presentation of data by the output manager may be performed according to various known techniques. As some examples, the data may include SQL, HTML, or XML documents, email or other files, or other formats of data. The data may also include Internet URL addresses so that the user can retrieve additional SQL, HTML, XML, or other documents or data from remote sources. The one or more platforms present in the subject system may be of any type of known or future-developed computer platform, but they are typically of a class of computers commonly referred to as servers. However, they may also be mainframe computers, workstations, or other computer types. They may be connected via any known or future type of cabling or other communication system, including wireless systems, and may or may not be networked. They may be co-located or physically separated.In some cases, various operating systems may be employed on any computer platform depending on the type and / or manufacturer of the computer platform selected. Suitable operating systems include Windows NT, Windows XP, Windows 7, Windows 8, Windows 10, iOS, macOS, Linux, Ubuntu, Fedora, OS / 400, i5 / OS, IBM i, Android™, SGI IRIX, Oracle Solaris, etc.

[0135] FIG. 10 illustrates the general architecture of an exemplary computing device 1000 according to certain embodiments. The general architecture of the computing device 1000 illustrated in FIG. 10 includes an arrangement of computer hardware and software components. However, not all of these generally conventional elements need be shown to form an enabling disclosure. As illustrated, the computing device 1000 includes a processing unit 1010, a network interface 1020, a computer-readable medium drive 1030, an input / output device interface 1040, a display 1050, and input devices 1060, all of which can communicate with each other via a communication bus. The network interface 1020 can provide connections to one or more networks or computing systems. Thus, the processing unit 1010 can receive information and instructions from other computing systems or services via a network. The processing unit 1010 also communicates with a memory 1070 and can further communicate output information to an optional display 1050 via the input / output device interface 1040. For example, analysis software (e.g., data analysis software or program such as FlowJo™) stored as executable instructions in non-transitory memory of the analysis system can display flow cytometry event data to a user. The input / output device interface 1040 can also accept input from optional input devices 1060, such as a keyboard, mouse, digital pen, microphone, touch screen, gesture recognition system, voice recognition system, gamepad, accelerometer, gyroscope, or other input device.

[0136] The memory 1070 may include computer program instructions (grouped in some embodiments as modules or components) that the processing unit 1010 executes to implement one or more embodiments. The memory 1070 generally includes RAM, ROM, and / or other persistent, secondary, or non-transitory computer-readable media. The memory 1070 may store an operating system 1072 that communicates computer program instructions used by the processing unit 1010 in the general management and operation of the computing device 1000. Data may be stored on the data storage device 1090. The memory 1070 may further include computer program instructions and other information for implementing aspects of the present disclosure.

[0137] Utilities The present methods, systems, and computer-readable media can be employed when it is desirable to minimize the need to collect all samples upfront and concatenate them to calculate dimensionality reduction. The present invention enables a workflow in which a user collects one or more reference samples, calculates dimensionality reduction for this / these samples, and uses them as training (references) to calculate compatible dimensionality reductions for new samples in the future. The methods, systems, and computer-readable media can also be employed when it is desirable to have a quality score designed to indicate whether data is sufficiently stable over time.

[0138] Embodiments of the present invention are used in applications where cells prepared from biological samples may be desired for use in research, laboratory testing, or therapy. In some embodiments, the subject methods and devices may facilitate obtaining individual cells prepared from target fluid or tissue biological samples. For example, the subject methods and systems may facilitate obtaining cells from fluid or tissue samples used as research or diagnostic specimens for diseases such as cancer. Similarly, the subject methods and systems may facilitate obtaining cells from fluid or tissue samples used for therapy. The disclosed methods and devices enable the separation and collection of cells from biological samples (e.g., organs, tissues, tissue fragments, body fluids) with improved efficiency and lower cost compared to conventional flow cytometry systems.

[0139] kit Aspects of the present disclosure further include kits, which include storage media such as magneto-optical disks, CD-ROMs, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, DVD-ROMs, Blu-ray discs, solid-state disks, and network-attached storage (NAS). Any of these program storage media, or others currently in use or that may be later developed, may be included in the subject kits. In embodiments, the program storage media include instructions for dimensionality reduction according to the methods described herein. In embodiments, the instructions included on the computer-readable media provided in the subject kits, or portions thereof, may be implemented as software components of software for analyzing data. In these embodiments, a computer-controlled system according to the present disclosure may function as a software "plug-in" to an existing software package (e.g., FlowJo™).

[0140] In addition to the above components, the subject kits may (in some embodiments) further include instructions for, for example, installing a plug-in into an existing software package. These instructions may be present in a variety of forms, one or more of which may be present in the subject kits. One form in which these instructions may be present is information printed on a suitable medium or substrate, such as one or more pieces of paper on which the information is printed, kit packaging, a package insert, etc. Another form in which these instructions may be present is a computer-readable medium on which the information is recorded, such as a diskette, a compact disc (CD), a portable flash drive, etc. Another form in which these instructions may be present is a website address that can be used via the Internet to access the information at the removed site.

[0141] Notwithstanding the appended claims, the present invention may also be defined by the following clauses.

[0142] 1. A computer-implemented method of dimensionality reduction, comprising, via a processor: receiving a secondary data set including data points collected from the secondary sample; and 1. A computer-implemented method comprising: calculating a second dimensionality reduction of a secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample, wherein the second dimensionality reduction is compatible with the dimensionality-reduced reference dataset. 2. The computer-implemented method of clause 1, wherein calculating the second dimensionality reduction includes determining the data point collected from the reference sample that is closest to each data point collected from the secondary sample with respect to the given data parameter. 3. The computer-implemented method of clause 2, wherein computing the second dimensionality reduction includes computing k nearest neighbors in the reference dataset for each data point collected from the secondary sample. 4. The computer-implemented method of clause 3, wherein k ranges from 1 to m and m is half the number of data points in the reference sample. 5. The computer-implemented method of clause 4, wherein k is in the range of 2 to 5.

[0143] 6. The computer-implemented method of any one of clauses 3-5, comprising calculating the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, locality-sensitive hashing, a hierarchical navigable small world, approximate nearest neighbors with random projection trees, a GPU-based KNN search, or a brute-force KNN search. 7. The computer-implemented method of any one of clauses 3-6, further comprising calculating a distance between each data point collected from the secondary sample and each of the k nearest neighbors. 8. The computer-implemented method of clause 7, wherein the distance is a Manhattan distance. 9. The computer-implemented method of clause 7, wherein the distance is a Euclidean distance. 10. The computer-implemented method of clause 7, wherein the distance is a Chebyshev distance.

[0144] 11. The computer-implemented method of clause 7, wherein the distance is a Minkowski distance. 12. The computer-implemented method of clause 7, wherein the distance is a cosine distance. 13. The computer-implemented method of any one of clauses 7-12, comprising computing the second dimensionality reduction based on a weighted average of the k nearest neighbor positions. 14. The computer-implemented method of clause 13, further comprising calculating a weighted average using the weight obtained for each of the k nearest neighbors based on distance. 15. The computer-implemented method of clause 14, comprising obtaining the weights as follows:

[0145]

number

[0146] 16. The computer-implemented method of clause 14, comprising obtaining the weights as follows:

[0147]

number

[0148] 17. The computer-implemented method of claim 14, comprising obtaining the weights as follows:

[0149]

number

[0150] 18. The computer-implemented method of any one of clauses 1-17, wherein the secondary dataset comprises data points relating to a plurality of parameters that match the reference dataset. 19. The computer-implemented method of clause 18, further comprising generating derived parameters for the secondary dataset based on parameters of the dimensionality-reduced reference dataset. 20. The computer-implemented method of clause 19, wherein the derived parameter is a fluorescence parameter, a scattering parameter, an imaging parameter, or a categorical parameter. 21. The computer-implemented method of clause 20, wherein the derived parameter is a fluorescence parameter. 22. The computer-implemented method of any one of clauses 19-21, wherein generating derived parameters for the secondary data set comprises linear interpolation.

[0151] 23. The computer-implemented method of any one of clauses 7-22, further comprising calculating an input quality score indicating the degree to which each data point collected from the secondary sample is related to a data point in the reference dataset. 24. The computer-implemented method of clause 23, wherein calculating the input quality score includes obtaining a normalized average distance to k nearest neighbors of each data point collected from the secondary sample. 25. The computer-implemented method of any one of clauses 7-24, further comprising calculating an output quality score that measures separation of k nearest neighbors in the dimensionally reduced first data set. 26. The computer-implemented method of clause 25, wherein calculating the output quality score includes obtaining a normalized average distance of each of the k nearest neighbors in the dimensionality-reduced reference dataset. 27. The computer-implemented method of any one of clauses 1-26, wherein the reference dataset and the secondary dataset consist of flow cytometer data.

[0152] 28. The computer-implemented method of any one of clauses 1-27, wherein the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled reference dataset. 29. The computer-implemented method of any one of clauses 1 to 28, further comprising: receiving a reference dataset; and performing a first dimensionality reduction. 30. The computer-implemented method of any one of clauses 1-29, further comprising transforming the reference dataset and the secondary dataset. 31. The computer-implemented method of clause 30, comprising transforming the parameters of the reference data set and the parameters of the secondary data set with a linear function, a logarithmic function, a hyperbolic arcsine function, or a biexponential function. 32. The computer-implemented method of any one of clauses 1-31, comprising calculating a dimensionality reduction for a plurality of secondary data sets comprising data points collected from a plurality of secondary samples.

[0153] 33. A system including a processor including a memory operatively coupled to the processor, the memory storing instructions that, when executed by the processor, cause the processor to: receiving a secondary data set including data points collected from the secondary sample; and The system calculates a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample, wherein the second dimensionality reduction is compatible with the dimensionality-reduced reference dataset. 34. The system of clause 33, wherein calculating the second dimensionality reduction includes determining a data point collected from the reference sample that is closest to each data point collected from the secondary sample with respect to a given data parameter. 35. The system of clause 34, wherein computing the second dimensionality reduction includes computing k nearest neighbors in the reference dataset for each data point collected from the secondary sample. 36. The system of clause 35, wherein k ranges from 1 to m, and m is half the number of data points in the reference sample. 37. The system of clause 36, wherein k is in the range of 2 to 5.

[0154] 38. The system of any one of clauses 35-37, wherein the processor is configured to calculate the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, locality sensitive hashing, a hierarchical navigable small world, approximate nearest neighbors with random projection trees, a GPU-based KNN search, or a brute force KNN search. 39. The system of any one of clauses 35 to 38, wherein the processor is configured to calculate the distance between each data point collected from the secondary sample and each of the k nearest neighbors. 40. The system of clause 39, wherein the distance is Manhattan distance. 41. The system of clause 39, wherein the distance is a Euclidean distance. 42. The system of clause 39, wherein the distance is the Chebyshev distance.

[0155] 43. The system of clause 39, wherein the distance is a Minkowski distance. 44. The system of clause 39, wherein the distance is a cosine distance. 45. The system of any one of clauses 39-44, wherein the processor is configured to calculate the second dimensionality reduction based on a weighted average of the positions of the k nearest neighbors. 46. ​​The system of clause 45, wherein the processor is configured to calculate a weighted average using the weight obtained for each of the k nearest neighbors based on distance.

[0156] 47. The system of clause 46, wherein the processor is configured to obtain the weights as follows:

[0157]

number

[0158] 48. The system of clause 46, wherein the processor is configured to obtain the weights as follows:

[0159]

number

[0160] 49. The system of clause 46, wherein the processor is configured to obtain the weights as follows:

[0161]

number

[0162] 50. A system according to any one of clauses 33 to 49, wherein the secondary dataset comprises data points relating to a plurality of parameters that match the reference dataset. 51. The system of clause 50, wherein the processor is configured to generate derived parameters for the secondary dataset based on parameters of the dimensionality-reduced reference dataset. 52. The system of clause 51, wherein the derived parameter is a fluorescence parameter, a scattering parameter, an imaging parameter, or a categorical parameter. 53. The system of clause 52, wherein the derived parameter is a fluorescence parameter. 54. The system of any one of clauses 51 to 53, wherein generating derived parameters for the secondary data set includes linear interpolation. 55. A system described in any one of clauses 39 to 54, wherein the processor is configured to calculate an input quality score indicating the degree to which each data point collected from the secondary sample is related to a data point in the reference dataset.

[0163] 56. The system of clause 55, wherein calculating the input quality score includes obtaining a normalized average distance to k nearest neighbors of each data point collected from the secondary sample. 57. The system of any one of clauses 39 to 56, wherein the processor is configured to calculate an output quality score that measures separation of k nearest neighbors in the dimensionality-reduced first data set. 58. The system of clause 57, wherein calculating the output quality score includes obtaining a normalized average distance for each of the k nearest neighbors in the dimensionality-reduced reference dataset. 59. A system according to any one of clauses 33 to 58, wherein the processor is in a data-receiving relationship with a flow cytometer, and the reference data set and the secondary data set are comprised of flow cytometer data. 60. The system of any one of clauses 33 to 59, wherein the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled reference dataset. 61. The system of any one of clauses 33 to 60, wherein the processor: Receive a reference dataset, and The system is configured to perform a first dimensionality reduction.

[0164] 62. A system according to any one of clauses 33 to 61, wherein the processor is configured to transform the reference dataset and the secondary dataset. 63. The system of clause 62, wherein the processor is configured to transform the parameters of the reference data set and the parameters of the secondary data set with a linear function, a logarithmic function, a hyperbolic arcsine function, or a biexponential function. 64. The system of any one of clauses 33-63, wherein the processor is configured to calculate a dimensionality reduction of a plurality of secondary data sets comprising data points collected from a plurality of secondary samples.

[0165] 65. A non-transitory computer-readable storage medium having stored thereon instructions for dimensionality reduction by a method including: receiving a secondary data set including data points collected from the secondary sample; and A non-transitory computer-readable storage medium having stored thereon instructions for dimensionality reduction by a method comprising: calculating a second dimensionality reduction of a secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample, wherein the second dimensionality reduction is compatible with the dimensionally reduced reference dataset. 66. The non-transitory computer-readable storage medium of clause 65, wherein calculating the second dimensionality reduction includes determining a data point collected from the reference sample that is closest to each data point collected from the secondary sample with respect to a given data parameter. 67. The non-transitory computer-readable storage medium of clause 66, wherein calculating the second dimensionality reduction includes calculating k nearest neighbors in the reference dataset for each data point collected from the secondary sample. 68. The non-transitory computer-readable storage medium of clause 67, wherein k ranges from 1 to m and m is half the number of data points in the reference sample. 69. The non-transitory computer-readable storage medium of clause 68, wherein k is in the range of 2 to 5.

[0166] 70. The non-transitory computer-readable storage medium of any one of clauses 67-69, wherein the method includes calculating the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, locality sensitive hashing, a hierarchical navigable small world, approximate nearest neighbors with random projection trees, a GPU-based KNN search, or a brute force KNN search, or a brute force KNN search. 71. The non-transitory computer-readable storage medium of any one of clauses 67-70, wherein the method includes calculating a distance between each data point collected from the secondary sample and each of the k nearest neighbors. 72. The non-transitory computer-readable storage medium of clause 71, wherein the distance is a Manhattan distance. 73. The non-transitory computer-readable storage medium of clause 71, wherein the distance is a Euclidean distance. 74. The non-transitory computer-readable storage medium of clause 71, wherein the distance is a Chebyshev distance. 75. The non-transitory computer-readable storage medium of clause 71, wherein the distance is a Minkowski distance. 76. The non-transitory computer-readable storage medium of clause 71, wherein the distance is a cosine distance. 77. The non-transitory computer-readable storage medium of any one of clauses 71-76, wherein the method includes calculating the second dimensionality reduction based on a weighted average of the positions of the k nearest neighbors. 78. The non-transitory computer-readable storage medium of clause 77, wherein the method further comprises calculating a weighted average using the weight obtained for each of the k nearest neighbors based on the distance.

[0167] 79. The non-transitory computer-readable storage medium of clause 78, wherein the processor is configured to obtain the weights as follows:

[0168]

number

[0169] 80. The non-transitory computer-readable storage medium of clause 78, wherein the method includes obtaining the weights as follows:

[0170]

number

[0171] 81. The non-transitory computer-readable storage medium of clause 78, wherein the method includes obtaining the weights as follows:

[0172]

number

[0173] 82. A non-transitory computer-readable storage medium according to any one of clauses 65 to 81, wherein the secondary dataset comprises data points relating to a plurality of parameters that match the reference dataset. 83. The non-transitory computer-readable storage medium of clause 82, wherein the method further comprises creating derived parameters of the secondary data set based on parameters of the dimensionality-reduced reference data set. 84. The non-transitory computer-readable storage medium of clause 83, wherein the derived parameter is a fluorescence parameter, a scattering parameter, an imaging parameter, or a categorical parameter. 85. The non-transitory computer-readable storage medium of clause 89, wherein the derived parameter is a fluorescence parameter. 86. The non-transitory computer-readable storage medium of any one of clauses 83 to 85, wherein creating derived parameters of the secondary data set includes linear interpolation.

[0174] 87. The non-transitory computer-readable storage medium of any one of clauses 71 to 86, wherein the method further comprises calculating an input quality score indicating the degree to which each data point collected from the secondary sample is related to a data point in the reference dataset. 88. The non-transitory computer-readable storage medium of clause 87, wherein calculating the input quality score includes obtaining a normalized average distance to k nearest neighbors of each data point collected from the secondary samples. 89. The non-transitory computer-readable storage medium of any one of clauses 71-88, wherein the method further comprises calculating an output quality score that measures separation of k nearest neighbors in the dimensionality-reduced first data set. 90. The non-transitory computer-readable storage medium of clause 89, wherein calculating the output quality score includes obtaining a normalized average distance of each of the k nearest neighbors in the dimensionality-reduced reference dataset. 91. The non-transitory computer-readable storage medium of any one of clauses 65 to 90, wherein the reference data set and the secondary data set are comprised of flow cytometer data. 92. The non-transitory computer-readable storage medium of any one of clauses 65 to 91, wherein the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled dataset.

[0175] 93. The computer-implemented method of any one of clauses 65-92, wherein the method comprises: receiving a reference dataset; and performing a first dimensionality reduction. 94. The non-transitory computer-readable storage medium of any one of clauses 65 to 93, wherein the method further comprises transforming the reference dataset and the secondary dataset. 95. The non-transitory computer-readable storage medium of clause 94, wherein the method includes transforming the parameters of the reference data set and the parameters of the secondary data set with a linear function, a logarithmic function, a hyperbolic arcsine function, or a bi-exponential function. 96. The non-transitory computer-readable storage medium of any one of clauses 65-95, wherein the method includes calculating a dimensionality reduction of a plurality of secondary data sets including data points collected from a plurality of secondary samples.

[0176] The following are offered by way of example and not by way of limitation.

[0177] experiment Flow cytometer data for the first sample (sample S1) and the second sample (sample S2) were obtained from the public FlowRepository dataset FR-FCM-ZYX4. tSNE and UMAP were calculated for sample S1 (Figure 11A), and compatible tSNE and UMAP were derived for sample S2 (Figure 11B) using the algorithms described in Figures 2-3. A lymphocyte gate on FS / SS, followed by examination of the lymphocyte population in tSNE and UMAP space, was performed on both samples as a "sanity check." Plots showing the matching embedding using overlay are shown in Figures 12A-12B. Plots showing the matching embedding using overlay with switched sample order are shown in Figures 13A-13B. Input and output quality scores were also calculated for tSNE (Figure 14A) and UMAP (Figure 14B). As shown in Figures 14A-14B, the input and output quality scores can be utilized to gate events with high-quality embedding results. These scores allow users to focus further downstream analysis on data with high-quality embedding results, i.e., events for which the derived embedding represents a true approximation of the event's ideal location in a compatible embedding. Specifically, there are two main reasons why this may not be true for a particular event: (i) the algorithm does not have enough data to place such an event in a compatible embedding because a similar event population was not present in the reference sample S1, or (ii) a similar event population from the reference sample S1 is not consistently placed in the reduced-dimensionality embedding. The former is indicated by a poor input quality score for such an event, and the latter is indicated by a poor output score for such an event. Furthermore, the output quality score can be used as a quality indicator of the embedding performed on the reference sample S1 using any of the state-of-the-art dimensionality reduction techniques.Specifically, poor output quality scores for most events may indicate that the dimensionality reduction performed on the reference sample S1 using such state-of-the-art dimensionality reduction techniques is not a faithful representation of the data with respect to preserving what is usually referred to as the local structure of the data, i.e., events R that are close to each other are not close to each other in D.

[0178] Although the foregoing invention has been described in some detail by way of illustration and example for clarity of understanding, it will be readily apparent to those skilled in the art in light of the teachings of the invention that certain changes and modifications can be made without departing from the spirit or scope of the appended claims.

[0179] Accordingly, the foregoing merely illustrates the principles of the present invention. It will be appreciated that those skilled in the art will be able to devise various configurations, not explicitly described or shown herein, which embody the principles of the present invention and are within its spirit and scope. Furthermore, all examples and conditional language recited herein are intended primarily to aid the reader in understanding the principles of the present invention and the concepts the inventors have contributed to advancing the art, and should not be construed as being limited to such specifically recited examples and conditions. Furthermore, all statements herein reciting principles, aspects, and embodiments of the present invention, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof. Furthermore, such equivalents are intended to include both currently known equivalents and equivalents developed in the future, i.e., any elements developed that perform the same function, regardless of structure. Furthermore, nothing disclosed herein is intended as a dedication to the public, regardless of whether such disclosure is expressly recited in the claims.

[0180] Accordingly, the scope of the present invention is not intended to be limited to the exemplary embodiments shown and described herein. Rather, the scope and spirit of the present invention are embodied by the appended claims. In the claims, 35 U.S.C. §112(f) or 35 U.S.C. §112(6) is expressly defined as being invoked for a limitation in a claim only if the exact phrase "means for" or the exact phrase "step" appears at the beginning of such limitation in the claim. If such exact phrases are not used in a claim limitation, 35 U.S.C. §112(f) or 35 U.S.C. §112(6) is not invoked.

[0181] CROSS-REFERENCE TO RELATED APPLICATIONS Pursuant to 35 U.S.C. § 119(e), this application claims priority to the filing date of U.S. Provisional Patent Application No. 63 / 421,839, filed November 2, 2022, the disclosure of which is incorporated herein by reference in its entirety.

Claims

1. 1. A computer-implemented method of dimensionality reduction, comprising, via a processor: receiving a secondary data set including data points collected from the secondary sample; and 1. A computer-implemented method comprising: calculating a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset comprising data points collected from a reference sample, wherein the second dimensionality reduction is compatible with the dimensionally reduced reference dataset.

2. 2. The computer-implemented method of claim 1, wherein computing the second dimensionality reduction comprises determining a data point collected from the reference sample that is closest to each data point collected from the secondary sample with respect to a given data parameter.

3. 3. The computer-implemented method of claim 2, wherein computing the second dimensionality reduction comprises computing k nearest neighbors in the reference data set for each data point collected from the secondary sample.

4. 4. The computer-implemented method of claim 3, wherein k ranges from 1 to m, and m is half the number of data points in the reference sample.

5. 5. The computer-implemented method of claim 4, wherein k is in the range of 2 to 5.

6. 6. The computer-implemented method of claim 3, comprising computing the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, a locality-sensitive hash, a hierarchical navigable small world, an approximate nearest neighbor with randomly projected trees, a GPU-based KNN search, or a brute-force KNN search.

7. The computer-implemented method of any one of claims 3 to 6, further comprising calculating a distance between each data point collected from the secondary sample and each of the k nearest neighbors.

8. The computer-implemented method of claim 7 , wherein the distance is a Manhattan distance.

9. The computer-implemented method of claim 7 , wherein the distance is a Euclidean distance.

10. The computer-implemented method of claim 7 , wherein the distance is a Chebyshev distance.

11. The computer-implemented method of claim 7 , wherein the distance is a Minkowski distance.

12. The computer-implemented method of claim 7 , wherein the distance is a cosine distance.

13. The computer-implemented method of any one of claims 7 to 12, comprising calculating the second dimensionality reduction based on a weighted average of the k nearest neighbor positions.

14. The computer-implemented method of claim 13 , further comprising calculating the weighted average using a weight obtained for each of the k nearest neighbors based on the distance.

15. 15. The computer-implemented method of claim 14, comprising obtaining the weights as follows: [Equation 1]

16. 15. The computer-implemented method of claim 14, comprising obtaining the weights as follows: [Equation 2]

17. 15. The computer-implemented method of claim 14, comprising obtaining the weights as follows: [Equation 3]

18. The computer-implemented method of any one of claims 1 to 17, wherein the secondary data set comprises data points associated with a plurality of parameters that match the reference data set.

19. 20. The computer-implemented method of claim 18, further comprising generating derived parameters for the secondary data set based on parameters of the dimensionally reduced reference data set.

20. 20. The computer-implemented method of claim 19, wherein the derived parameter is a fluorescence parameter, a scattering parameter, an imaging parameter, or a categorical parameter.

21. 21. The computer-implemented method of claim 20, wherein the derived parameter is a fluorescence parameter.

22. The computer-implemented method of any one of claims 19 to 21, wherein creating the derived parameters of the secondary data set comprises linear interpolation.

23. 23. The computer-implemented method of claim 7, further comprising calculating an input quality score indicating the degree to which each data point collected from the secondary sample is related to the data point in the reference dataset.

24. 24. The computer-implemented method of claim 23, wherein calculating the input quality score comprises taking a normalized average distance of each data point collected from the secondary sample to the k nearest neighbors.

25. 25. The computer-implemented method of claim 7, further comprising calculating an output quality score that measures separation of the k nearest neighbors in the dimensionally reduced first data set.

26. 26. The computer-implemented method of claim 25, wherein calculating the output quality score comprises obtaining a normalized average distance of each of the k nearest neighbors in the dimensionality-reduced reference dataset.

27. 27. The computer-implemented method of any one of claims 1 to 26, wherein the reference data set and the secondary data set consist of flow cytometer data.

28. The computer-implemented method of any one of claims 1 to 27, wherein the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled reference dataset.

29. receiving the reference dataset; and performing said first dimensionality reduction; A computer-implemented method according to any preceding claim, comprising:

30. The computer-implemented method of any one of claims 1 to 29, further comprising transforming the reference dataset and the secondary dataset.

31. 31. The computer-implemented method of claim 30, comprising transforming the parameters of the reference data set and the parameters of the secondary data set with a linear function, a logarithmic function, a hyperbolic arcsine function, or a bi-exponential function.

32. 32. A computer-implemented method according to any preceding claim, comprising calculating a dimensionality reduction of a plurality of secondary data sets comprising data points collected from a plurality of secondary samples.

33. 1. A system including a processor including a memory operatively coupled to the processor, the memory storing instructions that, when executed by the processor, cause the processor to: receiving a secondary data set including data points collected from the secondary sample; and and computing a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample, the second dimensionality reduction being compatible with the dimensionality-reduced reference dataset.

34. 34. The system of claim 33, wherein calculating the second dimensionality reduction comprises determining a data point collected from the reference sample that is closest to each data point collected from the secondary sample for a given data parameter.

35. 35. The system of claim 34, wherein computing the second dimensionality reduction comprises computing k nearest neighbors in the reference data set for each data point collected from the secondary sample.

36. 36. The system of claim 35, wherein k ranges from 1 to m, and m is half the number of data points in the reference sample.

37. 37. The system of claim 36, wherein k is in the range of 2 to 5.

38. 38. The system of any one of claims 35 to 37, wherein the processor is configured to compute the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, locality sensitive hashing, a hierarchical navigable small world, approximate nearest neighbors with randomly projected trees, a GPU-based KNN search, or a brute force KNN search.

39. A system according to any one of claims 35 to 38, wherein the processor is configured to calculate the distance between each data point collected from the secondary samples and each of the k nearest neighbors.

40. 40. The system of claim 39, wherein the distance is a Manhattan distance.

41. 40. The system of claim 39, wherein the distance is a Euclidean distance.

42. 40. The system of claim 39, wherein the distance is a Chebyshev distance.

43. 40. The system of claim 39, wherein the distance is a Minkowski distance.

44. 40. The system of claim 39, wherein the distance is a cosine distance.

45. The system of any one of claims 39 to 44, wherein the processor is configured to calculate the second dimensionality reduction based on a weighted average of the k nearest neighbor positions.

46. 46. ​​The system of claim 45, wherein the processor is configured to calculate the weighted average using a weight obtained for each of the k nearest neighbors based on the distance.

47. 47. The system of claim 46, wherein the processor is configured to obtain the weights as follows: [Equation 4]

48. 47. The system of claim 46, wherein the processor is configured to obtain the weights as follows: [Equation 5]

49. 47. The system of claim 46, wherein the processor is configured to obtain the weights as follows: [Equation 6]

50. 50. The system of any one of claims 33 to 49, wherein the secondary data set comprises data points associated with a plurality of parameters that match the reference data set.

51. 51. The system of claim 50, wherein the processor is configured to generate derived parameters for the secondary data set based on parameters of the dimensionally reduced reference data set.

52. 52. The system of claim 51, wherein the derived parameter is a fluorescence parameter, a scattering parameter, an imaging parameter, or a categorical parameter.

53. 53. The system of claim 52, wherein the derived parameter is a fluorescence parameter.

54. 54. The system of any one of claims 51 to 53, wherein generating the derived parameters of the secondary data set comprises linear interpolation.

55. 55. The system of any one of claims 39 to 54, wherein the processor is configured to calculate an input quality score indicating the degree to which each data point collected from the secondary sample is related to the data point in the reference dataset.

56. 56. The system of claim 55, wherein calculating the input quality score comprises obtaining a normalized average distance of each data point collected from the secondary sample to its k nearest neighbors.

57. 57. The system of any one of claims 39 to 56, wherein the processor is configured to calculate an output quality score measuring separation of the k nearest neighbors in the dimensionality-reduced first data set.

58. 58. The system of claim 57, wherein calculating the output quality score comprises obtaining a normalized average distance of each of the k nearest neighbors in the dimensionality-reduced reference dataset.

59. 59. The system of any one of claims 33 to 58, wherein the processor is in data-receiving relationship with a flow cytometer, and wherein the reference data set and the secondary data set comprise flow cytometer data.

60. The system of any one of claims 33 to 59, wherein the reference dataset is a uniformly downsampled reference dataset or a density-based downsampled reference dataset.

61. the processor: receiving the reference dataset; and configured to perform the first dimensionality reduction; A system according to any one of claims 33 to 60.

62. The system of any one of claims 33 to 61, wherein the processor is configured to transform the reference data set and the secondary data set.

63. 63. The system of claim 62, wherein the processor is configured to transform the parameters of the reference data set and the parameters of the secondary data set with a linear function, a logarithmic function, a hyperbolic arcsine function, or a bi-exponential function.

64. 64. The system of any one of claims 33 to 63, wherein the processor is configured to calculate a dimensionality reduction of a plurality of secondary data sets comprising data points collected from a plurality of secondary samples.

65. receiving a secondary data set including data points collected from the secondary sample; and calculating a second dimensionality reduction of the secondary dataset based on a first dimensionality reduction of a reference dataset including data points collected from a reference sample, wherein the second dimensionality reduction is compatible with the dimensionality-reduced reference dataset; A non-transitory computer-readable storage medium having stored thereon instructions for dimensionality reduction by a method comprising:

66. 66. The non-transitory computer-readable storage medium of claim 65, wherein calculating the second dimensionality reduction comprises determining a data point collected from the reference sample that is closest to each data point collected from the secondary sample with respect to a given data parameter.

67. 67. The non-transitory computer-readable storage medium of claim 66, wherein computing the second dimensionality reduction comprises computing k nearest neighbors in the reference data set for each data point collected from the secondary sample.

68. 68. The non-transitory computer-readable storage medium of claim 67, wherein k ranges from 1 to m, and m is half the number of data points in the reference sample.

69. 69. The non-transitory computer-readable storage medium of claim 68, wherein k is in the range of 2 to 5.

70. 70. The non-transitory computer-readable storage medium of any one of claims 67 to 69, wherein the method comprises computing the k nearest neighbors using a vantage point tree, a k-dimensional tree, a ball tree, a cover tree, locality sensitive hashing, a hierarchical navigable small world, approximate nearest neighbors with randomly projected trees, a GPU-based KNN search, or a brute force KNN search, or a brute force KNN search.

71. 71. The non-transitory computer-readable storage medium of any one of claims 67 to 70, wherein the method comprises calculating a distance between each data point collected from the secondary sample and each of the k nearest neighbors.

72. 72. The non-transitory computer-readable storage medium of claim 71, wherein the distance is a Manhattan distance.

73. 72. The non-transitory computer-readable storage medium of claim 71, wherein the distance is a Euclidean distance.

74. 72. The non-transitory computer-readable storage medium of claim 71, wherein the distance is a Chebyshev distance.

75. 72. The non-transitory computer-readable storage medium of claim 71, wherein the distance is a Minkowski distance.

76. 72. The non-transitory computer-readable storage medium of claim 71, wherein the distance is a cosine distance.

77. 77. The non-transitory computer-readable storage medium of any one of claims 71 to 76, wherein the method comprises calculating the second dimensionality reduction based on a weighted average of the k nearest neighbor locations.

78. 78. The non-transitory computer-readable storage medium of claim 77, wherein the method further comprises calculating the weighted average using a weight obtained for each of the k nearest neighbors based on the distance.

79. 79. The non-transitory computer-readable storage medium of claim 78, wherein the processor is configured to obtain the weights as follows: [Equation 7]

80. 79. The non-transitory computer-readable storage medium of claim 78, wherein the method comprises obtaining the weights as follows: [Equation 8]

81. 79. The non-transitory computer-readable storage medium of claim 78, wherein the method comprises obtaining the weights as follows: [Equation 9]

82. 82. The non-transitory computer-readable storage medium of any one of claims 65 to 81, wherein the secondary data set comprises data points associated with a plurality of parameters that match the reference data set.

83. 83. The non-transitory computer-readable storage medium of claim 82, wherein the method further comprises creating derived parameters of the secondary data set based on parameters of the dimensionally reduced reference data set.

84. 84. The non-transitory computer-readable storage medium of claim 83, wherein the derived parameter is a fluorescence parameter, a scattering parameter, an imaging parameter, or a categorical parameter.

85. 90. The non-transitory computer-readable storage medium of claim 89, wherein the derived parameter is a fluorescence parameter.

86. 86. The non-transitory computer-readable storage medium of any one of claims 83 to 85, wherein creating the derived parameters of the secondary data set comprises linear interpolation.

87. 87. The non-transitory computer-readable storage medium of any one of claims 71 to 86, wherein the method further comprises calculating an input quality score indicating the degree to which each data point collected from the secondary sample is related to the data point in the reference dataset.

88. 88. The non-transitory computer-readable storage medium of claim 87, wherein calculating the input quality score comprises obtaining a normalized average distance of each data point collected from the secondary sample to the k nearest neighbors.

89. 89. The non-transitory computer-readable storage medium of any one of claims 71 to 88, wherein the method further comprises calculating an output quality score measuring separation of the k nearest neighbors in the dimensionality-reduced first data set.

90. 90. The non-transitory computer-readable storage medium of claim 89, wherein calculating the output quality score comprises obtaining a normalized average distance of each of the k nearest neighbors in the dimensionality-reduced reference data set.

91. 91. The non-transitory computer-readable storage medium of any one of claims 65 to 90, wherein the reference data set and the secondary data set consist of flow cytometer data.

92. 92. The non-transitory computer-readable storage medium of any one of claims 65 to 91, wherein the reference dataset is a uniformly down-sampled reference dataset or a density-based down-sampled dataset.

93. receiving the reference dataset; and performing said first dimensionality reduction; 93. A computer-implemented method according to any one of claims 65 to 92, comprising:

94. 94. The non-transitory computer-readable storage medium of any one of claims 65 to 93, wherein the method further comprises transforming the reference data set and the secondary data set.

95. 95. The non-transitory computer-readable storage medium of claim 94, wherein the method comprises transforming parameters of the reference data set and parameters of the secondary data set with a linear function, a logarithmic function, a hyperbolic arcsine function, or a bi-exponential function.

96. 96. The non-transitory computer-readable storage medium of any one of claims 65 to 95, wherein the method comprises computing a dimensionality reduction of a plurality of secondary data sets comprising data points collected from a plurality of secondary samples.

Images