Information processing method and system for flow cytometer and flow cytometer

Abstract

A method and system for information processing in flow cytometry, as well as a flow cytometer, are disclosed. The information processing method includes: performing a first clustering process based on detection data of a sample loaded in the flow cytometer to determine multiple cell populations in the sample; identifying one of the multiple cell populations as a target cell population according to predetermined settings; and determining negative and positive cell populations by performing a second clustering process on the cells in the target cell population.

Description

Technical Field

[0001] This disclosure relates to flow cytometry detection, and more specifically, to information processing methods and systems for use in flow cytometers, and to flow cytometers themselves. Background Technology

[0002] Flow cytometry is a biological technique used to count and sort tiny particles suspended in a fluid. In a typical flow cytometer, cells stained with fluorescent dyes are illuminated using a suitable light source. The cells are excited and emit fluorescence, which is collected and converted into an electrical signal by a photoelectric converter. The electrical signal is then amplified and input into a computer analyzer for quantitative analysis and sorting of the cells.

[0003] Typically, multiple fluorescent dyes are used to stain the sample to be tested. Because the cells in the sample have different characteristics, different cells can be stained with different fluorescence, thus emitting different fluorescence when excited. By collecting and analyzing the total fluorescence emitted by all excited cells, the cells stained by each fluorescent dye and their number can be identified, thereby enabling cell sorting and counting. In practice, users need to manually select the target cell population (e.g., lymphocytes) in the sample, and manually select positive and negative cell populations for each fluorescence. This is usually not a problem in traditional flow cytometers, as they only support a small number of fluorescences. However, for spectroscopic flow cytometers that support dozens of fluorescences, selecting positive and negative cell populations one by one for dozens of fluorescences will consume a significant amount of time and effort. For example, for 40 fluorescences, it may take one or two hours for a user to complete the selection of positive and negative cell populations.

[0004] On the other hand, autofluorescence of samples has always been a challenge in flow cytometry because it can adversely affect the detection of fluorescence emitted by excited samples. Users may sometimes not even be aware that multiple autofluorescences may exist in a sample, or even if they are aware that multiple autofluorescences may exist, it is difficult for them to identify and remove the autofluorescence. Summary of the Invention

[0005] This disclosure is intended to provide information processing techniques for flow cytometers that substantially avoid one or more problems caused by the limitations and drawbacks of the prior art.

[0006] According to one aspect of this disclosure, an information processing method for a flow cytometer is provided, comprising: performing a first clustering process based on detection data of a sample loaded in the flow cytometer to determine a plurality of cell populations in the sample; determining one of the plurality of cell populations as a target cell population according to a predetermined setting; and determining a negative cell population and a positive cell population by performing a second clustering process on the cells in the target cell population.

[0007] According to another aspect of this disclosure, an information processing method for a flow cytometer is provided, comprising: performing clustering processing based on detection data of samples loaded in the flow cytometer, wherein the samples are not stained with fluorescent dyes and the samples emit at least one autofluorescence; and determining at least one population of autofluorescent cells emitting the at least one autofluorescence based on the result of the clustering processing.

[0008] According to another aspect of this disclosure, an information processing system for a flow cytometer is provided, comprising a processing unit configured to: perform a first clustering process based on detection data of a sample loaded in the flow cytometer to determine a plurality of cell populations in the sample; determine one of the plurality of cell populations as a target cell population according to a predetermined setting; and determine a negative cell population and a positive cell population by performing a second clustering process on the cells in the target cell population.

[0009] According to another aspect of this disclosure, an information processing system for a flow cytometer is provided, comprising a processing unit configured to: perform clustering processing based on detection data of samples loaded in the flow cytometer, wherein the samples are not stained with fluorescent dyes and the samples emit at least one autofluorescence; and determine at least one population of autofluorescent cells emitting the at least one autofluorescence, based on the result of the clustering processing.

[0010] According to another aspect of this disclosure, a flow cytometer including the aforementioned information processing system is provided.

[0011] According to another aspect of this disclosure, a non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to perform the aforementioned information processing method.

[0012] In accordance with other aspects of this disclosure, computer program code and computer program products for implementing the above information processing methods are also provided. Attached Figure Description

[0013] Figure 1 (a) schematically shows a scatter plot of FSC-SSC. Figure 1 (b) schematically illustrates the target cell population.

[0014] Figure 2 A flowchart of a method for determining cell populations according to this disclosure is shown.

[0015] Figure 3 A curve illustrating the relationship between the number of groups and the profile coefficient according to this disclosure is shown schematically.

[0016] Figure 4 Histograms including negative and positive cell populations are schematically shown.

[0017] Figure 5 (a) schematically illustrates the negative and positive clusters obtained by clustering according to this disclosure. Figure 5 (b) schematically shows the final negative and positive cell populations.

[0018] Figure 6 A flowchart is shown of the overall method for determining the target cell population and negative and positive cell populations according to this disclosure.

[0019] Figure 7 (a) schematically illustrates the three groups that can be identified in the FSC-SSC scatter plot. Figure 7 (b) schematically illustrates the view obtained by mapping the dimensionality-reduced data onto a two-dimensional plane, in which more groups can be identified.

[0020] Figure 8 A flowchart of a method for detecting autofluorescence according to this disclosure is shown.

[0021] Figure 9 An exemplary configuration block diagram of computer hardware implementing the present disclosure is shown. Detailed Implementation

[0022] The specific implementation methods according to this disclosure are described in detail below with reference to the accompanying drawings.

[0023] After loading fluorescently stained cell samples into a flow cytometer, the flow cytometer can generate an FSC-SSC scatter plot of the sample. The FSC-SSC scatter plot can be used to group cells based on cell size and granularity, such as into lymphocyte populations, monocyte populations, neutrophil populations, etc. Figure 1 (a) schematically illustrates an FSC-SSC scatter plot. According to conventional techniques, users need to manually gating the target cell population (e.g., lymphocyte population) in the FSC-SSC scatter plot.

[0024] This disclosure provides a method to help users select target cell populations, eliminating the need for manual selection. For example... Figure 1 As shown in (a), cells in the FSC-SSC scatter plot typically form one or more roughly spherical clusters. Therefore, clustering algorithms such as K-Means, DBSCAN (Density-Based Spatial Clustering of Applications with Noise), GMM (Gaussian Mixed Model), and DPGMM (Dirichlet Process GMM) can be used to determine the individual cell clusters. In the following description, the K-Means algorithm will be used primarily as an example to illustrate the techniques according to this disclosure, as K-Means is a simple clustering algorithm well-suited for spherical clusters. However, it should be noted that this disclosure is not limited to the K-Means algorithm, and those skilled in the art can employ other suitable known clustering algorithms. The K-Means algorithm essentially includes the following steps:

[0025] - Initialization steps: Randomly select K data points from the data point set as the initial K group centers;

[0026] - Allocation steps: For each data point in the data point set, calculate its distance to each group center and assign it to the nearest group center, thereby forming K groups;

[0027] - Update steps: Recalculate the center of each cluster, i.e., the mean of all points within the cluster;

[0028] - Repeat the assignment and update steps until the group center no longer changes or the preset number of iterations is reached, at which point the algorithm converges.

[0029] Using the K-Means algorithm requires knowledge of the number of cell populations (i.e., the value of K in the initialization step). However, the number of cell populations is often unknown before clustering cells in an FSC-SSC scatter plot. To address this issue, this disclosure provides a method for determining cell populations based on the K-Means algorithm and the silhouette coefficient, such as... Figure 2 As shown.

[0030] See Figure 2In step S210, during the initialization step of the K-Means algorithm, K is set to the first value (e.g., 2) among a predetermined plurality of values ​​(e.g., 2, 3, ..., 10). It should be noted that predetermined values ​​such as 2, 3, ..., 10 are given as examples, and this disclosure is not limited thereto. Those skilled in the art can set these values ​​according to actual needs, such as based on experience or through experimentation.

[0031] In step S220, the subsequent steps of the K-Means algorithm are executed based on the set value of K to obtain the clustering result, i.e., K clusters.

[0032] In step S230, silhouette coefficients are calculated for the K clusters obtained after clustering. The silhouette coefficients S are defined by the following mathematical formulas (1) and (2):

[0033] S(i)=(b(i)-–a(i)) / (max{(a(i),b(i))}-(1)

[0034]

[0035] Where a(i) represents the average distance from the i-th point to all other points in the same cluster, b(i) represents the average distance from the i-th point to all points in its nearest cluster, and N represents the number of points in the K clusters. The silhouette coefficient is an index used to evaluate the quality of clustering, combining the cohesion and separation of the clusters. The silhouette coefficient value is in the interval [-1, 1], with values ​​closer to 1 indicating better clustering (i.e., better cohesion and separation), and values ​​closer to -1 indicating worse clustering.

[0036] Next, if K has not yet been set to the next of a predetermined set of values, i.e., "No" in step S240, the method returns to step S210, sets K to the next of the set of values ​​(e.g., 3), and performs steps S220 and S230 based on the newly set value of K to calculate the profile coefficient when K = 3.

[0037] Steps S210-S240 are repeated in this manner until K has been set to each of the predetermined values ​​and the corresponding silhouette coefficient has been calculated, i.e., "Yes" in step S240. At this point, the quality of the clustering results under different values ​​of K (i.e., different numbers of clusters) can be evaluated based on the silhouette coefficient. For example, Figure 3 The curve illustrating the relationship between the number of groups and the silhouette coefficient is shown schematically. Figure 3The horizontal axis of the dots on the curves shown corresponds to the number of clusters, which are 2, 3, 4, 5, and 6 respectively, while the vertical axis corresponds to the silhouette coefficient. It can be seen that the silhouette coefficient is largest when the number of clusters is 4, indicating that clustering cells into 4 clusters is optimal in the FSC-SSC scatter plot.

[0038] Therefore, as Figure 2 As shown in step S250, the cell population in the sample is determined based on the clustering results corresponding to the largest silhouette coefficient. Figure 3 In the example shown, the four cell populations can be determined based on the clustering results of the K-Means algorithm when there are four populations.

[0039] Next, based on the user's pre-set parameters regarding the target cell population of interest (e.g., as described below), Figure 6 In the pretreatment process, the target cell population can be identified from the four resulting cell populations, such as... Figure 2 Step S260 is shown. For example, Figure 1 (b) schematically illustrates the identified target cell population, which corresponds to the lymphocyte population.

[0040] For a selected target cell population, flow cytometry can generate histograms that include both negative and positive cell populations, such as... Figure 4 As shown. However, due to the influence of noise and other factors in flow cytometry, Figure 4 The negative and positive cell populations shown are often inaccurate, so according to conventional techniques, users need to manually set gates on the histogram to further define the negative and positive cell populations.

[0041] This disclosure provides a method for automatically identifying negative and positive cell populations. Specifically, with a population size of 2, the K-Means algorithm is used to cluster the selected target cell population. Clustering yields results such as… Figure 5 (a) schematically illustrates the negative and positive clusters. Next, with a cluster size of 2, the negative cluster is clustered into two clusters (not shown) using the K-Means algorithm, and one of these clusters is selected as the final negative cell population. Similarly, with a cluster size of 2, the positive cluster is clustered into two clusters (not shown) using the K-Means algorithm, and one of these clusters is selected as the final positive cell population. Figure 5 (b) schematically illustrates the final negative and positive cell populations. The two populations selected as the final negative and positive cell populations are farther apart than the two populations that were discarded.

[0042] Using the above methods, this disclosure can automatically identify negative and positive cell populations, eliminating the need for users to manually set gates on histograms.

[0043] Figure 6 A flowchart illustrating the overall method for determining the target cell population and negative and positive cell populations according to this disclosure is shown. Figure 6 As shown, in step S610, preprocessing is performed. Preprocessing may include, for example, specifying a single-stained sample by the user, setting the target cell population of interest, and setting microspheres. In step S620, clustering is performed based on the detection data of the sample loaded in the flow cytometer to automatically determine multiple cell populations in the sample. In step S630, one of the multiple cell populations is automatically determined as the target cell population according to the user's settings in the preprocessing. In step S640, negative and positive cell populations are automatically determined by performing clustering on the cells in the target cell population. Additionally, optionally, Figure 6 The method also includes step S650, whereby the user can manually modify the cell population if, based on their professional knowledge and experience, they deem the cell population automatically determined by the method unsatisfactory. Since defining the cell population by gating is a technique known to those skilled in the art, a detailed description of step S650 will be omitted herein.

[0044] The implementation of automatically identifying target cell populations and negative and positive cell populations for stained samples has been described above. The following describes an implementation according to this disclosure when loading unstained blank samples into a flow cytometer. The purpose of loading unstained blank samples is to detect autofluorescence in the samples.

[0045] Unstained samples may contain one or more types of autofluorescence, but the quantity of autofluorescence is difficult to predict, and the cell populations emitting various types of autofluorescence cannot be identified. To address this problem, this disclosure provides a method for automatically identifying autofluorescent cell populations.

[0046] After loading unstained samples, the detectors of multiple channels in a flow cytometer can provide multidimensional detection data of the samples. For example, an FSC-SSC scatter plot can display detection data in two dimensions: FSC channels (forward scattering channels) and SSC channels (side scattering channels). However, observing only two dimensions results in the loss of information in other dimensions. Therefore, this disclosure utilizes dimensionality reduction algorithms from machine learning to reduce the dimensionality of high-dimensional detection data. Examples of dimensionality reduction algorithms include t-SNE (t-Distributed Stochastic Neighbor Embedding), opt-SNE (optimized t-SNE), and PCA (Principal Component Analysis).

[0047] Figure 7 A schematic visualization of the dimensionality reduction effect when using the PCA algorithm is shown. Figure 7 (a) shows three populations with autofluorescence that can be identified in the FSC-SSC scatter plot. Figure 7 (b) shows the view obtained by mapping the dimensionality-reduced data onto a two-dimensional plane. Since all dimensions of information can be observed through projection in this two-dimensional plane, therefore... Figure 7 More groups can be identified in the figure shown in (b).

[0048] Based on the dimensionality-reduced data, the clustering method described above (using silhouette coefficient and K-Means algorithm) is performed to determine the cell populations in the unstained samples that correspond to various autofluorescence types.

[0049] Specifically, in the initialization step of the K-Means algorithm, the number K of autofluorescent cell populations is set to one of several predetermined values ​​(e.g., 1, 2, 3, ...), and subsequent steps are performed to obtain clustering results. Silhouette coefficients are calculated for the resulting clusters, and one or more cell populations emitting one or more autofluorescent patterns in the unstained sample are determined based on the clustering results (including one or more clusters) corresponding to the largest silhouette coefficient. This process is similar to that described above. Figure 2 The described process is similar. For example, assuming the largest silhouette coefficient is obtained when K=3, it indicates that the unstained sample contains three cell populations that produce three different types of autofluorescence. In this case, the details of these three cell populations can be determined based on the clustering results of the K-Means algorithm.

[0050] After identifying the various cell populations emitting different types of autofluorescence, the similarity of spectral characteristics between each autofluorescent cell population and each of the other autofluorescent cell populations can be calculated. Cosine similarity will be used as an example in the following description, but this disclosure is not limited to cosine similarity, and those skilled in the art can use other suitable similarity metrics, such as the Pearson correlation coefficient.

[0051] Cosine similarity is defined by the following mathematical formula (3):

[0052]

[0053] Where, x i and y i These are two vectors representing the spectral characteristics of the autofluorescence emitted by two cell populations. The value of cos(θ) reflects the similarity of the spectral characteristics of the autofluorescence emitted by the two cell populations. Specifically, if the value of cos(θ) is 1, it means that the autofluorescence produced by the two cell populations is exactly the same. If the value of cos(θ) is 0, it means that the autofluorescence produced by the two cell populations is completely different.

[0054] Since a conventional spillover matrix only includes elements corresponding to multiple predetermined fluorescence (fluorescent dyes) and does not include elements corresponding to autofluorescence, autofluorescence can be used to expand the spillover matrix in this disclosure. Specifically, when the calculated cos(θ) value is less than or equal to a predetermined threshold (e.g., 0.1), the autofluorescence produced by the two cell populations can be considered different from each other (both have unique spectral characteristics), and therefore the spillover matrix can be expanded using these two types of autofluorescence. When the calculated cos(θ) value is greater than the predetermined threshold, the autofluorescence produced by the two cell populations can be considered similar, and therefore the autofluorescence of one of the cell populations (with unique spectral characteristics) can be added to the spillover matrix.

[0055] Furthermore, the cosine similarity of spectral features between each autofluorescent cell population and an excited cell population (e.g., a positive cell population in a single-stained sample) can be calculated according to mathematical formula (3). In this case, when the calculated value of cos(θ) is less than or equal to a predetermined threshold, the autofluorescence emitted by the autofluorescent cell population can be considered to be different from the fluorescence already included in the overflow matrix, and therefore has unique spectral features, thus allowing it to be added to the overflow matrix. When the calculated value of cos(θ) is greater than the predetermined threshold, the autofluorescence is not added.

[0056] The following section details the extensions to the spillover matrix. The spillover matrix is ​​used to unmix multiple fluorescence signals detected by flow cytometry, thereby distinguishing them from one another. The spillover matrix is ​​typically defined by the following mathematical expression (4):

[0057]

[0058] Where, m ij The coefficients in the matrix are represented by . ...

[0059] When considering the autofluorescence of the sample, the number of fluorescence signals will be greater than the number of fluorescent dyes. In this disclosure, autofluorescence with unique spectral characteristics can be identified based on the calculated cosine similarity, matrix coefficients are determined for such autofluorescence, and these coefficients are added to the spillover matrix. The expanded spillover matrix can be represented by the following mathematical formula (5):

[0060]

[0061] Matrix M e The rightmost column This represents the newly added coefficient corresponding to a specific type of autofluorescence. It's easy to understand that when multiple types of autofluorescence are added, they can be represented in matrix M. e Add multiple columns to the right of it.

[0062] By using the expanded overflow matrix to unmix the detection data for actual samples, not only can the various fluorescence signals generated by the actual samples be distinguished from each other, but also each type of spontaneous fluorescence generated by the sample itself can be identified, thereby improving the detection accuracy.

[0063] Figure 8 A flowchart of a method for detecting autofluorescence according to this disclosure is shown. Figure 8As shown, in step S810, clustering is performed based on the detection data of the unstained sample loaded in the flow cytometer to automatically identify at least one population of autofluorescent cells emitting at least one type of autofluorescence. This step can be implemented based on the silhouette coefficient and K-Means algorithm described above. In step S820, the similarity of spectral features between autofluorescent cell populations, and / or the similarity of spectral features between autofluorescent cell populations and other cell populations, is calculated. In step S830, autofluorescence cells with unique spectral features are identified based on the calculated similarity, and these autofluorescence cells are added to the overflow matrix to expand the overflow matrix. In step S840, the expanded overflow matrix is ​​used to unmix the detection data for the actual sample to identify one or more types of autofluorescence present in the actual sample.

[0064] use Figure 8 The method shown can automatically identify autofluorescent cell populations in an unstained sample. Furthermore, the similarity of the autofluorescence emitted by these cell populations can be automatically compared, and based on the comparison results, the overflow matrix can be expanded using those distinctive autofluorescences. When unmixing is performed using the expanded overflow matrix, not only can each type of fluorescence emitted by the sample upon excitation be identified, but also the autofluorescence produced by that sample.

[0065] The technology according to this disclosure has been described above in conjunction with specific embodiments. This disclosure provides a method for automatically identifying target cell populations and negative and positive cell populations, eliminating the need for users to manually gating and selecting cell populations, thus improving processing efficiency. Furthermore, addressing the problem of users having difficulty identifying autofluorescence that may be present in a sample, this disclosure provides a method for automatically identifying autofluorescent cell populations in a sample. In addition, this disclosure proposes using autofluorescence with unique spectral characteristics to expand the overflow matrix, thereby enabling the detection of autofluorescence generated by the sample using the expanded overflow matrix, thereby improving the detection accuracy of flow cytometry.

[0066] It should be noted that the methods described in this disclosure are not necessarily to be executed in the order shown in the flowchart. Where technically feasible, some steps in the methods may be executed in different orders or in parallel.

[0067] The information processing methods for flow cytometers described above can be implemented by software, hardware, or a combination of both. Programs included in the software can be stored beforehand in a storage medium located internally or externally to the device. As an example, during execution, these programs are written to random access memory (RAM) and executed by a processor (e.g., a CPU) to implement the various methods and processes described herein. Therefore, this disclosure also includes an information processing system for flow cytometers, comprising a processing unit configured to perform the methods described above. Furthermore, flow cytometers including such an information processing system are also included within the scope of this disclosure.

[0068] In addition, this disclosure also includes computer program code and computer program products for implementing the methods described above, and computer-readable storage media on which the computer program code is recorded.

[0069] Figure 9 An example configuration block diagram of computer hardware is shown for performing the methods of this disclosure according to a program.

[0070] like Figure 9 As shown, in computer 900, central processing unit (CPU) 901, read-only memory (ROM) 902 and random access memory (RAM) 903 are connected to each other via bus 904.

[0071] The input / output interface 905 is further connected to the bus 904. The input / output interface 905 is connected to the following components: an input device 906 formed by a keyboard, mouse, microphone, etc.; an output device 907 formed by a display, speaker, etc.; a storage device 908 formed by a hard disk, non-volatile memory, etc.; a communication device 909 formed by a network interface card (such as a local area network (LAN) card, modem, etc.); and a driver 910 for driving a removable medium 911, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory.

[0072] In a computer with the above structure, the CPU 901 loads the program stored in the storage device 908 into the RAM 903 via the input / output interface 905 and the bus 904, and executes the program to perform the method described above.

[0073] The program to be executed by the computer (CPU 901) can be recorded on a removable medium 911, which is formed as a packaging medium, such as a magnetic disk (including a floppy disk), an optical disk (including a compact optical disk-read-only memory (CD-ROM)), a digital multifunction optical disk (DVD), etc.), a magneto-optical disk, or a semiconductor memory. Furthermore, the program to be executed by the computer (CPU 901) can also be provided via wired or wireless transmission media such as a local area network, the Internet, or digital satellite broadcasting.

[0074] When the removable medium 911 is installed in the drive 910, the program can be installed in the storage device 908 via the input / output interface 905. Alternatively, the program can be received by the communication device 909 via a wired or wireless transmission medium and installed in the storage device 908. Alternatively, the program can be pre-installed in the ROM 902 or the storage device 908.

[0075] The modules or systems described in this disclosure are for logical purposes only and do not strictly correspond to physical devices or entities. For example, the function of each module described in this disclosure may be implemented by multiple physical entities, or the function of multiple modules described in this disclosure may be implemented by a single physical entity. Furthermore, features, components, elements, steps, etc., described in one embodiment are not limited to that embodiment, but can also be applied to other embodiments, such as replacing specific features, components, elements, steps, etc., in other embodiments, or in combination with them.

[0076] The scope of this disclosure is not limited to the specific embodiments described herein. Those skilled in the art will understand that various modifications or variations can be made to the embodiments described herein, depending on design requirements and other factors, without departing from the principles of this disclosure. The scope of this disclosure is defined by the appended claims and their equivalents.

Claims

1. An information processing method for flow cytometers, comprising: A first clustering process is performed based on the detection data of the sample loaded in the flow cytometer to identify multiple cell populations in the sample; According to the predetermined settings, one of the multiple cell populations is identified as the target cell population; Negative and positive cell populations are determined by performing a second clustering process on the cells in the target cell population.

2. The information processing method according to claim 1, further comprising: The cell population obtained through clustering is manually modified by the user.

3. The information processing method according to claim 1, wherein The step of identifying multiple cell populations in the sample further includes: a. Set the number of the plurality of cell populations to the first of a predetermined plurality of values; b. Perform a clustering algorithm based on the set number of the multiple cell populations to obtain multiple populations as clustering results; c. Calculate the profile coefficients for the obtained multiple groups; d. Set the number of the plurality of cell populations to the next value among the plurality of values; e. Repeat steps b through d until the number of the plurality of cell populations has been set to each of the plurality of values; f. Determine multiple cell populations in the sample based on multiple populations corresponding to the largest silhouette coefficient.

4. The information processing method according to any one of claims 1-3, wherein The steps for identifying negative and positive cell populations also include: The cells in the target cell population are clustered into negative and positive groups; The cells in the negative group are clustered into a first group and a second group, and the cells in the positive group are clustered into a third group and a fourth group; The first group was selected as the negative cell population, and the fourth group was selected as the positive cell population. Among them, the distance between the first group and the fourth group is larger than the distance between the second group and the third group.

5. An information processing method for flow cytometers, comprising: Clustering is performed based on detection data of samples loaded in the flow cytometer, wherein the samples are not stained with fluorescent dyes and emit at least one type of autofluorescence. Based on the results of the clustering process, at least one population of autofluorescent cells emitting at least one autofluorescence is identified.

6. The information processing method according to claim 5, wherein The step of determining the at least one population of autofluorescent cells further includes: a. Set the number of the autofluorescent cell population to the first of a predetermined plurality of values; b. Perform a clustering algorithm based on the set number of autofluorescent cell populations to obtain at least one population as the clustering result; c. Calculate the profile coefficients for at least one obtained group; d. Set the number of the autofluorescent cell population to the next value among the plurality of values; e. Repeat steps b through d until the number of the autofluorescent cell population has been set to each of the plurality of values; f. Determine the at least one autofluorescent cell population based on at least one population corresponding to the largest silhouette coefficient.

7. The information processing method according to claim 5 or 6 further includes: For each autofluorescent cell population, calculate the first similarity of its spectral features with each other autofluorescent cell population. For each autofluorescent cell population, a second similarity of spectral characteristics between it and another cell population is calculated, wherein the other cell population is stained with a fluorescent dye and emits fluorescence corresponding to the fluorescent dye upon excitation; Based on the first similarity and the second similarity, autofluorescence with unique spectral characteristics is determined; The spontaneous fluorescence with unique spectral characteristics is used to expand the overflow matrix.

8. The information processing method according to claim 7, further comprising: An expanded overflow matrix is ​​used to unmix detection data for actual samples to identify one or more types of autofluorescence present in the actual samples.

9. The information processing method according to claim 5, further comprising: Dimensionality reduction processing is performed on the detection data; The clustering process is performed based on the dimensionality-reduced data to determine the at least one population of autofluorescent cells.

10. An information processing system for a flow cytometer, comprising a processing unit configured to: A first clustering process is performed based on the detection data of the sample loaded in the flow cytometer to identify multiple cell populations in the sample; According to the predetermined settings, one of the multiple cell populations is identified as the target cell population; Negative and positive cell populations are determined by performing a second clustering process on the cells in the target cell population.

11. The information processing system according to claim 10, wherein The processing unit is also configured to modify the cell population obtained through clustering based on user actions.

12. The information processing system according to claim 10, wherein The processing unit is also configured to perform the following operations to determine multiple cell populations in the sample: a. Set the number of the plurality of cell populations to the first of a predetermined plurality of values; b. Perform a clustering algorithm based on the set number of the multiple cell populations to obtain multiple populations as clustering results; c. Calculate the profile coefficients for the obtained multiple groups; d. Set the number of the plurality of cell populations to the next value among the plurality of values; e. Repeat steps b through d until the number of the plurality of cell populations has been set to each of the plurality of values; f. Determine multiple cell populations in the sample based on multiple populations corresponding to the largest silhouette coefficient.

13. The information processing system according to any one of claims 10-12, wherein, The processing unit is also configured to perform the following operations to determine negative and positive cell populations: The cells in the target cell population are clustered into negative and positive groups; The cells in the negative group are clustered into a first group and a second group, and the cells in the positive group are clustered into a third group and a fourth group; The first group was selected as the negative cell population, and the fourth group was selected as the positive cell population. Among them, the distance between the first group and the fourth group is larger than the distance between the second group and the third group.

14. An information processing system for a flow cytometer, comprising a processing unit configured to: performing a clustering process based on the detection data for the sample loaded in the flow cytometer, wherein, The sample was not stained with a fluorescent dye, and the sample emitted at least one type of autofluorescence; Based on the results of the clustering process, at least one population of autofluorescent cells that emits at least one autofluorescence is identified.

15. The information processing system according to claim 14, wherein The processing unit is also configured to perform the following operations to determine the at least one autofluorescent cell population: a. Set the number of the autofluorescent cell population to the first of a predetermined plurality of values; b. Perform a clustering algorithm based on the set number of autofluorescent cell populations to obtain at least one population as the clustering result; c. Calculate the profile coefficients for at least one obtained group; d. Set the number of the autofluorescent cell population to the next value among the plurality of values; e. Repeat steps b through d until the number of the autofluorescent cell population has been set to each of the plurality of values; f. Determine the at least one autofluorescent cell population based on at least one population corresponding to the largest silhouette coefficient.

16. The information processing system according to claim 14 or 15, wherein The processing unit is further configured to: For each autofluorescent cell population, calculate the first similarity of its spectral features with each other autofluorescent cell population. For each autofluorescent cell population, a second similarity of spectral characteristics between it and another cell population is calculated, wherein the other cell population is stained with a fluorescent dye and emits fluorescence corresponding to the fluorescent dye upon excitation; Based on the first similarity and the second similarity, autofluorescence with unique spectral characteristics is determined; The spontaneous fluorescence with unique spectral characteristics is used to expand the overflow matrix.

17. The information processing system of claim 16, wherein, The processing unit is further configured to use an expanded overflow matrix to unmix detection data for the actual sample in order to identify one or more autofluorescences present in the actual sample.

18. The information processing system according to claim 14, wherein The processing unit is further configured to: Dimensionality reduction processing is performed on the detection data; The clustering process is performed based on the dimensionality-reduced data to determine the at least one population of autofluorescent cells.

19. A flow cytometer comprising an information processing system according to any one of claims 10 to 18.

20. A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to perform the information processing method according to any one of claims 1 to 9.

Images