A dual cell recognition method, system, device and storage medium

Abstract

The application provides a double cell recognition method, comprising the following steps: step one) obtaining a data set to be subjected to double cell recognition, and then clustering and grouping the data set to be subjected to double cell recognition to obtain a plurality of cell populations; step two) determining mutually exclusive protein pairs according to the cell types of the data set to be subjected to double cell recognition, and then determining the expression negative cell population and the expression positive cell population corresponding to each protein in the mutually exclusive protein pairs by using the COSG method of cyclic iteration in the cell populations obtained in step one); step three) calculating the threshold value of the positive and negative expression of each protein in the mutually exclusive protein pairs according to the data in the expression negative cell population and the expression positive cell population; and step four) recognizing double cell data in the data set to be subjected to double cell recognition according to the threshold value. The double cell recognition method based on the COSG method of cyclic iteration is more stable and more effective in the recognition of positive and negative cell population modules.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of single-cell omics, and particularly relates to a double-cell identification method, system, device and storage medium. BACKGROUND

[0002] The development of single-cell omics technology has played a key role in promoting people's understanding of health and disease, and has provided important basis for discovering disease mechanisms and potential therapeutic targets. Quality control (QC) of single-cell omics is crucial to ensure the reliability and accuracy of data.

[0003] In single-cell sequencing experiments, if a reaction contains exactly two cells, double cells will be formed. Since double cells are not real cells, their existence will seriously interfere with the analysis of single-cell sequencing data. Therefore, it is necessary to develop some computational methods to identify double-cell data in single-cell sequencing data.

[0004] However, the existing double-cell identification methods usually have certain limitations in use. Single-cell transcriptome analysis is prone to lose information due to the sparsity of data, resulting in a blind area of quality control. ADT data of single-cell RNA and ADT joint sequencing is bimodal, and the missing rate is not so serious. Therefore, for multi-omic RNA+ADT joint sequencing data, it is necessary to further identify and supplement the identification of double cells from the protein level. SUMMARY

[0005] The application provides a double-cell identification method, system, device and storage medium to more stably and accurately identify double-cell data in single-cell sequencing data. The application adopts the following technical solutions:

[0006] A double-cell identification method, comprising the following steps:

[0007] Step 1) obtaining a data set to be subjected to double-cell identification, wherein the data set to be subjected to double-cell identification includes protein data generated by single-cell transcriptome sequencing and single-cell membrane protein sequencing; then clustering and grouping the data set to be subjected to double-cell identification to obtain a plurality of cell populations;

[0008] Step 2) determining mutually exclusive protein pairs according to the cell types of the data set to be subjected to double-cell identification, and then determining the expression-negative cell population and the expression-positive cell population corresponding to each protein in the mutually exclusive protein pairs by using the COSG method of cyclic iteration in the cell populations obtained in step 1);

[0009] Step 3) calculating the positive-negative expression threshold of each protein in the mutually exclusive protein pairs according to the data in the expression-negative cell population and the expression-positive cell population;

[0010] Step four) identifying doublet data in the data set to be identified according to the boundary threshold.

[0011] The mutually exclusive protein pairs are defined according to biological cell types, for example, the mutually exclusive protein pairs of B and T cells are CD19 and CD3.

[0012] Optionally, the data set to be identified is a single cell data set generated by RNA+ADT joint sequencing technology.

[0013] Optionally, the double cell identification method only identifies double cells on the protein data level of the single cell data set generated by RNA+ADT joint sequencing technology.

[0014] Optionally, the clustering and grouping in step one) is based on single cell transcriptome sequencing data or single cell membrane protein sequencing data.

[0015] Preferably, the data set to be identified is a single cell data set generated by CITE-seq technology.

[0016] Optionally, in step two), whether to perform background denoising on the protein data in each cell population can be selected, and the corresponding expression negative cell population and expression positive cell population of each protein in the mutually exclusive protein pair can be calculated using the denoised data.

[0017] Optionally, the background denoising is performed by scCDC method.

[0018] Optionally, the method for calculating the corresponding expression negative cell population and expression positive cell population of each protein in the mutually exclusive protein pair in step two) comprises:

[0019] For each protein in the mutually exclusive protein pair, the percentage of detected proteins and the average expression level in each cell population are calculated; the percentage of proteins and the average expression level are multiplied to obtain the penalty mean of the cell population;

[0020] The cell population with the highest penalty mean is taken as the initial seed of the cell population combination, and the cell populations are iteratively added to the cell population combination from high to low according to the penalty mean of the cell population; in each iteration, the COSG score of the combination of the added cell populations is calculated using the COSG method; when the COSG score does not increase after iteration or the frequency of the corresponding marker gene in the current cell population is less than the frequency threshold, the iteration is stopped; the cell populations in the added cell population combination are expression positive cell populations, and the remaining cell populations are expression negative cell populations;

[0021] The frequency threshold is 0.1-0.2.

[0022] Optionally, the COSG score is calculated by using the COSG algorithm formula as disclosed below,

[0023]

[0024] wherein g i represents each protein, G k represents the combination of cell populations, λ k represents the expression pattern of the hypothetical marker gene of G k cell population combination, λ t represents the expression pattern of the hypothetical marker gene of other cell populations except G k , and u represents a penalty factor.

[0025] Optionally, the step three) comprises calculating the positive expression and negative expression threshold of each protein in the mutually exclusive protein pair according to the data in the expression negative cell population and the expression positive cell population.

[0026] The positive expression and negative expression threshold of each protein in the mutually exclusive protein pair is calculated by using at least one of the ROC method and the Naive Bayes method.

[0027] The ROC method uses the labels of the positive and negative cell populations of the protein generated in the step three) and the expression value of the protein to generate a ROC curve and its object, and uses the Youden index to seek the optimal threshold of the ROC curve, i.e. the positive expression and negative expression threshold of the protein.

[0028] The Naive Bayes method takes the labels of the positive and negative cell populations of the protein generated in the step three) as the dependent variable of the model, and takes the expression value of the protein as the independent variable of the model to construct a Naive Bayes classification model. Then the model is used to predict the expression value of the protein to obtain the newly predicted labels of the positive and negative cells, and then calculate the minimum value of the protein expression of the new positive population and the maximum value of the protein expression of the negative population, and take the average of the minimum value and the maximum value as the positive expression and negative expression threshold of the protein.

[0029] Optionally, the method for identifying the double cell data in the data set to be subjected to double cell identification according to the threshold comprises:

[0030] The cell data whose expression level of each protein in the mutually exclusive protein pair is higher than the respective threshold is identified as the double cell data.

[0031] A double cell identification system comprises:

[0032] A clustering module: configured to obtain a data set to be subjected to double cell identification, and to cluster and group the data set to obtain a plurality of cell populations.

[0033] Yin and yang cell group recognition module of mutually exclusive protein pairs: determining the expression negative cell group and the expression positive cell group corresponding to each protein in the mutually exclusive protein pairs in the cell group obtained in step one) by using the COSG method in a loop iteration;

[0034] Threshold calculation module: for calculating the threshold of positive expression and negative expression of each protein in the mutually exclusive protein pairs according to the data in the expression negative cell group and the expression positive cell group;

[0035] Double cell data recognition module: identifying double cell data in the data set to be recognized according to the threshold.

[0036] An electronic device comprising a memory and a processor, the memory and the processor being connected;

[0037] The memory stores computer instructions, and the processor executes the computer instructions to perform the double cell recognition method described above.

[0038] A computer readable storage medium, the computer readable storage medium stores computer instructions, the computer instructions are used to make the computer execute the double cell recognition method described above.

[0039] Compared with the prior art, the technical scheme of the present application has the following beneficial effects:

[0040] The double cell recognition method based on the COSG method of the present application is more stable and effective in yin and yang cell group module recognition, while the MLtiplet CITE-seq method in the prior art uses mean to distinguish yin and yang cell groups, which is easily affected by background noise, clustering and expression level. In addition, each mutually exclusive protein in the MLtiplet CITE-seq method in the prior art needs to contain a matching gene to recognize double cell groups. The present method does not need matching proteins and genes, but only needs ADT data to recognize and analyze double cells. In addition, it also provides a background correction strategy and a variety of statistical discrimination methods, which can better adapt to different user groups.

[0041] The loop iteration COSG+ statistical strategy of the present method can effectively divide the bimodal characteristics of proteins in different data sets and proteins, while the MLtiplet CITE-seq method in the prior art can only perform well in part of the samples and proteins. Secondly, the present method also maintains good division performance in correcting data. In addition, the loop iteration COSG+ statistical method strategy of the present application has more stable threshold fluctuations compared with the MLtiplet CITE-seq in different resolution clustering dimensions, which avoids the influence caused by subjective selection of resolution. Attached Figure Description

[0042] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 Scatter plots showing the protein expression thresholds under uncorrected background conditions; where A is the scatter plot showing the expression thresholds of CD3-CD20 and CD19-CD3 mutually exclusive protein pairs in TP1 samples, the three plots correspond to different method strategies, with red representing the thresholds of their respective methods, the first two being the strategies of this method, and the last being the MLtiplet CITE-seq method strategy, the red dots ambigous representing TB double cells identified at the VDJ level; B is the scatter plot showing the expression thresholds of mutually exclusive protein pairs in TP2 samples; C is the scatter plot showing the expression thresholds of mutually exclusive protein pairs in TP3 samples; D is the scatter plot showing the expression thresholds of mutually exclusive protein pairs in TP3-rep samples; E is the scatter plot showing the expression thresholds of mutually exclusive protein pairs in pbmc5K samples; and F is the scatter plot showing the expression thresholds of mutually exclusive protein pairs in pbmc10K samples.

[0044] Figure 2 Scatter plots are used to divide protein expression thresholds under background correction conditions; where A is a scatter plot of expression thresholds for CD3-CD20 and CD19-CD3 mutually exclusive protein pairs in TP1 samples, the two plots correspond to different method strategies, red represents the division threshold of each method, and red ambigous represents TB double cells identified at the VDJ level; B is a scatter plot of expression thresholds for mutually exclusive protein pairs in TP2 samples; C is a scatter plot of expression thresholds for mutually exclusive protein pairs in TP3 samples; D is a scatter plot of expression thresholds for mutually exclusive protein pairs in TP3-rep samples; E is a scatter plot of expression thresholds for mutually exclusive protein pairs in pbmc5K samples; and F is a scatter plot of expression thresholds for mutually exclusive protein pairs in pbmc10K samples.

[0045] Figure 3 The table shows the threshold variations for different methods at different clustering resolutions; A represents the threshold fluctuations for TP1 samples using different methods; B represents the threshold fluctuations for TP2 samples using different methods; C represents the threshold fluctuations for TP3 samples using different methods; D represents the threshold fluctuations for TP3-rep samples using different methods; E represents the threshold fluctuations for pbmc5K samples using different methods; and F represents the threshold fluctuations for pbmc10K samples using different methods.

[0046] Figure 4 is a structural block diagram of a double cell recognition device according to an embodiment of the present application. DETAILED DESCRIPTION

[0047] Various exemplary embodiments of the present application will now be described in detail, which should be considered in a descriptive sense only and not for purposes of limitation to the present application. It should be understood that the terms used herein are for the purpose of describing particular embodiments and are not intended to limit the present application.

[0048] As used herein, the terms "comprises", "comprising", "includes", "including", "has", "having" or the like are open-ended terms that are intended to mean including, but not limited to. Unless otherwise noted, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. Although preferred methods are described herein, any method similar or equivalent to those described herein can be used in the practice or testing of the present application.

[0049] Example 1

[0050] The present application proposes a double cell recognition method, comprising the following steps:

[0051] Step one) obtaining a data set to be subjected to double cell recognition, and then clustering and grouping the data set to obtain a plurality of cell populations. The data set is a single cell data set generated by RNA+ADT joint sequencing technology, and in this embodiment, a single cell data set generated by CITE-seq technology can generally be used.

[0052] Step two) determining the expression negative cell population and the expression positive cell population corresponding to each protein in the mutually exclusive protein pair in the cell population obtained in step one) by using the COSG method with cyclic iteration; the mutually exclusive protein pair is defined according to biological cell types, for example, the mutually exclusive protein pair of B and T cells is CD19 and CD3.

[0053] Specifically, it comprises:

[0054] For each protein in the mutually exclusive protein pair, the percentage of detected proteins and the average expression level in each cell population are calculated; the percentage of proteins and the average expression level are multiplied to obtain the penalty mean of the cell population;

[0055] The cell population with the highest penalty mean is taken as the initial seed, and the cell populations are added in turn according to the penalty mean of the cell populations from high to low; in each iteration, the COSG score of the combination of the added cell populations is calculated using the disclosed COSG method. This disclosed COSG method is to calculate the cosine similarity between the expression pattern of a feature in a certain cell population and its assumed marker pattern, thereby obtaining the specific COSG score of the feature in a certain cell population, and the specific formula is as follows,

[0056]

[0057] wherein g i represents each protein, G k represents the combination of the cell populations, λ k represents the expression pattern of the marker gene of G k The combination of the cell populations, λ t represents the expression pattern of the marker gene of other cell populations except G k represents the penalty factor (in this embodiment, u takes the value of 1). Thus, the COSG specific score of the combination of the cell populations of each protein can be obtained; when the COSG score does not increase after iteration or the frequency of the corresponding marker gene in the current cell population is less than 0.2 (data generally has sparsity, and it is generally considered in the industry that a marker less than 0.1 or 0.2 basically has no great effect, and in order to identify more accurately, 0.2 is selected), the iteration is stopped; the cell population in the combination of the added cell populations is the positive expression cell population, and the remaining cell populations are the negative expression cell populations.

[0058] The COSG algorithm (Cosine Similarity-based Gene Identification) is a method for identifying marker genes based on cosine similarity. It is mainly used for single-cell data analysis to identify cell marker genes more accurately and quickly. The COSG algorithm uses cosine similarity to measure the relationship between gene expression vectors. Cosine similarity evaluates the similarity between two vectors in a vector space by calculating the cosine of the angle between them. Unlike traditional statistical testing methods, cosine similarity compares the direction of two vectors, not their absolute values, which means that even if two genes have the same expression pattern but different expression abundances, cosine similarity will consider them equivalent. The COSG algorithm is particularly suitable for cell type annotation in single-cell data analysis. In single-cell sequencing technology, cell type annotation is a key step, and traditional statistical methods may encounter challenges in identifying cell marker genes, as statistical tests tend to identify candidate genes with systematic differences between two groups rather than true cell markers. The COSG algorithm can more accurately identify cell marker genes by calculating the cosine similarity of gene expression vectors, thereby improving the accuracy of cell type annotation.

[0059] Step three) Calculate the positive expression and negative expression threshold of each protein in the mutually exclusive protein pair according to the data in the expression negative cell population and the expression positive cell population.

[0060] Step four) Identify double-cell data in the data set to be identified according to the threshold. In general, the data of cells whose expression levels of each protein in the mutually exclusive protein pair are higher than the respective threshold are identified as double-cell data.

[0061] Further, the clustering in step one) adopts RNA clustering or ADT clustering. The present application supports RNA and ADT single-omic clustering. When the number of ADT omics is small, RNA clustering can be used instead, and ADT clustering will be used by default.

[0062] Further, in step two), it can be selected whether to use the scCDC method to denoise the protein data in each cell population first, and the corresponding expression negative cell population and expression positive cell population of each protein in the mutually exclusive protein pair can be calculated using the denoised data.

[0063] Further, step three) includes calculating the positive expression and negative expression threshold of each protein in the mutually exclusive protein pair according to the data in the expression negative cell population and the expression positive cell population.

[0064] adopting at least one of ROC method and Naive Bayes method to calculate the threshold value of positive expression and negative expression of each protein in the mutually exclusive protein pair; the ROC method adopts the labels of the positive and negative groups of the protein generated in step three) and the expression value of the protein to generate a ROC curve and its object, and seeks the optimal threshold value of the ROC curve, i.e. the threshold value of positive expression and negative expression of the protein, by using the Youden index. The Naive Bayes method takes the labels of the positive and negative groups of the protein generated in step three) as the dependent variable of the model, and takes the expression value of the protein as the independent variable of the model to construct a Naive Bayes classification model. Then the model is used to predict the expression value of the protein to obtain the newly predicted labels of the positive and negative cells, and then the minimum value of the protein expression of the new positive group and the maximum value of the protein expression of the negative group are calculated, and the mean value of the minimum value and the maximum value is taken as the threshold value of positive expression and negative expression of the protein.

[0065] To better compare the pros and cons of the methods, we tested them on four self-tested data sets and two public CITE-seq data sets. The four self-tested data sets include TP1, TP2, TP3 and TP3-rep (originating from the same healthy donor, with three peripheral blood mononuclear cell (PBMCs) samples obtained at intervals of three days, namely TP1, TP2 and TP3. The third node sample (TP3) failed the QC initially, and then TP3-rep was obtained by re-measuring it), and the public data includes pbmc5K and pbmc10K, which are from the 10x Genomics website. Since the six CITE-seq data sets all contain matching VDJ sequencing data, in order to better reflect the performance difference, we selected the common mutually exclusive protein pairs of B and T cells, CD19-CD3 and CD20-CD3. It can be seen that the method of the application can effectively divide the bimodal characteristics of the proteins in different data sets and proteins, while the MLtiplet CITE-seq can only perform well in part of the samples and proteins. Figure 1 Secondly, the method of the application also maintains good division performance on the corrected data of the scCDC method. Figure 2

[0066] Moreover, the application further analyzes the threshold value fluctuation of different proteins of the COSG+ statistical method strategy with circulation iteration on different clustering dimensions of different resolutions, which is more stable than the MLtiplet CITE-seq Figure 3 , which avoids the influence caused by subjective selection of resolution.

[0067] ​Compared with the existing MLtiplet CITE-seq method, the method is more stable in identifying positive and negative cell population modules, and MLtiplet CITE-seq uses mean value to distinguish positive and negative cell population modules, which is easily affected by background noise, clustering and expression level. In addition, each mutually exclusive protein of MLtiplet CITE-seq needs to contain a matching gene to identify double cell population, and the method does not need matching protein and gene, and only needs ADT data to identify double cell analysis, and further provides a background correction strategy and a variety of statistical discrimination methods, which can better adapt to different user groups.

[0068] The application further provides a double cell identification system, comprising:

[0069] The clustering and grouping module is configured to obtain a data set to be identified by double cells, and perform clustering and grouping on the data set to obtain a plurality of cell populations.

[0070] The positive and negative cell population identification module of the mutually exclusive protein pair is configured to obtain the cell population obtained by the clustering and grouping module, and calculate the corresponding expression negative cell population and expression positive cell population of each protein in the mutually exclusive protein pair by using the COSG method of cyclic iteration.

[0071] The threshold calculation module is configured to calculate the positive expression and negative expression threshold of each protein in the mutually exclusive protein pair according to the data in the expression negative cell population and the expression positive cell population.

[0072] The double cell data identification module is configured to identify double cell data in the data set to be identified by double cells according to the threshold.

[0073] The further function description of each module and unit is the same as the above corresponding embodiment, and will not be repeated here.

[0074] The double cell identification device in the embodiment is in the form of a functional unit, and the unit refers to an ASIC (Application Specific Integrated Circuit, Application Specific Integrated Circuit) circuit, a processor and a memory executing one or more software or fixed programs, and / or other devices that can provide the above functions.

[0075] The embodiment of the application further provides a computer device with the double cell identification device.

[0076] Please refer to Figure 4Fig. 1 is a structural schematic diagram of a computer device according to an optional embodiment of the present application. The computer device includes one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. The various components communicate with each other using different buses, and can be mounted on a common main board or mounted in other manners as needed. The processor can process instructions executed within the computer device, including instructions stored in the memory or on the memory to display graphical information of a GUI on an external input / output device, such as a display device coupled to the interface.

[0077] In some optional embodiments, multiple processors and / or buses can be used together with multiple memories and multiple memory banks, if needed. Also, multiple computer devices can be connected, each providing part of the necessary operations (for example, as a server array, a set of blade servers, or a multi-processor system). Figure 4 The processor 10 is taken as an example in the figure.

[0078] The processor 10 can be a central processor, a network processor, or a combination thereof. The processor 10 can further include a hardware chip. The hardware chip can be an application specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device can be a complex programmable logic device, a field programmable logic device, a general array logic, or any combination thereof.

[0079] The memory 20 stores instructions executable by the at least one processor 10, so that the at least one processor 10 performs the double cell recognition method shown in the above embodiments.

[0080] The memory 20 can include a program storage area and a data storage area. The program storage area can store an operating system and application programs required by at least one function; the data storage area can store data created according to the use of the computer device, and the like. In addition, the memory 20 can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some optional embodiments, the memory 20 can include a memory remotely arranged with respect to the processor 10, and these remote memories can be connected to the computer device through a network. Examples of the network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0081] The memory 20 can include a volatile memory, such as a random access memory; the memory can also include a non-volatile memory, such as a flash memory, a hard disk, or a solid-state disk; and the memory 20 can further include a combination of the above kinds of memories.

[0082] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 can be connected through a bus or other means.

[0083] The input device 30 can receive input digital or character information, and generate key signal input with respect to a user setting and a function control of the computer device, such as a touch screen, a keypad, a mouse, a trackpad, a touchpad, a pointing stick, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 can include a display device, an auxiliary lighting device (e.g., an LED), a tactile feedback device (e.g., a vibration motor), etc. The display device includes, but is not limited to, a liquid crystal display, a light emitting diode, a display, and a plasma display. In some alternative embodiments, the display device can be a touch screen.

[0084] The computer device also includes a communication interface for communication of the computer device with other devices or communication networks.

[0085] The embodiments of the present application also provide a computer readable storage medium, and the above method according to the embodiments of the present application can be implemented in hardware, firmware, or recorded in a storage medium, or be implemented through computer code originally stored in a remote storage medium or a non-transitory machine readable storage medium and downloaded to a local storage medium, so that the method described herein can be processed by such software on a storage medium using a general purpose computer, a special purpose processor, or programmable or special purpose hardware.

[0086] The storage medium can be a magnetic disc, an optical disc, a read-only memory, a random access memory, a flash memory, a hard disk, or a solid state disk, etc. Further, the storage medium can also include a combination of the above-mentioned types of memories. It can be understood that the computer, the processor, the microprocessor controller, or the programmable hardware includes a storage component that can store or receive software or computer code, when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the method shown in the above embodiments is implemented.

[0087] Embodiments of the present application can also provide a computer program product including computer program instructions, which, when executed by a processor, cause the processor to perform the steps of the above method. Wherein the computer program product can be written in any combination of one or more programming languages to perform the operations of the embodiments of the present disclosure, including object-oriented programming languages such as Java, C++, etc., and conventional procedural programming languages such as "C" language 10 or similar programming languages. The program code can be executed entirely on a user computing device, partially on a user device, as a separate software package, partially on a user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0088] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the above embodiments of the present application are described in detail, those skilled in the art should understand that the technical solutions recorded in the above embodiments can be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A two-cell recognition method, characterized in that, Includes the following steps: Step 1) Obtain the dataset to be used for dual-cell identification, which includes protein data generated by single-cell transcriptome sequencing and single-cell membrane protein sequencing; then, cluster the dataset to be used for dual-cell identification to obtain multiple cell populations. Step 2) Determine mutually exclusive protein pairs based on the cell types of the dataset to be identified for dual-cell recognition. Then, in the cell population obtained in Step 1), use the iterative COSG method to determine the negative expression cell population and positive expression cell population corresponding to each protein in the mutually exclusive protein pair. The method for calculating the corresponding negative expression cell population and positive expression cell population for each protein in the mutually exclusive protein pair includes: For each protein in the mutually exclusive protein pair, calculate the percentage of the detected protein and the average expression level in each cell population; multiply the percentage of the protein and the average expression level to obtain the penalty mean of the cell population; The cell population with the highest penalty mean is used as the initial seed for cell population combinations, and cell populations are added to the cell population combinations in descending order of penalty mean. In each iteration, the COSG score of the added cell population combinations is calculated using the COSG method. If the COSG score does not increase after iteration or the frequency of the corresponding marker gene in the current cell population is less than the frequency threshold, the iteration stops. The cell populations in the added cell population combinations are positive expression cell populations, and the remaining cell populations are negative expression cell populations. The COSG score is calculated using the following publicly available COSG algorithm formula. in, Representing each protein, A combination representing a cell population, represent The expression patterns of presumed marker genes in cell population composition. Representative except The expression patterns of putative marker genes in other cell populations, where u represents a penalty factor; Step 3) Based on the data from the negative expression cell population and the positive expression cell population, calculate the boundary threshold for the positive and negative expression of each protein in the mutually exclusive protein pair; Step 4) Identify the two-cell data in the dataset to be identified based on the boundary threshold.

2. The dual-cell recognition method according to claim 1, characterized in that, In step one), clustering is performed based on single-cell transcriptome sequencing data or single-cell membrane protein sequencing data.

3. The dual-cell recognition method according to claim 2, characterized in that, The dataset to be used for dual-cell identification is a single-cell dataset generated by CITE-seq technology.

4. The dual-cell recognition method according to claim 1, characterized in that, In step two, the protein data in each cell population is first denoised, and the denoised data is used to calculate the corresponding negative expression cell population and positive expression cell population for each protein in the mutually exclusive protein pair.

5. The dual-cell recognition method according to claim 4, characterized in that, The background denoising was performed using the scCDC method.

6. The dual-cell recognition method according to claim 1, characterized in that, The frequency threshold is 0.1~0.

2.

7. The dual-cell recognition method according to claim 1, characterized in that, The method for calculating the boundary threshold between positive and negative expression of each protein in the mutually exclusive protein pair based on data from the negative expression cell population and the positive expression cell population in step three) includes: using at least one of ROC and Naive Bayes to calculate the boundary threshold between positive and negative expression of each protein in the mutually exclusive protein pair. The ROC method uses the labels of the positive and negative populations of the protein generated in step two) and the expression value of the protein to generate ROC curves and their objects. The Youden index is used to find the optimal boundary threshold of the ROC curve, that is, the boundary threshold between positive and negative expression of the protein. The Naive Bayes method uses the labels of the positive and negative populations of the protein generated in step two) as the dependent variable of the model and the protein expression value as the independent variable of the model to construct a Naive Bayes classification model. Then, the model is used to predict the protein expression value to obtain the labels of the newly predicted positive and negative cells. Then, the minimum value of protein expression in the new positive population and the maximum value of protein expression in the negative population are calculated, and the mean of the minimum and maximum values ​​is taken as the boundary threshold of protein expression between positive and negative.

8. The dual-cell recognition method according to claim 1, characterized in that, The method for identifying two-cell data in the dataset to be identified based on the boundary threshold includes: Cell data where the expression level of each protein in a mutually exclusive protein pair is above its respective threshold are identified as two-cell data.

9. A dual-cell recognition system, characterized in that, The dual-cell recognition system is used to perform the dual-cell recognition method according to any one of claims 1 to 8, including: Clustering module: used to acquire the dataset to be identified by two cells, which includes protein data generated by single-cell transcriptome sequencing and single-cell membrane protein sequencing; then the dataset to be identified by two cells is clustered to obtain multiple cell populations; The module for identifying the positive and negative cell populations of mutually exclusive protein pairs is used to determine the mutually exclusive protein pairs based on the cell types of the dataset to be identified in dual cell recognition, and then to determine the expression negative cell population and expression positive cell population corresponding to each protein in the mutually exclusive protein pair using the iterative COSG method in the cell population obtained in step one). Boundary threshold calculation module: used to calculate the boundary threshold between positive and negative expression of each protein in the mutually exclusive protein pair based on data from the negative expression cell population and the positive expression cell population; Dual-cell data recognition module: Identifies dual-cell data in the dataset to be recognized based on the boundary threshold.

10. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory and the processor are connected. The memory stores computer instructions, and the processor executes the computer instructions to perform the dual-cell recognition method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform the dual-cell recognition method according to any one of claims 1 to 8.

Images