An intelligent anomaly detection system for single-cell expression profiles

Abstract

The present invention relates to the field of abnormal detection of cell expression profiles, and specifically to an intelligent abnormal detection system for single cell expression profiles. The system first obtains the gene expression profiles of preset regulatory genes and preset kinase genes of each cell and the distance between two cells, analyzes the change trend of the gene expression level of the preset regulatory genes of the target cells at each local moment, obtains the initial abnormal characteristic value, analyzes the correlation of the gene expression level between the preset regulatory genes and the preset kinase genes of the target cells, and corrects the initial abnormal characteristic value of the target cell through the obtained correlation parameters, obtains the real abnormal characteristic value, analyzes the difference and distance of the real abnormal characteristic value between the target cell and other cells, and performs abnormal detection on the gene expression profile of the target cell through the obtained gene expression abnormality index. The present invention can improve the accuracy of abnormal detection of the gene expression profile of a single cell.

Description

Technical Field

[0001] The present invention relates to the field of abnormality detection of cell expression profiles, and in particular to an intelligent abnormality detection system for single cell expression profiles. Background Art

[0002] Single-cell expression profiling is a technology that studies gene expression at the single-cell level. It detects abnormalities in the gene expression profiles of single cells obtained through single-cell sequencing technology. It can analyze and identify variations or differences in gene expression at the single-cell level, helping relevant personnel to discover early signs of disease at the molecular level. It is of great significance in the field of biological cell gene research.

[0003] In the related art, an anomaly detection model is usually constructed based on the data in the gene expression profile of a single cell, and anomaly detection of the gene expression profile of a single cell is achieved through the anomaly detection model. However, since the changing relationship between different data in the gene expression profile of a single cell is relatively complex, the detection accuracy of the constructed traditional anomaly detection model is low, thereby reducing the accuracy of anomaly detection of the gene expression profile of a single cell. Summary of the invention

[0004] In order to solve the technical problem that the detection accuracy of the traditional anomaly detection model is low, thereby reducing the accuracy of anomaly detection of single-cell gene expression profiles, the purpose of the present invention is to provide an intelligent anomaly detection system for single-cell expression profiles. The technical solutions adopted are as follows:

[0005] The present invention proposes an intelligent anomaly detection system for single-cell expression profiles, the system comprising:

[0006] A data acquisition module, used to acquire gene expression profiles of different cells in the tissue to be tested and the distances between any two cells, wherein the gene expression profiles include the gene expression levels of preset regulatory genes and preset kinase genes at each moment in a preset time period;

[0007] The abnormality analysis module is used to take any cell as a target cell, and obtain the change rate value at each moment according to the change of the gene expression level of the preset regulatory gene of the target cell at each moment; and obtain the initial abnormal characteristic value of the target cell according to the difference of the change rate value of the preset regulatory gene of the target cell at each moment in the preset window at each moment;

[0008] A correction module is used to analyze the correlation between the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell at each moment to obtain the correlation parameter of the target cell; based on the correlation parameter of the target cell, the initial abnormal characteristic value is corrected to obtain the real abnormal characteristic value of the target cell; according to the distance between the target cell and other cells except the target cell, the real abnormal characteristic value of the target cell, and the real abnormal characteristic value of other cells except the target cell, the gene expression abnormality index of the target cell is obtained;

[0009] The abnormality detection module is used to perform abnormality detection on the gene expression profile of the target cell based on the gene expression abnormality index of the target cell.

[0010] Further, obtaining the change rate value at each moment according to the change of the gene expression level of the preset regulatory gene of the target cell at each moment comprises:

[0011] Using each moment in the gene expression profile of the target cell and the two-dimensional data points consisting of the gene expression level of the preset regulatory gene at each moment as the sample points of the preset regulatory gene of the target cell at each moment;

[0012] According to the change in relative position between sample points of the preset regulatory gene of the target cell at each moment, the change rate value at each moment is obtained.

[0013] Further, obtaining the initial abnormal characteristic value of the target cell according to the difference in the change rate value of the preset regulatory gene of the target cell at each moment in the preset window at each moment includes:

[0014] Taking any moment as a target moment, obtaining the abnormal factor at the target moment according to the difference in the change rate values ​​of the preset regulatory gene of the target cell at adjacent moments in the preset window of the target moment;

[0015] Based on the abnormal factors at each moment, screening out the abnormal moments of the preset regulatory genes of the target cells from all moments;

[0016] The abnormal factors of the target cell at all the abnormal moments are integrated to obtain the initial abnormal characteristic value of the target cell.

[0017] Further, obtaining the abnormal factor at the target moment according to the difference in the change rate values ​​of the preset regulatory gene of the target cell at adjacent moments in the preset window at the target moment includes:

[0018] Any two adjacent moments in the preset window of the target moment are taken as a moment group of the target moment;

[0019] According to the difference of the change rate values ​​at two moments in each moment group, obtaining the rate change difference value of each moment group;

[0020] Analyze the discreteness of the rate change difference values ​​of all time groups of the target time to obtain the rate change disorder degree of the target time;

[0021] The rate change disorder at the target moment is normalized to obtain the abnormal factor at the target moment.

[0022] Furthermore, the abnormal moments of the preset regulatory genes of the target cells are screened out from all moments based on the abnormal factors at each moment, including:

[0023] The moment when the abnormal factor is greater than the preset first threshold is regarded as the abnormal moment of the preset regulatory gene of the target cell.

[0024] Furthermore, the analysis of the correlation between the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cells at each time to obtain the correlation parameters of the target cells includes:

[0025] The Pearson correlation coefficient of the gene expression levels between the preset regulatory gene and the preset kinase gene of the target cell at each moment is used as the correlation parameter of the target cell.

[0026] Furthermore, the correcting the initial abnormal characteristic value based on the correlation parameter of the target cell to obtain the real abnormal characteristic value of the target cell includes:

[0027] Normalizing the difference between the correlation parameter of the target cell and a constant 1 to obtain a non-positive correlation degree of the target cell;

[0028] The non-positive correlation degree of the target cell is used to perform weighted adjustment on the initial abnormal characteristic value to obtain the true abnormal characteristic value of the target cell.

[0029] Furthermore, obtaining the abnormal gene expression index of the target cell according to the distance between the target cell and other cells except the target cell, the true abnormal characteristic value of the target cell, and the true abnormal characteristic value of other cells except the target cell comprises:

[0030] According to the distance between the target cell and other cells except the target cell, the true abnormal characteristic values ​​of all other cells except the target cell are adjusted to obtain a reference abnormal characteristic value of the target cell;

[0031] The difference between the true abnormal characteristic value of the target cell and the reference abnormal characteristic value is normalized to obtain the abnormal gene expression index of the target cell.

[0032] Further, adjusting the true abnormal characteristic values ​​of all other cells except the target cell according to the distance between the target cell and other cells except the target cell to obtain the reference abnormal characteristic value of the target cell includes:

[0033] Any other cell except the target cell is used as a reference cell, and the distance between the reference cell and the target cell is negatively correlated and normalized to obtain a reference value parameter of the reference cell;

[0034] The reference value parameters of each other cell except the target cell are used to perform weighted adjustment on the true abnormal characteristic value of each other cell to obtain the reference abnormal characteristic value of the target cell, wherein the cumulative value of the reference value parameters of each other cell except the target cell is equal to 1.

[0035] Furthermore, the abnormality detection of the gene expression profile of the target cell based on the abnormal gene expression index of the target cell includes:

[0036] If the gene expression abnormality index of the target cell is greater than a preset second threshold, it is considered that the gene expression profile of the target cell is abnormal.

[0037] The present invention has the following beneficial effects:

[0038] The present invention takes into account that the changing relationship between different data in the gene expression spectrum of a single cell is relatively complex, resulting in low detection accuracy of the constructed traditional anomaly detection model, thereby reducing the accuracy of anomaly detection of the gene expression spectrum of a single cell. Therefore, the present invention first obtains the gene expression spectrum of different cells in the tissue to be tested and the distance between each two cells. Considering that in the normal cell division process, the gene expression level of its preset regulatory gene gradually increases, the change rate value can be used to first reflect the change trend of the gene expression level of the preset regulatory gene of the target cell at each moment. Considering that under normal circumstances, the growth rate of the local gene expression level of the preset regulatory gene of the target cell gradually increases, and the growth rate changes relatively evenly, when the expression of the preset regulatory gene is abnormal, the growth rate changes. The uniformity is poor, so the initial abnormal characteristic value obtained can preliminarily reflect the possibility of abnormal gene expression in the target cell. At the same time, under normal circumstances, the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell are positively correlated. Therefore, the initial abnormal characteristic value of the target cell can be corrected by the obtained correlation parameter to obtain a more accurate real abnormal characteristic value. Since the target cell and other cells are cells of the same type belonging to the same tissue, the gene expression of each cell is similar under normal circumstances. Therefore, other cells can be used as a reference, and the closer the other cells are to the target cell, the greater the reference value. Then, the abnormal gene expression index can accurately reflect the possibility of abnormal gene expression in the target cell, thereby improving the accuracy of subsequent abnormal detection of gene expression profiles of single cells. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 A block diagram of an intelligent anomaly detection system for single-cell expression profiles provided by an embodiment of the present invention;

[0041] Figure 2 A flow chart of a method for obtaining initial abnormal characteristic values ​​of target cells provided by one embodiment of the present invention;

[0042] Figure 3 This is a flow chart of a method for obtaining abnormal gene expression indicators of target cells provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, characteristics and effects of an intelligent abnormality detection system for single-cell expression profile proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0044] Unless defined otherwise, 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.

[0045] The following is a detailed description of a specific scheme of an intelligent anomaly detection system for single-cell expression profiles provided by the present invention in conjunction with the accompanying drawings.

[0046] See also Figure 1 , which shows a block diagram of an intelligent anomaly detection system for single-cell expression profile provided by an embodiment of the present invention, the system includes: a data acquisition module 101, an anomaly analysis module 102, a correction module 103, and an anomaly detection module 104.

[0047] The data acquisition module 101 is used to acquire the gene expression profiles of different cells in the tissue to be tested and the distances between any two cells. The gene expression profiles include the gene expression levels of the preset regulatory genes and the preset kinase genes at each moment.

[0048] Single-cell expression profiling is a technology that studies gene expression at the level of a single cell. Anomaly detection is performed on the gene expression profile of a single cell obtained by single-cell sequencing technology, which can analyze and identify variations or differences in gene expression at the level of a single cell, helping relevant personnel to discover early signs of disease at the molecular level, etc. It is of great significance in the field of biological cell gene research. In related technologies, anomaly detection models are usually constructed based on data in the gene expression profile of a single cell, and anomaly detection of the gene expression profile of a single cell is achieved through the anomaly detection model. However, due to the complex relationship between the changes in different data in the gene expression profile of a single cell, the detection accuracy of the constructed traditional anomaly detection model is low, thereby reducing the accuracy of anomaly detection of the gene expression profile of a single cell. Therefore, an embodiment of the present invention proposes an intelligent anomaly detection system for single-cell expression profile to solve this problem.

[0049] Since the regulatory genes in cells play an important role in the process of cell division, and their function is to promote the transition of cell division from one stage to another, the gene expression of the regulatory genes in the cells is analyzed in the subsequent steps of the embodiment of the present invention. The embodiment of the present invention first uses ImageJ software to obtain the distance between each cell in the tissue to be tested, such as human oral tissue. Subsequently, the distance can be used as a measure of the reference value of the cell to achieve more accurate abnormality detection of the single cell expression spectrum.

[0050] At the same time, a cell separation technique, such as an enzymatic cell dissociation method, is used to separate multiple cells of the tissue to be tested, thereby obtaining multiple single cells. The cell separation technique is a technical field well known to those skilled in the art, which will not be elaborated or limited herein. For each cell, the preset regulatory gene of each cell is processed using the existing single-cell sequencing technology to obtain the gene expression level of the preset regulatory gene at each moment in a preset time period, wherein in the embodiment of the present invention, the preset regulatory gene can be set to, for example, CCNB1 gene or Hox gene, etc., which will not be limited herein. At the same time, considering that the complex formed by the product of regulatory gene expression and the product of kinase gene expression in the cell plays an important role in the process of cell division, and the degree of demand for the two for cell division is relatively close, Therefore, the embodiment of the present invention also needs to use single-cell sequencing technology to process the preset kinase gene of each cell, and obtain the gene expression level of the preset kinase gene at each moment in the preset time period, so as to facilitate subsequent auxiliary analysis. In the embodiment of the present invention, the preset regulatory gene can be set to, for example, the CDK1 gene or the MAPK gene, etc., which is not limited here. The value range of the preset time period is generally [6, 12], and the unit is hours. The value range of the time interval for obtaining the gene expression level is generally [10, 30], and the unit is seconds. In one embodiment of the present invention, the preset time period is set to 8 hours, and the time interval is set to 20 seconds. The specific values ​​of the preset time period and the time interval can also be set by the implementer according to the specific implementation scenario, which is not limited here.

[0051] After obtaining the gene expression levels of the preset regulatory genes and the preset kinase genes in each cell at each moment, the gene expression levels of the preset regulatory genes in each cell at each moment and the gene expression levels of the preset kinase genes at each moment can be mapped to a two-dimensional coordinate system with time as the horizontal axis and the gene expression level as the vertical axis, thereby obtaining the gene expression profile of each cell, which records the gene expression status of the preset regulatory genes and the preset kinase genes of the cell.

[0052] It should be noted that gene expression level is an existing concept. The gene expression level of a preset regulatory gene or a preset kinase gene in a cell at each moment can be measured by the mRNA production of the corresponding gene at each moment. The gene expression level of the preset regulatory gene or the preset kinase gene in each cell at each moment can be obtained by existing methods such as FPKM, TPM, etc., or by the protein production of the corresponding gene at each moment. The gene expression level of the preset regulatory gene or the preset kinase gene in each cell at each moment can be obtained by existing methods such as Western blotting, ELISA, and mass spectrometry analysis, which will not be elaborated or limited here.

[0053] The abnormal analysis module 102 is used to take any cell as a target cell, and obtain the change rate value at each moment according to the change of the gene expression level of the preset regulatory gene of the target cell at each moment; and obtain the initial abnormal characteristic value of the target cell according to the difference of the change rate value of the preset regulatory gene of the target cell at each moment in the preset window at each moment.

[0054] In the subsequent steps of the embodiment of the present invention, only a single cell is analyzed, so any cell can be first used as a target cell. Since the function of the preset regulatory gene is to promote cell division, in the early stage of target cell division, the gene expression level of the preset regulatory gene in the target cell is relatively low. As the division proceeds, the target cell's demand for the product expressed by the preset regulatory gene increases. Therefore, in the normal cell division process, the gene expression level of the preset regulatory gene in the target cell gradually increases, and the closer to the late stage of cell division, the faster the growth rate of the gene expression level of the preset regulatory gene. Therefore, the changes in the gene expression level of the preset regulatory gene of the target cell at each moment can be first analyzed, and the obtained change rate value reflects the degree of change trend of the gene expression level of the preset regulatory gene of the target cell at each moment, providing a data basis for the subsequent analysis of the initial abnormal characteristic value of the target cell.

[0055] Preferably, in one embodiment of the present invention, the method for obtaining the change rate value at each moment specifically includes:

[0056] The two-dimensional data points composed of each moment in the gene expression spectrum of the target cell and the gene expression level of the preset regulatory gene at each moment are used as the sample points of the preset regulatory gene of the target cell at each moment; according to the change in the relative position between the sample points of the preset regulatory gene of the target cell at each moment, the change rate value at each moment is obtained.

[0057] In one embodiment of the present invention, the sample points at each moment are connected to the sample points at two adjacent moments to obtain two line segments, and the average of the absolute values ​​of the slopes of the two line segments is used as the change rate value at that moment. For the first moment or the last moment, since there is only one adjacent moment, the absolute value of the slope of the line connecting it and the sample points at the adjacent moments can be used as the change rate value at that moment.

[0058] In other embodiments of the present invention, curve fitting can be performed on the sample points at each moment to obtain a fitting curve, and the absolute value of the slope of the fitting curve at each moment is used as the change rate value at each moment. Curve fitting can be implemented using existing methods such as polynomial fitting or least squares method, which will not be elaborated or limited here.

[0059] Under normal circumstances, as cell division proceeds, the growth rate of the gene expression level of the preset regulatory gene in the target cell gradually increases over time, and its growth rate changes relatively evenly. When the expression of the preset regulatory gene in the target cell is abnormal, its gene expression level will fluctuate, resulting in poor uniformity in the growth rate of the gene expression level of the preset regulatory gene over time. Specifically, as time goes by, the change in the growth rate of the gene expression level of the preset regulatory gene has the characteristics of sometimes fast and sometimes slow. Therefore, the difference in the change rate values ​​of the preset regulatory gene of the target cell at each moment in the preset window at each moment can be analyzed, and the initial abnormal characteristic value obtained can be used to preliminarily reflect the degree of abnormal expression of the preset regulatory gene in the gene expression spectrum of the target cell. The initial abnormal characteristic value can be further adjusted later to improve the accuracy of abnormal detection of the gene expression spectrum of the target cell. The length of the preset window is set to 20, that is, the preset window at each moment includes the moment and the 19 other moments closest to the moment. The length of the preset window can also be set by the implementer according to the specific implementation scenario, and is not limited here.

[0060] Preferably, in one embodiment of the present invention, the method for obtaining the initial abnormal characteristic value of the target cell specifically includes:

[0061] See also Figure 2 , which shows a flow chart of a method for obtaining initial abnormal characteristic values ​​of target cells provided by one embodiment of the present invention.

[0062] Step S201: taking any moment as a target moment, and obtaining the abnormal factor at the target moment according to the difference in the change rate values ​​of the preset regulatory genes of the target cells at adjacent moments in the preset window of the target moment.

[0063] From the above analysis, it can be seen that when the expression of the preset regulatory gene in the target cell is abnormal, the uniformity of the growth rate of the gene expression level of the preset regulatory gene over time will be poor. Therefore, any moment can be first taken as the target moment, and the difference in the change rate values ​​of the preset regulatory gene of the target cell at adjacent moments in the preset window of the target moment is analyzed. The greater the difference in the change rate values ​​between adjacent moments in the preset window, the poorer the uniformity of the growth rate of the gene expression level of the preset regulatory gene of the target cell at each local moment of the target moment, and then the degree of abnormal expression of the preset regulatory gene at the target moment can be reflected by the obtained abnormal factor. The larger the abnormal factor, the greater the possibility of abnormal expression of the preset regulatory gene of the target cell at the target moment. Subsequently, the abnormal moment can be screened out from all moments by the abnormal factor, thereby realizing accurate analysis of the initial abnormal characteristic value of the target cell.

[0064] Preferably, in one embodiment of the present invention, the method for obtaining the abnormal factor at the target moment specifically includes:

[0065] Any two adjacent moments in the preset window of the target moment are taken as a moment group of the target moment. According to the difference in the change rate values ​​between the two moments in each moment group, the rate change difference value of each moment group is obtained. The rate change difference value reflects the difference in the change rate values ​​between the two moments in each moment group.

[0066] In the embodiment of the present invention, the absolute value of the difference between the change rate values ​​of two moments in each moment group or the square value of the difference can be used as the rate change difference value of each moment group, which is not limited here.

[0067] The greater the degree of dispersion of the rate change difference values ​​of each moment group, the less uniform the growth rate of the gene expression level of the preset regulatory gene of the target cell at each local moment of the target moment, which further indicates that the possibility of abnormal expression of the preset regulatory gene of the target cell at the target moment is greater. Therefore, the degree of dispersion of the rate change difference values ​​of all moment groups at the target moment is analyzed to obtain the degree of disorder of the rate change at the target moment.

[0068] In an embodiment of the present invention, statistics such as the variance, standard deviation or range of the rate change difference values ​​of all moment groups at the target moment can be used as the rate change disorder of the target moment to analyze the degree of discreteness of the rate change difference values ​​of all moment groups at the target moment, which is not limited here.

[0069] The rate change disorder at the target moment is normalized, and the acquired abnormal factor of the target moment is limited to the range of [0,1]. The larger the abnormal factor of the target moment, the greater the possibility that the preset regulatory gene of the target cell will be abnormally expressed at the target moment. Subsequently, the abnormal moment can be screened out from all moments through the abnormal factor, and provide a data basis for the calculation of the initial abnormal characteristic value of the target cell.

[0070] In one embodiment of the present invention, the normalization processing can be specifically, for example, maximum and minimum value normalization processing, and the normalization in subsequent steps can all adopt maximum and minimum value normalization processing. In other embodiments of the present invention, other normalization methods can be selected according to the specific range of numerical values, which will not be described in detail.

[0071] By using the same method as above, the abnormal factors of the target cells at each moment can be obtained.

[0072] Step S202: Based on the abnormal factors at each moment, the abnormal moments of the preset regulatory genes of the target cells are screened out from all moments.

[0073] The larger the abnormal factor, the greater the possibility that the preset regulatory gene of the target cell will be abnormally expressed at the target moment. Therefore, based on the abnormal factor at each moment, the abnormal moments of the preset regulatory gene of the target cell can be screened out from all moments. Subsequently, based on the abnormal factors at each abnormal moment, the initial abnormal characteristic value of the target cell can be analyzed and calculated.

[0074] Preferably, in one embodiment of the present invention, the moment when the abnormal factor is greater than a preset first threshold can be used as the abnormal moment of the preset regulatory gene of the target cell. The preset first threshold is set to 0.5. The specific value of the preset first threshold can also be set by the implementer according to the specific implementation scenario, which is not limited here.

[0075] Step S203: integrating the abnormal factors of the target cell at all abnormal moments to obtain the initial abnormal characteristic value of the target cell.

[0076] The more abnormal moments of the preset regulatory genes of the target cells and the larger the abnormal factor at each abnormal moment, the greater the degree of abnormality in the expression of the preset regulatory genes of the target cells. Therefore, the abnormal factors of all abnormal moments of the target cells can be integrated, and the initial abnormal characteristic values ​​obtained can preliminarily reflect the degree of abnormal expression of the preset regulatory genes in the gene expression spectrum of the target cells.

[0077] In the embodiment of the present invention, the abnormal factors of the target cells at all abnormal moments may be added together to achieve integration of the abnormal factors of the target cells at all abnormal moments.

[0078] The initial abnormal characteristic value of each cell can be obtained by the same method as above.

[0079] The correction module 103 is used to analyze the correlation between the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cells at each moment to obtain the correlation parameters of the target cells; based on the correlation parameters of the target cells, the initial abnormal characteristic values ​​are corrected to obtain the real abnormal characteristic values ​​of the target cells; according to the distance between the target cells and other cells except the target cells, the real abnormal characteristic values ​​of the target cells, and the real abnormal characteristic values ​​of other cells except the target cells, the gene expression abnormality index of the target cells is obtained.

[0080] Since the complex formed by the product expressed by the preset regulatory gene and the product expressed by the preset kinase gene in the cell plays an important role in the cell division process, the demand for the preset regulatory gene and the preset kinase gene is similar during the cell division process. Under normal circumstances, the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell have a certain positive correlation. Therefore, the correlation between the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell at each moment can be analyzed, and the correlation parameters obtained reflect the correlation between the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell. Subsequently, the initial abnormal characteristic value of the target cell can be corrected based on the correlation parameters, thereby improving the accuracy of the analysis of the abnormal expression of the preset regulatory gene of the target cell, thereby improving the accuracy of the abnormal detection of the gene expression spectrum of the target cell.

[0081] Preferably, in one embodiment of the present invention, based on a correlation analysis algorithm, the correlation between the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cells at each moment can be analyzed to obtain a correlation parameter of the target cells. The value range of the correlation parameter is [-1,1]. The closer the correlation parameter is to the constant 1, the stronger the positive correlation between the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cells. The closer the correlation parameter is to the constant -1, the weaker the positive correlation between the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cells. This further indicates that the expressions of the preset regulatory genes and the preset kinase genes in the target cells are abnormal, that is, the gene expression profile of the target cells is abnormal. Subsequently, the initial abnormal characteristic value can be further corrected based on the correlation parameter of the target cells to improve the accuracy of abnormal detection of the gene expression profile of the target cells.

[0082] In an embodiment of the present invention, existing correlation analysis algorithms such as the Pearson correlation coefficient or the Spearman rank correlation coefficient can be used to implement correlation analysis of the gene expression levels at each moment between the preset regulatory gene and the preset kinase gene of the target cell, which is not limited here.

[0083] As can be seen from the above analysis, under normal circumstances, the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cells have a certain positive correlation. The weaker the positive correlation between the two, the greater the possibility that the gene expressions of the preset regulatory genes and the preset kinase genes in the target cells are abnormal. Therefore, the correlation parameters of the target cells can be used to correct the initial abnormal characteristic values. The actual abnormal characteristic values ​​obtained can more accurately reflect the possibility or degree of abnormality in the gene expression spectrum of the target cells, providing a data basis for the subsequent calculation and analysis of abnormal gene expression indicators of the target cells.

[0084] Preferably, in one embodiment of the present invention, the method for obtaining the true abnormal characteristic value of the target cell specifically includes:

[0085] From the above analysis, it can be seen that the closer the correlation parameter of the target cell is to the constant -1 and the farther it is from the constant 1, the weaker the positive correlation between the gene expression levels of the preset regulatory genes and the preset kinase genes of the target cell is, which further indicates that the expression of the preset regulatory genes and the preset kinase genes in the target cell is more likely to be abnormal, and the gene expression spectrum of the target cell is more likely to be abnormal. Therefore, the difference between the correlation parameter of the target cell and the constant 1 can be normalized to obtain the degree of non-positive correlation of the target cell.

[0086] In one embodiment of the present invention, the correlation parameter of the target cell can be subtracted from the constant 1 and the absolute value can be taken to analyze the difference between the two.

[0087] The greater the degree of non-positive correlation of the target cells, the weaker the positive correlation between the gene expression profiles of the preset regulatory genes and the preset kinase genes of the target cells, which further indicates that the gene expression profiles of the target cells are more likely to be abnormal. In order to improve the accuracy of abnormal detection of target cell gene expression profiles, the degree of non-positive correlation of the target cells can be used to perform weighted adjustment on the initial abnormal feature values ​​to obtain the true abnormal feature values ​​of the target cells, thereby providing a data basis for the subsequent calculation of abnormal gene expression indicators of the target cells.

[0088] In one embodiment of the present invention, the non-positive correlation degree of the target cells may be multiplied by the initial abnormal characteristic value, so as to achieve weighted adjustment of the initial abnormal characteristic value by utilizing the non-positive correlation degree of the target cells.

[0089] The expression of the actual abnormal characteristic value of the target cell can be specifically, for example, as follows:

[0090] A ′ =norm(|P-1|)×A

[0091] Among them, A ′represents the true abnormal characteristic value of the target cell; P represents the correlation parameter of the target cell; A represents the initial abnormal characteristic value of the target cell; norm(|P-1|) represents the non-positive correlation degree of the target cell; norm() represents the normalization function.

[0092] The same method as above can be used to obtain each cell. Since the multiple single cells obtained in the embodiment of the present invention belong to the same tissue to be tested, under normal circumstances, the gene expression of the preset regulatory genes and the preset kinase genes of each single cell have similar characteristics, and the abnormal phenomenon only occurs in a small number of cells. Therefore, the difference in the real abnormal characteristic value between the target cell and other cells except the target cell can be analyzed. The greater the difference in the real abnormal characteristic value between the target cell and other cells, it means that the target cell has abnormal gene expression, and then it means that the gene expression spectrum of the target gene is abnormal. At the same time, considering that the smaller the distance between other cells and the target cell, the greater its reference value in the calculation and analysis process, the distance between the target cell and other cells can be further combined, so as to accurately reflect the possibility of abnormal expression of the preset regulatory genes and preset kinase genes of the target cell through the obtained gene expression abnormality index, thereby improving the accuracy of subsequent abnormal detection of the gene expression spectrum of the target cell.

[0093] Preferably, in one embodiment of the present invention, the method for obtaining the abnormal gene expression index of the target cell specifically includes:

[0094] See also Figure 3 , which shows a flow chart of a method for obtaining abnormal gene expression indicators of target cells provided by one embodiment of the present invention.

[0095] Step S301: According to the distance between the target cell and other cells except the target cell, the true abnormal characteristic values ​​of all other cells except the target cell are adjusted to obtain a reference abnormal characteristic value of the target cell.

[0096] Since the closer the distance between cells other than the target cell and the target cell is, the greater the reference value of the other cell in the calculation and analysis is, the true abnormal characteristic values ​​of all cells other than the target cell can be weighted and adjusted according to the distance between the target cell and the other cells other than the target cell to obtain the reference abnormal characteristic value of the target cell. The reference abnormal characteristic value reflects the abnormal conditions of all cells other than the target cell, which is used for subsequent comparison with the true abnormal characteristic value of the target cell, and then the abnormal gene expression index of the target cell is calculated and analyzed.

[0097] Preferably, in one embodiment of the present invention, the method for obtaining the reference abnormal characteristic value of the target cell specifically includes:

[0098] Any cell other than the target cell is used as a reference cell. The closer the distance between the reference cell and the target cell is, the greater the reference value of the reference cell in the calculation. Therefore, the distance between the reference cell and the target cell can be first negatively correlated and normalized to obtain the reference value parameter of the reference cell. The larger the reference value parameter is, the greater the reference value of the reference cell is.

[0099] In one embodiment of the present invention, the inverse of the distance between the reference cell and the target cell can be first used as the initial reference value of the reference cell, and the initial reference value of each cell other than the target cell can be obtained in the same way. Then, the initial reference value of the reference cell is used as the numerator, the cumulative value of the initial reference values ​​of all cells is used as the denominator, and the ratio is used as the reference value parameter of the reference cell, thereby achieving negative correlation normalization processing of the distance between the reference cell and the target cell, and ensuring that the subsequently calculated cumulative value of the reference value parameter of each cell other than the target cell is equal to 1, which is convenient for subsequent weighted adjustment.

[0100] In other embodiments of the present invention, negative correlation mapping may be performed on the distance between the reference cell and the target cell using a negative correlation function such as a negative exponential function with the natural exponent e as the base, which is not limited here.

[0101] The reference value parameters of each cell except the target cell can be obtained by the same method as above.

[0102] Then, the reference value parameter of each other cell except the target cell is used to perform weighted adjustment on the true abnormal characteristic value of each other cell to obtain the reference abnormal characteristic value of the target cell.

[0103] In one embodiment of the present invention, the reference value parameter of each other cell except the target cell can be used to perform weighted summation on the real abnormal characteristic value of each other cell, thereby obtaining the reference abnormal characteristic value of the target cell.

[0104] Step S302: normalize the difference between the real abnormal characteristic value of the target cell and the reference abnormal characteristic value to obtain the abnormal gene expression index of the target cell.

[0105] The larger the difference between the true abnormal characteristic value of the target cell and the reference abnormal characteristic value, the greater the possibility that the target cell has abnormal gene expression. Therefore, the difference between the true abnormal characteristic value of the target cell and the reference abnormal characteristic value can be normalized, and the possibility of abnormal gene expression in the target cell can be reflected by the obtained abnormal gene expression index. The larger the abnormal gene expression index, the greater the possibility that the gene expression spectrum of the target cell is abnormal, which is convenient for improving the accuracy of abnormal detection of cell gene expression spectrum.

[0106] The expression of the abnormal gene expression index of the target cell can be specifically, for example, as follows:

[0107]

[0108] Among them, U represents the abnormal gene expression index of the target cell; A ′ Indicates the true abnormal characteristic value of the target cell; D n represents the distance between the nth cell other than the target cell and the target cell; D m represents the distance between the mth cell and the target cell except the target cell; A ′ n represents the true abnormal characteristic value of the nth cell other than the target cell; N represents the number of cells other than the target cell; represents the initial reference value of the nth cell except the target cell; represents the initial reference value of the mth cell except the target cell; represents the cumulative value of the initial reference values ​​of all other cells except the target cell; Represents the reference value parameter of the nth cell other than the target cell; Represents the reference abnormal feature value of the target cell; norm() represents the normalization function.

[0109] The abnormal gene expression index of each cell can be obtained by the same method as above.

[0110] The abnormality detection module 104 is used to perform abnormality detection on the gene expression profile of the target cell based on the abnormal gene expression index of the target cell.

[0111] The abnormal gene expression index can accurately reflect the possibility of abnormal expression of preset regulatory genes and preset kinase genes in target cells. The larger the abnormal gene expression index is, the greater the possibility that the gene expression profile of the target cell is abnormal. Therefore, based on the abnormal gene expression index of the target cell, the gene expression profile of the target cell can be detected for abnormality, thereby improving the accuracy of abnormal detection of the gene expression profile of the target cell.

[0112] Preferably, in one embodiment of the present invention, the method for detecting abnormalities in the gene expression profile of target cells specifically includes:

[0113] Since the larger the abnormal gene expression index is, the greater the possibility that the gene expression spectrum of the target cell is abnormal, a threshold can be set. If the abnormal gene expression index of the target cell is greater than the preset second threshold, it is considered that the gene expression spectrum of the target cell is abnormal. Based on the same method, the abnormality of the gene expression spectrum of other cells can be detected, wherein the preset second threshold is set to 0.8. The specific value of the preset second threshold can also be set by the implementer according to the specific implementation scenario, and is not limited here.

[0114] In summary, the embodiment of the present invention first obtains the gene expression profiles of different cells in the tissue to be tested and the distance between two cells, wherein the gene expression profile includes the gene expression levels of preset regulatory genes and preset kinase genes at each moment in a preset time period; any cell is taken as a target cell, and the change rate value at each moment is obtained according to the change of the gene expression level of the preset regulatory gene of the target cell at each moment; the initial abnormal characteristic value of the target cell is obtained according to the difference of the change rate values ​​of the preset regulatory gene of the target cell at each moment in a preset window at each moment; the correlation of the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell at each moment is analyzed to obtain the correlation parameter of the target cell; based on the correlation parameter of the target cell, the initial abnormal characteristic value is corrected to obtain the real abnormal characteristic value of the target cell; according to the difference and distance of the real abnormal characteristic values ​​between the target cell and other cells except the target cell, the gene expression abnormality index of the target cell is obtained; based on the gene expression abnormality index of the target cell, the gene expression profile of the target cell is detected for abnormality.

[0115] It should be noted that the sequence of the above embodiments of the present invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0116] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. An intelligent anomaly detection system for single-cell expression profiles, characterized in that: The system comprises: A data acquisition module, used to acquire gene expression profiles of different cells in the tissue to be tested and the distances between any two cells, wherein the gene expression profiles include the gene expression levels of preset regulatory genes and preset kinase genes at each moment in a preset time period; The abnormality analysis module is used to take any cell as a target cell, and obtain the change rate value at each moment according to the change of the gene expression level of the preset regulatory gene of the target cell at each moment; and obtain the initial abnormal characteristic value of the target cell according to the difference of the change rate value of the preset regulatory gene of the target cell at each moment in the preset window at each moment; A correction module is used to analyze the correlation between the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell at each moment to obtain the correlation parameter of the target cell; based on the correlation parameter of the target cell, the initial abnormal characteristic value is corrected to obtain the real abnormal characteristic value of the target cell; according to the distance between the target cell and other cells except the target cell, the real abnormal characteristic value of the target cell, and the real abnormal characteristic value of other cells except the target cell, the gene expression abnormality index of the target cell is obtained; An abnormality detection module, used for performing abnormality detection on the gene expression profile of the target cell based on the abnormal gene expression index of the target cell; The step of obtaining the abnormal gene expression index of the target cell according to the distance between the target cell and other cells except the target cell, the true abnormal characteristic value of the target cell, and the true abnormal characteristic value of other cells except the target cell comprises: According to the distance between the target cell and other cells except the target cell, the true abnormal characteristic values ​​of all other cells except the target cell are adjusted to obtain a reference abnormal characteristic value of the target cell; Normalizing the difference between the true abnormal characteristic value of the target cell and the reference abnormal characteristic value to obtain a gene expression abnormality index of the target cell; The step of adjusting the true abnormal characteristic values ​​of all other cells except the target cell according to the distance between the target cell and other cells except the target cell to obtain the reference abnormal characteristic value of the target cell comprises: Any other cell except the target cell is used as a reference cell, and the distance between the reference cell and the target cell is negatively correlated and normalized to obtain a reference value parameter of the reference cell; The reference value parameters of each other cell except the target cell are used to perform weighted adjustment on the true abnormal characteristic value of each other cell to obtain the reference abnormal characteristic value of the target cell, wherein the cumulative value of the reference value parameters of each other cell except the target cell is equal to 1.

2. The intelligent anomaly detection system for single cell expression profile according to claim 1, characterized in that: The obtaining of the change rate value at each moment according to the change of the gene expression level of the preset regulatory gene of the target cell at each moment comprises: Using each moment in the gene expression profile of the target cell and the two-dimensional data points consisting of the gene expression level of the preset regulatory gene at each moment as the sample points of the preset regulatory gene of the target cell at each moment; According to the change in relative position between sample points of the preset regulatory gene of the target cell at each moment, the change rate value at each moment is obtained.

3. The intelligent anomaly detection system for single cell expression profile according to claim 1, characterized in that: The obtaining of the initial abnormal characteristic value of the target cell according to the difference of the change rate value of the preset regulatory gene of the target cell at each moment in the preset window at each moment comprises: Taking any moment as a target moment, obtaining the abnormal factor at the target moment according to the difference in the change rate values ​​of the preset regulatory gene of the target cell at adjacent moments in the preset window of the target moment; Based on the abnormal factors at each moment, screening out the abnormal moments of the preset regulatory genes of the target cells from all moments; The abnormal factors of the target cell at all the abnormal moments are integrated to obtain the initial abnormal characteristic value of the target cell.

4. The intelligent anomaly detection system for single cell expression profile according to claim 3, characterized in that: The step of obtaining the abnormal factor at the target moment according to the difference in the change rate values ​​of the preset regulatory gene of the target cell at adjacent moments in the preset window at the target moment comprises: Any two adjacent moments in the preset window of the target moment are taken as a moment group of the target moment; According to the difference of the change rate values ​​between two moments in each moment group, obtaining the rate change difference value of each moment group; Analyze the discreteness of the rate change difference values ​​of all time groups of the target time to obtain the rate change disorder degree of the target time; The rate change disorder at the target moment is normalized to obtain the abnormal factor at the target moment.

5. The intelligent anomaly detection system for single cell expression profile according to claim 3, characterized in that: The abnormal moments of the preset regulatory genes of the target cells are screened out from all the moments based on the abnormal factors at each moment, including: The moment when the abnormal factor is greater than the preset first threshold is regarded as the abnormal moment of the preset regulatory gene of the target cell.

6. The intelligent anomaly detection system for single cell expression profile according to claim 1, characterized in that: The analysis of the correlation between the gene expression levels of the preset regulatory gene and the preset kinase gene of the target cell at each moment to obtain the correlation parameter of the target cell includes: The Pearson correlation coefficient of the gene expression levels between the preset regulatory gene and the preset kinase gene of the target cell at each moment is used as the correlation parameter of the target cell.

7. The intelligent anomaly detection system for single cell expression profile according to claim 1, characterized in that: The correcting the initial abnormal characteristic value based on the correlation parameter of the target cell to obtain the real abnormal characteristic value of the target cell comprises: Normalizing the difference between the correlation parameter of the target cell and a constant 1 to obtain a non-positive correlation degree of the target cell; The non-positive correlation degree of the target cell is used to perform weighted adjustment on the initial abnormal characteristic value to obtain the true abnormal characteristic value of the target cell.

8. The intelligent anomaly detection system for single cell expression profile according to claim 1, characterized in that: The abnormal detection of the gene expression profile of the target cell based on the abnormal gene expression index of the target cell comprises: If the gene expression abnormality index of the target cell is greater than a preset second threshold, it is considered that the gene expression profile of the target cell is abnormal.

Images