Micro-lesion residual detection method, device and equipment based on single-cell sequencing technology and storage medium

Abstract

The application discloses a micro-lesion residual detection method and device based on single-cell sequencing technology, equipment and a storage medium. The method comprises the following steps: acquiring single-cell transcriptome sequencing data of a sample to be detected; performing cell clustering by using the single-cell transcriptome sequencing data; performing cell subtype annotation on the clustered cells; calculating the gene copy number variation value of the annotated cells; and determining whether the cells are micro-lesion residual positive according to the gene copy number variation value (CNV). The application designs a micro-lesion residual detection method based on single-cell sequencing technology, analyzes the CNV value of a single cell, thereby identifying MRD at the level of a large number of single cells, and is beneficial to inferring the malignant degree, heterogeneity, source similarity and drug resistance mutation of cells and the like, can be used for exploring the mechanism of the out-of-control body inflammation reaction of MRD patients, and guides the personalized treatment scheme suitable for the patients.

Description

Technical Field

[0001] This invention belongs to the field of gene detection technology and relates to methods, devices, equipment and storage media for detecting residual microlesions based on single-cell sequencing technology. Background Technology

[0002] Minimal residual disease (MRD) refers to a small number of malignant cells remaining in a patient after treatment remission, without any clinical symptoms or obvious signs of disease. The presence of MRD in a patient means that residual lesions still exist after cancer treatment, and the residual cancer cells may become active and begin to proliferate, which may lead to disease recurrence. Therefore, MRD detection can be used in clinical practice to guide treatment, measure the effectiveness of treatment, predict which patients are at risk of recurrence, assist in dynamic monitoring and confirmation of disease remission, detect signs of recurrence early, and start treatment as early as possible (LEUNG W, PUI CH, COUSTAN-SMITHE, et al. De:ectable minimal residual disease before hematopoietic cell transplantation is prognosli bul does not precludecure for child:en with very-high-risk leukemia[J].Blood,2):2,120(2):468-472).

[0003] Existing MRD detection technologies include flow cytometry, ddPCR, and NGS sequencing. Among them, ddPCR and NGS sequencing detect malignant cell-specific nucleic acid sequences (ctDNA) with a sensitivity as high as 0.01% (Della Starza I, Nunes V, Cavalli M, et al. Comparative analysis between RQ-PCR and digital-droplet-PCR of immunoglobulin / T cell receptor gene rearrangements to monitor minimal residual disease in acute lymphoblastic leukaemia[J]. Br. J. Haematol. 2016; 174: 541-549.), but cannot obtain information at the live cell level. Flow cytometry can detect circulating tumor cells (CTCs) and can obtain basic information on cell number and surface proteins to a certain extent. However, the ctDNA of malignant cells is highly heterogeneous, and the tumor mutations carried by different cancer types, and even different patients with the same cancer type, vary greatly. Flow cytometry has difficulty analyzing these differences.

[0004] In conclusion, developing comprehensive and highly sensitive MRD detection methods is one of the most pressing issues to be addressed in the field of MRD. Summary of the Invention

[0005] To address the shortcomings of existing technologies and practical needs, this invention provides a method, device, equipment, and storage medium for detecting residual microlesions based on single-cell sequencing technology, enabling comprehensive and highly sensitive identification of MRD at the single-cell level.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for detecting residual microlesions based on single-cell sequencing technology for purposes other than disease diagnosis and / or treatment, the method comprising:

[0008] Obtain single-cell transcriptome sequencing data of the sample to be tested;

[0009] Cell clustering was performed using the single-cell transcriptome sequencing data.

[0010] Cell subtype annotation was performed on the cells after clustering.

[0011] Calculate the gene copy number variation value of annotated cells, and determine whether the cells are positive for microfocal residual disease based on the gene copy number variation value;

[0012] The method for judgment includes: comparing the gene copy number variation values ​​of each cell, and determining that cells with gene copy number variation values ​​that are significantly different from those of other cells are positive for microfocal residual disease.

[0013] Preferably, the criteria for determining significant differences are as follows: analysis is performed using the inferCNV function, and the relative gene expression levels on each chromosome are displayed in the form of a heatmap, allowing for a direct calculation of the CNV value for each cell. Compared to normal cells, tumor cells exhibit large-scale overexpression or underexpression of genes. Therefore, the average gene expression value across the entire genome of normal cells (non-tumor cell types) is used as a control reference value. The CNV value represented by abnormal gene expression in tumor cells is calculated, and cells are sorted from high to low CNV values. Cells with significantly high CNV values ​​are considered minimal residual disease positive.

[0014] In this invention, normal and malignant cells are distinguished from samples by analyzing single-cell transcriptome data. Single-cell copy number variation (CNV) analysis, based on transcriptome data, uses gene expression levels across different samples or cell types to predict large-scale chromosome-level CNVs, identifying cells with abnormal copy numbers, thereby determining whether a patient has minimum residual disease (MRD), and subsequently predicting recurrence prognosis and the effectiveness of postoperative adjuvant therapy. Our strategy can accurately infer tumor cells through single-cell CNV prediction analysis, analyze tumor cell heterogeneity, detect minimum residual disease (MRD), and assist in predicting patient prognosis.

[0015] This invention presents a method for detecting residual microlesions (MRDs) based on single-cell sequencing technology. It utilizes single-cell transcriptome sequencing data for cell clustering and subtype annotation, and further analyzes copy number variation (CNV) values. Based on the CNV values ​​of individual cells, MRDs are identified at a massive single-cell level, comprehensively and accurately obtaining the number of malignant cells and their heterogeneity. This overcomes the seemingly insurmountable tumor heterogeneity problem encountered in conventional tumor tissue biopsies, providing information on cellular gene expression and functional characteristics. This information is closely related to tumor development, evolution, dormancy and drug resistance, metastasis, and recurrence. The detection method is simple, real-time, precise down to the cellular level, and non-invasive. Single-cell transcriptome detection can increase the likelihood and sensitivity of biomarkers responding to disease. It is of great significance for early disease diagnosis, prognostic assessment, disease monitoring, and personalized treatment.

[0016] The method for detecting residual microlesions of the present invention can be used to guide personalized treatment plans suitable for patients, and can also be applied to applications for non-disease diagnosis and / or treatment purposes, such as helping to infer characteristics such as the degree of malignancy, heterogeneity, and similarity of origin of cells. At the same time, by detecting the cell types and characteristics of other immune cells in body fluids, the immune level and characteristics of MRD patients can be analyzed from multiple single-cell information dimensions, and basic scientific research can be conducted to explore the mechanism of uncontrolled inflammatory response in MRD patients.

[0017] Preferably, the method for cell clustering includes:

[0018] use The program generates a gene expression matrix from the raw single-cell transcriptome sequencing data. The generated gene expression matrix is ​​imported into the Seurat program using the read.table function. The data undergoes quality control and filtering. The data is standardized using SCTransform. The FindVariableFeatures function is used to identify highly variable genes. The ScaleData function is used to scale the data. The RunPCA function is used for linear dimensionality reduction. After determining the data dimensionality, the KNN algorithm is used for clustering based on the Seurat program. The FindClusters function is used for cell clustering. Then, non-linear dimensionality reduction (UMAP / tSNE) is performed.

[0019] Preferably, the method for annotating cell subtypes includes:

[0020] Cell identity is identified using the Hypergeometric Test (HGT) of the Cell-ID program to obtain the most characteristic gene set of the cell. The characteristic gene set of known cell types in the SynEcoSys database is manually corrected. When the cell type has the smallest HGT p value, the identity of each cell is determined. The frequency of cell types corresponding to each cell identity in the cell cluster is calculated, and the cell type with the highest frequency is selected as the annotation of the cell cluster subtype.

[0021] Preferably, the method for calculating the gene copy number variation value includes:

[0022] Using the InferCNV software package, with non-malignant cells as the baseline control, the cell gene copy number variation value was calculated to estimate the frequency of large chromosomal insertions or deletions.

[0023] In this invention, the inferCNV function is used for analysis. Across the entire genome, the gene expression of each tumor cell is compared to the average expression or "normal" reference gene expression, and the relative gene expression levels on each chromosome are displayed in the form of a heatmap. This allows for intuitive calculation of whether tumor cell genomes exhibit large-scale overexpression or underexpression compared to normal cells.

[0024] Preferably, the sample to be tested can be ascites from a gastric cancer patient.

[0025] It is understood that this invention is based on the analysis of single-cell transcriptome data, and all commonly used single-cell transcriptome sequencing analysis methods in this field are applicable to this invention, such as those utilizing... The scRNA-seq library was constructed using the Singleron Biotechnologies cellular RNA library kit according to the manufacturer's instructions. The library was diluted to 4 nM and sequenced using an Illumina Novaseq 6000 sequencing platform in 150 bp paired-end sequencing mode.

[0026] In a second aspect, the present invention provides a device for detecting residual microlesions, the device being used to perform the steps of the method for detecting residual microlesions described in the first aspect, including:

[0027] Data input module:

[0028] Single-cell transcriptome sequencing data used to input the sample to be tested;

[0029] This is used to perform cell clustering using the single-cell transcriptome sequencing data, annotate cell subtypes after clustering, and calculate gene copy number variation values ​​of annotated cells.

[0030] Results analysis and output module:

[0031] This is used to determine whether a cell is positive for residual microlesions based on gene copy number variation values, and outputs the residual microlesion detection results.

[0032] The methods for making the determination include:

[0033] By comparing the gene copy number variation values ​​of each cell, cells with gene copy number variation values ​​that are significantly different from those of other cells are identified as having positive residual microlesions.

[0034] Preferably, the criteria for determining significant differences are as follows: analysis is performed using the inferCNV function, and the relative gene expression levels on each chromosome are displayed in the form of a heatmap, allowing for a direct calculation of the CNV value for each cell. Compared to normal cells, tumor cells exhibit large-scale overexpression or underexpression of genes. Therefore, the average gene expression value across the entire genome of normal cells (non-tumor cell types) is used as a control reference value. The CNV value represented by abnormal gene expression in tumor cells is calculated, and cells are sorted from high to low CNV values. Cells with significantly high CNV values ​​are considered minimal residual disease positive.

[0035] Preferably, the method for cell clustering includes:

[0036] use The program generates a gene expression matrix from the raw single-cell transcriptome sequencing data. The generated gene expression matrix is ​​imported into the Seurat program using the read.table function. The data undergoes quality control and filtering. The data is standardized using SCTransform. The FindVariableFeatures function is used to identify highly variable genes. The ScaleData function is used to scale the data. The RunPCA function is used for linear dimensionality reduction. After determining the data dimensionality, the KNN algorithm is used for clustering based on the Seurat program. The FindClusters function is used for cell clustering. Then, non-linear dimensionality reduction (UMAP / tSNE) is performed.

[0037] Preferably, the method for annotating cell subtypes includes:

[0038] Cell identity is identified using the Hypergeometric Test (HGT) of the Cell-ID program to obtain the most characteristic gene set of the cell. The characteristic gene set of known cell types in the SynEcoSys database is manually corrected. When the cell type has the smallest HGT p value, the identity of each cell is determined. The frequency of cell types corresponding to each cell identity in the cell cluster is calculated, and the cell type with the highest frequency is selected as the annotation of the cell cluster subtype.

[0039] Preferably, the method for calculating the gene copy number variation value includes:

[0040] Using the InferCNV software package, with non-malignant cells as the baseline control, the cell gene copy number variation value was calculated to estimate the frequency of large chromosomal insertions or deletions.

[0041] In this invention, the inferCNV function is used for analysis. Across the entire genome, the gene expression of each tumor cell is compared to the average expression or "normal" reference gene expression, and the relative gene expression levels on each chromosome are displayed in the form of a heatmap. This allows for intuitive calculation of whether tumor cell genomes exhibit large-scale overexpression or underexpression compared to normal cells.

[0042] Thirdly, the present invention provides an electronic device, comprising: one or more processors; and a storage device having stored one or more programs thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the steps in the method for detecting residual microlesions as described in the first aspect.

[0043] Fourthly, the present invention provides a computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps in the method for detecting residual micro-lesions described in the first aspect.

[0044] Compared with the prior art, the present invention adopts the following technical solution:

[0045] This invention designs a method for detecting residual microlesions based on single-cell sequencing technology. It utilizes single-cell transcriptome sequencing data for cell clustering and cell subtype annotation, and further analyzes the copy number variation (CNV) value. Based on the CNV value of a single cell, MRD can be identified at the massive single-cell level. This method can be used to guide personalized treatment plans suitable for patients and to conduct basic scientific research to explore the mechanism of uncontrolled inflammatory response in MRD patients. Attached Figure Description

[0046] Figure 1 This is a diagram showing the CNV analysis results of epithelial cells from Example 1 of the present invention.

[0047] Figure 2 This is a graph showing the CNV values ​​of various cell types in the sample of Example 1 of this invention;

[0048] Figure 3 This is a graph showing the CNV value results and significance analysis results of epithelial cells with high CNV values ​​and other cell types in Example 1 of this invention;

[0049] Figure 4 This is a graph showing the CNV analysis results of various cell types in Example 2 of this invention;

[0050] Figure 5 This is a graph showing the CNV values ​​of various cell types in the sample of Example 2 of this invention. Detailed Implementation

[0051] To further illustrate the technical means and effects of this invention, the following description, in conjunction with embodiments and accompanying drawings, provides a further explanation of the invention. It is understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it.

[0052] Where specific techniques or conditions are not specified in the examples, they shall be performed in accordance with the techniques or conditions described in the literature in this field, or in accordance with the product instructions. Reagents or instruments whose manufacturers are not specified are all conventional products that can be purchased through legitimate channels.

[0053] Example 1

[0054] This embodiment performs MRD detection.

[0055] Single-cell sequencing was performed on ascites fluid from a gastric cancer patient who had undergone immunotherapy. The library preparation and sequencing methods for the single-cell transcriptome were as follows: First, using… The cellular RNA library kit (Singleron Biotechnologies) was used to construct scRNA-seq libraries according to the instructions. The libraries were diluted to 4 nM and sequenced using the Illumina Novaseq 6000 sequencing platform in 150 bp paired-end sequencing mode.

[0056] Bioinformatics analysis of the sequencing results begins with clustering the cells. Specific methods include:

[0057] use The program generates a gene expression matrix from the raw single-cell transcriptome sequencing data. The generated gene expression matrix is ​​imported into the Seurat program using the `read.table` function, and cell clustering is performed using the `FindClusters` function. The hypergeometric test (HGT) of the Cell-ID program is used to identify cell identities, obtaining the most characteristic gene set for each cell. The characteristic gene sets of known cell types in the SynEcoSys database are manually refined; the cell type with the lowest HGT p-value is identified as the cell type. The frequency of cell types corresponding to each cell identity in the cell cluster is calculated, and the cell type with the highest frequency is selected as the subtype annotation for that cell cluster. The results are shown in Table 1.

[0058] Table 1

[0059]

[0060]

[0061] Using myeloid cells, T cells, plasma cells, and B cells as controls, the copy number variation (CNV) of epithelial cells was analyzed, and the results are as follows: Figure 1 As shown, the CNV values ​​of various cell types are displayed using a violin plot. Figure 2 Further analysis was conducted on the four epithelial cells with the highest CNV values, and the differences in CNV values ​​between them and the remaining cells were analyzed. Figure 3 (Among them, the up group was positive for minimal residual disease, and the other groups were negative for minimal residual disease).

[0062] The ascites cell clustering analysis revealed a total of 6164 cells, the majority of which were immune cells (99.64%), including myeloid cells, T cells, plasma cells, B cells, monocytes (PMCs), and a small number of epithelial cells (0.36% of total cells, Table 1). Epithelial cell CNV analysis, using myeloid cells, T cells, plasma cells, and B cells as controls, showed a small number of epithelial cells (…). Figure 1 The frequency of chromosome insertions and deletions was high. Using a violin plot to display the CNV values ​​of each cell subset, it was found that the CNV of epithelial cells was clearly divided into two parts. Figure 2 Among them, the CNV values ​​of the four epithelial cells in the upper frame were 541.42, 452.31, 435.38, and 549.38, respectively, while the CNV values ​​of the remaining 18 epithelial cells were below 300, showing significant differences. The four cells with high CNV values ​​in the upper frame also showed significant differences in CNV values ​​compared to other control cell types. These four epithelial cells with high CNV values ​​were identified as MRD (Mutual Disorder of Chromosomal Aberration), and their chromosomal aberrations are shown in the CNV analysis results, accounting for 0.06% of the total cells. Figure 3 ).

[0063] Example 2

[0064] This embodiment performs MRD detection.

[0065] Single-cell sequencing was performed on an ascites sample from a patient with metastatic gastric cancer. Following the method in Example 1, bioinformatics analysis was performed on the sequencing results. First, the cells were clustered, and then each subpopulation was annotated to obtain the cell type composition of the sample, as shown in Table 2.

[0066] Table 2

[0067] Cell types Cell count percentage Myeloid cells 4897 95.35％ Mesothelial cells 179 3.49％ goblet cells 33 0.64％ plasma cells 27 0.53％

[0068] Using myeloid cells and plasma cells as controls, copy number variation (CNV) of genes in mesothelial and goblet cells was analyzed. (See attached data.) Figure 4 The CNV values ​​of various cell types are displayed using violin plots. Figure 5 The cell clustering annotation results from the patient's ascites fluid showed a total of 5136 cells, including myeloid cells, plasma cells, mesothelial cells, and goblet cells. CNV analysis of mesothelial cells and goblet cells was performed using myeloid cells and plasma cells as controls to demonstrate chromosomal insertions and deletions. Figure 4 Further analysis using violin plots to visualize CNV values ​​for each cell subpopulation revealed no significant differences in CNV values ​​among cell types. Figure 5 However, one goblet cell had a significantly higher CNV value than all other cells. This goblet cell with a high CNV value was identified as MRD, accounting for 0.02% of the total cells.

[0069] In summary, this invention designs a method for detecting residual microlesions based on single-cell sequencing technology. It utilizes single-cell transcriptome sequencing data for cell clustering and cell subtype annotation, and further analyzes the copy number variation (CNV) value. Based on the CNV value of a single cell, MRD can be identified at the massive single-cell level. This method can be used to guide personalized treatment plans suitable for patients and to conduct basic scientific research to explore the mechanism of uncontrolled inflammatory response in MRD patients.

[0070] The applicant declares that the detailed method of the present invention is illustrated by the above embodiments, but the present invention is not limited to the above detailed method, that is, it does not mean that the present invention must rely on the above detailed method to be implemented. Those skilled in the art should understand that any improvements to the present invention, equivalent substitutions of the raw materials of the product of the present invention, addition of auxiliary components, selection of specific methods, etc., all fall within the protection scope and disclosure scope of the present invention.

Claims

1. A method for detecting residual microlesions based on single-cell sequencing technology for purposes other than disease diagnosis and / or treatment, characterized in that, The method includes: Obtain single-cell transcriptome sequencing data of the sample to be tested; Cell clustering was performed using the single-cell transcriptome sequencing data. Cell subtype annotation was performed on the cells after clustering. Calculate the gene copy number variation value of annotated cells, and determine whether the cells are positive for microfocal residual disease based on the gene copy number variation value; The method of judgment includes: comparing the gene copy number variation values ​​of each cell, and determining that cells whose gene copy number variation values ​​are significantly different from those of other cells are positive for microfocal residual disease. The method for annotating cell subtypes includes: Cell identity is identified using the hypergeometric test of the Cell-ID program to obtain the most characteristic gene set of the cell. The characteristic gene set of known cell types in the SynEcoSys database is manually corrected. When the cell type has the minimum HGT p value, the identity of each cell is determined. The frequency of cell type corresponding to each cell identity in the cell cluster is calculated, and the cell type with the highest frequency is selected as the annotation of the cell cluster subtype. The method for calculating the gene copy number variation value includes: Using the InferCNV software package, with non-malignant cells as the baseline control, the cell gene copy number variation value was calculated to estimate the frequency of large chromosomal insertions or deletions.

2. The method for detecting residual microlesions based on single-cell sequencing technology for non-disease diagnosis and / or treatment purposes, as described in claim 1, is characterized in that... The cell clustering method includes: Using CeleScope ® The program generates a gene expression matrix from the raw single-cell transcriptome sequencing data. The generated gene expression matrix is ​​imported into the Seurat program using the read.table function. The data undergoes quality control and filtering. The data is standardized using SCTransform. The FindVariableFeatures function is used to identify highly variable genes. The ScaleData function is used to scale the data, i.e., the RunPCA function is used for linear dimensionality reduction. After determining the data dimensionality, the KNN algorithm is used for clustering based on the Seurat program. The FindClusters function is used for cell clustering, followed by nonlinear dimensionality reduction.

3. The method for detecting residual micro-lesions based on single-cell sequencing technology for non-disease diagnosis and / or treatment purposes, as described in claim 1, is characterized in that... The sample to be tested includes ascites.

4. The method for detecting residual micro-lesions based on single-cell sequencing technology for non-disease diagnosis and / or treatment purposes, as described in claim 1, is characterized in that... The gene copy number variation value is the gene copy number variation value of epithelial cells.

5. A device for detecting residual micro-lesions, characterized in that, The detection device is used to perform the steps in the method for detecting residual microlesions according to any one of claims 1-4, including: Data input module: Single-cell transcriptome sequencing data used to input the sample to be tested; This is used to perform cell clustering using the single-cell transcriptome sequencing data, annotate cell subtypes after clustering, and calculate gene copy number variation values ​​of annotated cells. Results analysis and output module: This is used to determine whether a cell is positive for residual microlesions based on gene copy number variation values, and outputs the residual microlesion detection results. The method of judgment includes: comparing the gene copy number variation values ​​of each cell, and determining that cells whose gene copy number variation values ​​are significantly different from those of other cells are positive for microfocal residual disease. The method for annotating cell subtypes includes: Cell identity is identified using the Hypergeometric Test (HGT) of the Cell-ID program to obtain the most characteristic gene set of the cell. The characteristic gene set of known cell types in the SynEcoSys database is manually corrected. When the cell type has the smallest HGT p value, the identity of each cell is determined. The frequency of cell type corresponding to each cell identity in the cell cluster is calculated, and the cell type with the highest frequency is selected as the annotation of the cell cluster subtype. The method for calculating the gene copy number variation value includes: Using the InferCNV software package, with non-malignant cells as the baseline control, the cell gene copy number variation value was calculated to estimate the frequency of large chromosomal insertions or deletions.

6. The device for detecting residual micro-lesions according to claim 5, characterized in that, The cell clustering method includes: Using CeleScope ® The program generates a gene expression matrix from the raw single-cell transcriptome sequencing data. The generated gene expression matrix is ​​imported into the Seurat program using the read.table function. Data quality control and filtering are performed, data standardization is performed using SCTransform, highly variable genes are identified using the FindVariableFeatures function, data scaling is performed using the ScaleData function, and linear dimensionality reduction is performed using the RunPCA function. After determining the data dimensionality, KNN algorithm is used for clustering based on the Seurat program, cell clustering is performed using the FindClusters function, and then nonlinear dimensionality reduction is performed.

7. An electronic device, characterized in that, include: One or more processors; A storage device having stored one or more programs thereon, which, when executed by one or more processors, cause the one or more processors to perform the steps of the method for detecting residual microlesions as described in any one of claims 1-4.

8. A computer storage medium, characterized in that, It stores a computer program, wherein when the computer program is executed by a processor, it implements the steps in the method for detecting residual microlesions as described in any one of claims 1-4.

Images