Prediction method and system for chemotherapy resistance related genes of ovarian cancer

Abstract

This invention discloses a method and system for predicting chemotherapy resistance-related genes in ovarian cancer. It integrates single-cell transcriptomics, spatial transcriptomics, and the TCGA large-sample clinical database to study the drug resistance mechanism of ovarian cancer. It defines drug resistance characteristics from multiple dimensions such as cell abundance, transcriptional heterogeneity, immune evolution trajectory, and spatial distribution. It overcomes the shortcomings of traditional single-mic screening, such as high false positive rate and lack of mechanism and spatial validation, and significantly improves the reliability of drug resistance target screening.

Description

Technical Field

[0001] This invention relates to the fields of bioinformatics and precision oncology, and in particular to a method and system for predicting genes related to chemotherapy resistance in ovarian cancer. Background Technology

[0002] Ovarian cancer is one of the deadliest malignant tumors of the female reproductive system. Due to the lack of obvious early symptoms and effective screening methods, 70% of patients are diagnosed at an advanced stage (stage III and IV). Furthermore, patients often develop drug resistance during treatment, resulting in a high recurrence rate and poor treatment outcomes. Drug resistance is a complex phenomenon involving multiple mechanisms, currently classified into four main types according to research: transmembrane transport abnormalities, altered DNA damage repair, dysregulation of cancer-related signaling pathways, and epigenetic alterations.

[0003] Traditional gene expression analysis is usually based on average data from a large number of cells, which makes it difficult to capture the complexity of the tumor microenvironment. Existing technologies have long experimental verification cycles and high costs, and there is a lack of accurate predictive indicators for chemotherapy resistance in ovarian cancer. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method and system for predicting genes related to chemotherapy resistance in ovarian cancer, thereby improving the reliability of drug resistance target screening, in order to address the shortcomings of the existing technology.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for predicting ovarian cancer chemotherapy resistance-related genes, comprising the following steps: S1. Obtain TCGA-OV public cohort transcriptome data, perform differential expression analysis on the chemotherapy-sensitive group and chemotherapy-resistant group in the TCGA-OV public cohort transcriptome data to obtain the TCGA-OV differential gene set; download GSE154600, GSE184880, and GSE211956 datasets from the GEO public database, obtain single-cell transcriptome data and spatial transcriptome data of the ovarian cancer chemotherapy-sensitive group and chemotherapy-resistant group, and construct single-cell gene expression matrix and spatial transcriptome expression matrix containing spatial location information, respectively. S2. Extract epithelial cells from single-cell transcriptome data, use the inferCNV algorithm to infer copy number variations, and identify malignant epithelial cells; perform subpopulation subgrouping and differential expression analysis on malignant epithelial cells, identify drug-resistant subpopulations, and construct a single-cell drug resistance differential gene set; S3. Using the CellphoneDB receptor-ligand database, calculate the interaction strength between the malignant epithelial cells and other cell populations; compare the calculation results of the chemotherapy-sensitive group and the chemotherapy-resistant group, extract cell communication pathways whose mean communication interaction value is upregulated in the chemotherapy-resistant samples and whose permutation test statistical probability value indicating inter-group communication differences is less than a first set value, and construct a microenvironmental drug resistance communication network; take the intersection of the TCGA-OV differential gene set and the single-cell drug resistance differential gene set to obtain common genes, and use the Kaplan-Meier algorithm to combine the common genes with overall survival and progression-free interval prognostic data for survival analysis, screen out common genes that are statistically associated with prognosis and whose log-rank test probability value of survival prognostic differences is less than a second set value, and use these common genes as candidate drug resistance target genes to predict and characterize the microenvironmental immunosuppressive features of ovarian cancer chemotherapy resistance status.

[0006] The specific process for calculating the interaction strength between the malignant epithelial cells and other cell populations includes: obtaining receptor-ligand pairs based on the CellphoneDB receptor-ligand database; extracting the average expression levels of the ligands and receptors in the malignant epithelial cells and other cell populations; using the mean of the ligand expression levels and receptor expression levels between the malignant epithelial cells and other cell populations as the interaction strength of a specific receptor-ligand pair between groups; calculating the probability value of the interaction using a permutation test; and constructing a random distribution by randomly shuffling the cell type labels of the malignant epithelial cells and other cell populations, and repeatedly calculating the interaction strength to obtain the statistical probability value of the interaction of the specific receptor-ligand pair.

[0007] The method of the present invention further includes: S4. Based on the spatial transcriptome sequencing, predict the proportion of cell types at each spatial site using a deconvolution algorithm; analyze the difference in tumor cell purity between chemotherapy-sensitive and chemotherapy-resistant data in space, map the candidate drug-resistant target genes to the spatial atlas, verify the enrichment of the candidate drug-resistant target genes in drug-resistant lesions, and output the confirmed chemotherapy-resistant target genes.

[0008] The confirmed chemotherapy resistance target genes are KRT15 and SCEL.

[0009] The method of the present invention further includes: extracting the expression abundance of candidate transcription factors based on the spatial transcriptome expression matrix containing spatial location information, mapping the expression abundance to the corresponding physical coordinates of tissue slices, performing spatial expression profile reconstruction and analysis, and obtaining transcription factors with a wide expression trend in drug-resistant samples.

[0010] The transcription factors with a wide expression trend in the drug-resistant samples were CEBPB and SOX15.

[0011] Both the first setting value and the second setting value are 0.05.

[0012] The method of the present invention further includes: Based on the single-cell gene expression matrix, the UMAP algorithm was used for dimensionality reduction visualization and unsupervised clustering, identifying multiple cell subpopulations. These cell subpopulations include B cells, endothelial cells, epithelial cells, fibroblasts, myeloid cells, NK cells, pericytes, plasma cells, stromal cells, and T cells. Based on the single-cell gene expression matrix, T cell and NK cell subsets in the immune microenvironment were extracted and subdivided. The differences in cell subset abundance between groups were analyzed. The evolutionary trajectory of T cells was analyzed through pseudo-time series analysis, and drug resistance-related immune exhaustion subsets were identified.

[0013] As an inventive concept, the present invention also provides a prediction system for ovarian cancer chemotherapy resistance-related genes, including a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the steps of the above method.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention integrates single-cell transcriptomics, spatial transcriptomics, and the TCGA large-sample clinical database to study the drug resistance mechanism of ovarian cancer. It defines drug resistance characteristics from multiple dimensions, including cell abundance, transcriptional heterogeneity, immune evolution trajectory, and spatial distribution, overcoming the shortcomings of traditional single-mic screening, such as high false positive rates and lack of mechanistic and spatial validation, significantly improving the reliability of drug resistance target screening. This invention screened for KRT15 and SCEL genes, and after survival analysis, found them to be important molecular markers for clinically predicting chemosensitivity and prognosis in ovarian cancer. Simultaneously, this invention discovered that drug-resistant malignant epithelial cells can induce tumor microenvironment immunosuppression through abnormally activated MDK-NCL interaction axes, leading to impaired T cell proliferation and differentiation, functional exhaustion, and ultimately, immune surveillance failure and chemosensitivity. This mechanism suggests that targeted intervention of the MDK pathway and restoration of T cell immune function on the basis of chemotherapy may reverse chemosensitivity in ovarian cancer and improve clinical treatment efficacy. Attached Figure Description

[0015] Figure 1 Analysis of cell subset composition among different samples of single-cell transcriptome of ovarian cancer tumor microenvironment; Figure 2 A scatter plot is used to show the expression of marker genes in different cell types; Figure 3 The results of inferCNV for ovarian cancer epithelial tumor epithelial cells; Figure 4Count the cell types for different T cell subsets; Figure 5 Violin diagrams of FOXP3 and MKI67 in different T cell subsets; (a) Violin diagram of FOXP3 in different T cell subsets, (b) Violin diagram of MKI67 in different T cell subsets. Figure 6 Pseudo-temporal analysis of T cell subsets; (a) Cell subset grouping diagram of pseudo-temporal analysis of T cell subsets; (b) Pseudo-temporal evolution diagram of pseudo-temporal analysis of T cell subsets; Figure 7 Heatmaps showing the communication strength of the MK pathway across different cell types; (a) Heatmap of the communication strength of the MK pathway across different cell types in the drug-resistant group, (b) Heatmap of the communication strength of the MK pathway across different cell types in the drug-sensitive group. Figure 8 The following diagrams illustrate the intercellular communication networks of MK pathway-related ligand-receptor pairs in different cell types: (a) Intercellular communication network of MDK-SDC1 ligand-receptor pair, (b) Intercellular communication network of MDK-NCL ligand-receptor pair, and (c) Intercellular communication network of MDK-LRP1 ligand-receptor pair. Figure 9 Survival analysis curves for KRT15; (a) Kaplan-Meier curve of KRT15 expression level versus overall survival (OS); (b) Kaplan-Meier curve of KRT15 expression level versus progression-free survival (PFS); Figure 10 Survival analysis curves for SCEL; (a) Kaplan-Meier curve of SCEL expression level versus overall survival (OS); (b) Kaplan-Meier curve of SCEL expression level versus progression-free survival (PFS); Figure 11 Spatial expression profiles of CEBPB and SOX15 in drug-sensitive and drug-resistant groups; (a) Spatial expression distribution of CEBPB in the sensitive group sample; (b) Spatial expression distribution of SOX15 in the sensitive group sample; (c) Spatial expression distribution of CEBPB in the drug-resistant group sample; (d) Spatial expression distribution of SOX15 in the drug-resistant group sample. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1 This embodiment provides a method for predicting genes related to chemotherapy resistance in ovarian cancer, and the specific implementation process includes the following steps.

[0018] Step 1: Multidimensional Data Acquisition and Grouping: Acquire TCGA-OV public cohort transcriptome data; the TCGA-OV public cohort transcriptome data is divided into chemotherapy-sensitive and chemotherapy-resistant groups based on the duration of chemotherapy use or a platinum-free interval (PFI) of 6 months (grouped according to the progression-free period of chemotherapy use (≥ 6 months defined as chemotherapy-sensitive group, < 6 months defined as chemotherapy-resistant group)); simultaneously, download the GSE154600, GSE184880, and GSE211956 datasets from the GEO public database to obtain single-cell transcriptome data and spatial transcriptome data for the ovarian cancer chemotherapy-sensitive and chemotherapy-resistant groups, and construct single-cell gene expression matrices and spatial transcriptome expression matrices containing spatial location information, respectively.

[0019] Single-cell transcriptome data: integrated from the GEO public dataset. Divided into three groups: ① Primary ovarian cancer group (priCancer): 7 high-grade serous ovarian cancer primary lesion samples from the GSE184880 dataset, without chemotherapy response annotation, serving as baseline tumor controls; ② Chemotherapy-sensitive metastatic group (Cancer-Meta-S): 2 ovarian cancer metastatic lesion samples from the GSE154600 dataset, clinically assessed as chemotherapy-sensitive; ③ Chemotherapy-resistant metastatic group (Cancer-Meta-R): 3 ovarian cancer metastatic lesion samples from the GSE154600 dataset, clinically assessed as chemotherapy-resistant.

[0020] Spatial transcriptome data: Spatial transcriptome sequencing samples from the GEO database GSE211956 dataset were selected and grouped according to clinical treatment response: samples with good treatment response were assigned to the chemotherapy-sensitive group, and samples with poor treatment response were assigned to the chemotherapy-resistant group.

[0021] After merging the aforementioned GEO single-cell transcriptome data, the Seurat software was used to filter out low-quality cells with nFeature_RNA counts between 200 and 7500 and mitochondrial percentages >25%. Subsequently, the LogNormalize algorithm was used to normalize gene expression levels, and the PCA algorithm was applied for linear dimensionality reduction. To eliminate technical biases introduced by multi-source datasets, the Harmony software package was used for batch-free integration. Further, the DoubletFinder algorithm was used to remove potential doublets, and then t-SNE or UMAP algorithms were executed to achieve non-linear dimensionality reduction visualization. This data ultimately served as the standard input for subsequent differential gene (FindMarkers) identification and cell subpopulation analysis.

[0022] Step 2: Panoramic Cell Segmentation and Annotation of the Tumor Microenvironment: Based on the standardized single-cell gene expression matrix constructed in Step 1, this embodiment of the invention utilizes the UMAP method for visualization analysis, identifying 10 major cell subpopulations, including B cells, endothelial cells, epithelial cells, fibroblasts, bone marrow cells, NK cells, pericytes, plasma cells, stromal cells, and T cells. Proportional quantitative statistical results show that these 10 cell subpopulations are commonly present in primary lesions, chemotherapy-sensitive metastatic lesions, and chemotherapy-resistant metastatic lesions (Figure 1). There is no significant difference in the composition of cell types, but the relative abundance (cell proportion) of each cell subpopulation differs significantly between the drug-resistant and drug-sensitive groups.

[0023] To further confirm the biological attributes of the aforementioned cell subpopulations, this embodiment uses a characteristic gene dot map (…). Figure 2 The expression of classic marker genes in the above subgroups was verified, and the results are as follows: (1) The epithelial / tumor cell subsets significantly overexpressed marker genes such as KRT18, EPCAM, CD24, KRT19 and PAX8; (2) The endothelial cell subsets significantly overexpressed VWF, PECAM1, CLDN5 and PLVAP; (3) The plasma cell subsets significantly overexpressed MZB1, IGKC and DERL3; (4) The pericyte subsets significantly overexpressed RGS5 and COL18A1; (5) The NK cell subsets significantly overexpressed KLRC1 and did not express IGKC; (6) The immune subsets such as T cells significantly overexpress CD3D, CD3E and CD8A.

[0024] Step 3: Identification of Malignant Subpopulations of Tumors: Epithelial cell data were extracted from the single-cell gene expression matrix described in Step 1, and copy number variations were inferred using the inferCNV algorithm to identify malignant epithelial cells. Subpopulation subgrouping and differential expression analysis were performed on the malignant epithelial cells to identify drug-resistant subpopulations and construct a single-cell drug-resistant differential gene set. This embodiment of the invention uses the inferCNV tool to infer large-scale CNVs from scRNA-seq data. This embodiment of the invention uses epithelial cells from the single-cell gene expression matrix described in Step 1 as reference normal cells to analyze the changes in gene expression intensity in different regions of the tumor epithelial cell genome, displaying the relative gene expression levels on each chromosome in the form of a heatmap. The results showed that all epithelial cells were malignant tumor cells (Figure 3).

[0025] The identified malignant epithelial cells were further subdivided into subpopulations. The gene expression differences between chemotherapy-sensitive and drug-resistant malignant epithelial cells were compared. Genes with significant differential expression in the drug-resistant group were screened, and a single-cell drug resistance differential gene set was constructed.

[0026] Step 4: Analysis of T cell and NK cell subsets in the immune microenvironment T cells and NK cells were extracted from the single-cell gene expression matrix in step 2, and further subgroup subdivision, abundance difference and evolutionary trajectory analysis were carried out to analyze the drug resistance-related characteristics of the immune microenvironment.

[0027] (1) Differences in cell subpopulation abundance: In this embodiment of the invention, a fine subpopulation identification of T cells extracted from single-cell expression moments was performed, and a total of 7 subpopulations were identified: CD4+ T cells, CD8+ T cells, CD4-ANXA1+ T cells, Tem-GZMK+ effector memory T cells, CD8-GPR183 T cells, CD4-MKI67+ proliferating T cells, and naive T cells. The proportions of each cell type were compared between drug-resistant and drug-sensitive patients and between primary cancers (Figure 4). It was found that CD8+ T cells accounted for a higher proportion in primary cancer samples; CD4-ANXA1+ T cells dominated in chemotherapy-resistant samples and were specifically highly expressed only in drug-resistant metastatic samples, suggesting that they are closely related to chemotherapy resistance; Tem-GZMK+ T cells accounted for a higher proportion in chemotherapy-sensitive samples, which was associated with a good treatment response. Furthermore, the expression distribution of FOXP3 and MKI67 in each subgroup was presented using a violin diagram (Figure 5). It was found that no significant expression of FOXP3 was detected in T cells of either the sensitive or resistant group, while the expression level of the proliferation marker gene MKI67 in T cells of the sensitive group was significantly higher than that in the resistant group.

[0028] (2) T cell evolution exhaustion analysis: The differentiation trajectory of T cells was reconstructed through pseudo-temporal analysis (Figure 6). The results showed that sensitive T cells were able to differentiate smoothly along the pseudo-temporal evolution trajectory, eventually forming a CD4-MKI67+ subset in a highly proliferating state; while drug-resistant T cells were mostly arrested at the leading edge of the evolution trajectory, exhibiting a state of functional exhaustion with high ANXA1 expression. This obstruction of immune cell proliferation and differentiation prevents them from effectively clearing tumor cells, leading to the failure of immune surveillance mechanisms and tumor immune escape.

[0029] Step 5: Reconstruction of Microenvironment Cell Communication Networks and Elucidation of Drug Resistance Mechanisms Based on the single-cell gene expression matrix described in step 1, T cell and NK cell subsets in the immune microenvironment were extracted and subdivided. Combined with the CellphoneDB receptor-ligand database, the interaction strength between the malignant epithelial cells identified in step 3 and other cell populations was calculated. Abnormally active cell communication pathways in the drug-resistant group were extracted (Efremova M, Vento-Tormo M, Teichmann SA, et al., CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes. Nature Protocols. 2020; 15(4): 1484-1506.), thereby elucidating the drug resistance mechanism induced by the malignant tumor cells to remodel the microenvironment and escape immunity. The results were analyzed as follows: (1) Significant weakening of immune activation communication: The heatmap shows the communication strength of the MK pathway among different cell types (Figure 7). By comparing the communication strength networks of the sensitive and resistant groups, it was found that the most significant feature in the microenvironment of the resistant group was the overall loss of communication signals between malignant epithelial cells and T cells. Although malignant tumor cells retain basic immune recognition signals mediated by pathways such as HLA-E-KLRC1, the overall effective immune activation communication network is significantly weakened in the resistant group.

[0030] (2) MDK-NCL Interaction Axis Mediating Immunosuppressive Network: Microenvironmental cell communication network analysis showed that the signal intensity of the MDK (Midkine) pathway was significantly upregulated in the drug-resistant group. Comparing the interaction strength of multiple ligand-receptor pairs related to the MDK pathway (Figure 8), although interactions such as MDK-SDC1 and MDK-LRP1 were present, the MDK-NCL interaction axis showed abnormally strong global activity. This result suggests that malignant epithelial cells induce and dominate abnormal microenvironmental remodeling through abnormally high expression of strong MDK-NCL communication signals, enabling tumor cells to evade effective immune surveillance. The abnormal MDK-NCL communication network can serve as a key microenvironmental feature label for assessing and predicting chemotherapy resistance status in ovarian cancer.

[0031] Step 6: Screening for drug resistance genes and prognostic validation based on the TCGA-OV large-scale dataset Cross-validation of drug resistance genes based on a large clinical sample was performed to screen candidate drug resistance target genes. The specific process is as follows: 1. Construction of the TCGA-OV differentially expressed gene set: Based on the TCGA-OV public cohort transcriptome data obtained in step 1, the limma algorithm package was used to perform differential expression analysis between the chemotherapy-resistant and chemotherapy-sensitive groups. The selection criteria were adjusted P < 0.05 and |log2FC| ≥ 0.5. The results showed that 469 genes were significantly upregulated and 698 genes were significantly downregulated in the resistant group. Based on these results, the TCGA-OV differentially expressed gene set was constructed as background data for subsequent cross-validation.

[0032] 2. Validation of Drug Resistance Biomarkers and Survival Prognosis: The intersection of the single-cell drug resistance differential gene set constructed in step 3 and the TCGA-OV differential gene set was obtained to identify common genes. Subsequently, the expression levels of these common genes were extracted. Combined with the overall survival (OS) and progression-free interval (PFI) follow-up data of patients in the TCGA-OV cohort, survival analysis was performed using the Kaplan-Meier algorithm. With a log-rank test P < 0.05 as the threshold, genes significantly associated with prognosis were screened as candidate drug resistance target genes.

[0033] Ultimately, the core drug resistance target genes KRT15 and SCEL were identified. Survival analysis curves confirmed that high expression of KRT15 and SCEL was significantly negatively correlated with low survival rates in ovarian cancer patients (Figures 9 and 10). This confirms that KRT15 and SCEL genes can serve as core biomarkers for determining chemotherapy resistance in ovarian cancer.

[0034] Step 7: Spatial Map Validation: Based on the spatial transcriptome data from Step 1, the proportion of cell types at each spatial site (Spot) is predicted using the RCTD deconvolution algorithm; the difference in tumor cell purity between the drug-resistant and drug-sensitive groups in space is analyzed, and the candidate drug-resistant target genes screened in Step 6 are mapped to the spatial map to verify their enrichment in the drug-resistant lesion region; at the same time, the expression abundance of candidate transcription factors is extracted, mapped to the physical coordinates of tissue sections, and the spatial expression profile is reconstructed.

[0035] The results showed that CEBPB and SOX15 exhibited a wider expression trend in the drug-resistant tumor tissues. Figure 11 Furthermore, KRT15 and SCEL were specifically enriched in the tumor cell regions of drug-resistant lesions, which verified the correlation between target genes and chemotherapy resistance from a spatial in situ level, and finally output confirmed chemotherapy resistance targets.

[0036] Example 2 Embodiment 2 of the present invention provides a system corresponding to Embodiment 1 above, including a memory, a processor, and a computer program stored in the memory; the processor executes the computer program in the memory to implement the steps of the method of Embodiment 1 above.

[0037] In some implementations, the memory may be high-speed random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device.

[0038] In other implementations, the processor can be any type of general-purpose processor, such as a central processing unit (CPU) or a digital signal processor (DSP), and there is no limitation here.

[0039] Example 3 Embodiment 3 of the present invention provides a computer-readable storage medium corresponding to Embodiment 1 above, on which a computer program / instructions are stored. When the computer program / instructions are executed by a processor, they implement the steps of the method of Embodiment 1 above.

[0040] A computer-readable storage medium can be a tangible device that holds and stores instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof.

[0041] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0042] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0043] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0044] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0045] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for predicting genes related to chemotherapy resistance in ovarian cancer, characterized in that, Includes the following steps: S1. Obtain TCGA-OV public cohort transcriptome data, perform differential expression analysis on the chemotherapy-sensitive group and chemotherapy-resistant group in the TCGA-OV public cohort transcriptome data to obtain the TCGA-OV differential gene set; download GSE154600, GSE184880, and GSE211956 datasets from the GEO public database, obtain single-cell transcriptome data and spatial transcriptome data of the ovarian cancer chemotherapy-sensitive group and chemotherapy-resistant group, and construct single-cell gene expression matrix and spatial transcriptome expression matrix containing spatial location information, respectively. S2. Extract epithelial cells from single-cell transcriptome data, use the inferCNV algorithm to infer copy number variations, and identify malignant epithelial cells; Subpopulation subgroups and differential expression analysis of malignant epithelial cells were performed to identify characteristic subpopulations of drug-resistant groups and to construct a single-cell drug resistance differential gene set; S3. Using the CellphoneDB receptor-ligand database, calculate the interaction strength between the malignant epithelial cells and other cell populations; compare the calculation results of the chemotherapy-sensitive group and the chemotherapy-resistant group, extract cell communication pathways whose mean communication interaction value is upregulated in the chemotherapy-resistant samples and whose permutation test statistical probability value indicating inter-group communication differences is less than a first set value, and construct a microenvironmental drug resistance communication network; take the intersection of the TCGA-OV differential gene set and the single-cell drug resistance differential gene set to obtain common genes, and use the Kaplan-Meier algorithm to combine the common genes with overall survival and progression-free interval prognostic data for survival analysis, screen out common genes that are statistically associated with prognosis and whose log-rank test probability value of survival prognostic differences is less than a second set value, and use these common genes as candidate drug resistance target genes to predict and characterize the microenvironmental immunosuppressive features of ovarian cancer chemotherapy resistance status.

2. The method for predicting ovarian cancer chemotherapy resistance-related genes according to claim 1, characterized in that, The specific process for calculating the interaction strength between the malignant epithelial cells and other cell populations includes: obtaining receptor-ligand pairs based on the CellphoneDB receptor-ligand database; extracting the average expression levels of the ligands and receptors in the malignant epithelial cells and other cell populations; using the mean of the ligand expression levels and receptor expression levels between the malignant epithelial cells and other cell populations as the interaction strength of a specific receptor-ligand pair between groups; calculating the probability value of the interaction using a permutation test; and constructing a random distribution by randomly shuffling the cell type labels of the malignant epithelial cells and other cell populations, and repeatedly calculating the interaction strength to obtain the statistical probability value of the interaction of the specific receptor-ligand pair.

3. The method for predicting ovarian cancer chemotherapy resistance-related genes according to claim 1, characterized in that, Also includes: S4. Based on the spatial transcriptome sequencing, predict the proportion of cell types at each spatial site using a deconvolution algorithm; The study analyzes the spatial differences in tumor cell purity between chemotherapy-sensitive and chemotherapy-resistant groups, maps the candidate drug-resistant target genes to a spatial map, verifies the enrichment of the candidate drug-resistant target genes in drug-resistant lesions, and outputs confirmed chemotherapy-resistant target genes.

4. The method for predicting ovarian cancer chemotherapy resistance-related genes according to claim 3, characterized in that, The confirmed chemotherapy resistance target genes are KRT15 and SCEL.

5. The method for predicting ovarian cancer chemotherapy resistance-related genes according to claim 1, characterized in that, Also includes: Based on the spatial transcriptome expression matrix containing spatial location information, the expression abundance of candidate transcription factors is extracted, and the expression abundance is mapped to the corresponding physical coordinates of tissue slices for spatial expression profile reconstruction and analysis to obtain transcription factors with broad expression trends in drug-resistant samples.

6. The method for predicting ovarian cancer chemotherapy resistance-related genes according to claim 5, characterized in that, The transcription factors with a wide expression trend in the drug-resistant samples were CEBPB and SOX15.

7. The method for predicting ovarian cancer chemotherapy resistance-related genes according to any one of claims 1 to 6, characterized in that, Both the first setting value and the second setting value are 0.

05.

8. The method for predicting ovarian cancer chemotherapy resistance-related genes according to any one of claims 1 to 6, characterized in that, Also includes: Based on the single-cell gene expression matrix, the UMAP algorithm was used for dimensionality reduction visualization and unsupervised clustering, identifying multiple cell subpopulations. These cell subpopulations include B cells, endothelial cells, epithelial cells, fibroblasts, myeloid cells, NK cells, pericytes, plasma cells, stromal cells, and T cells. Based on the single-cell gene expression matrix, T cell and NK cell subsets in the immune microenvironment were extracted and subdivided. The differences in cell subset abundance between groups were analyzed. The evolutionary trajectory of T cells was analyzed through pseudo-time series analysis, and drug resistance-related immune exhaustion subsets were identified.

9. A predictive system for ovarian cancer chemotherapy resistance-related genes, comprising a memory, a processor, and a computer program stored in the memory; characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

Images