Prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism

Through liver cancer gene classification and energy metabolism analysis, an accurate prognostic analysis model was established, which solved the problem of inaccurate prognosis analysis of liver cancer in the existing technology, provided guidance for personalized treatment, and revealed the energy metabolism pattern and microenvironment of liver cancer.

CN114822688BActive Publication Date: 2024-11-19THE FIRST AFFILIATED HOSPITAL ZHEJIANG UNIV COLLEGE OF MEDICINE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210291850.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-24
Publication Date
2024-11-19
Estimated Expiration
2042-03-24

AI Technical Summary

Technical Problem

The existing prognosis analysis methods for liver cancer cannot accurately determine the prognosis of patients, and the metabolic heterogeneity of liver cancer leads to poor treatment effects, making it difficult to guide personalized treatment.

Method used

Using prognostic analysis methods based on primary liver cancer gene classification and liver cancer tissue energy metabolism, by obtaining clinical phenotype data, expression profile data, gene CNV mutation data and SNV mutation data, GSVA, ConsensusClusterPlus, gistic2, mutect2, maftools and other software Components, calculate metabolic pathway sample scores, conduct consistent clustering and differential analysis, and establish a prognostic analysis model for primary liver cancer.

Benefits of technology

The primary liver cancer gene classifier was constructed, and the molecular subtype of liver cancer was determined. The difference in metabolic pathway expression, gene mutation differences, clinical manifestations and immune differences were analyzed. An accurate prognosis analysis model was established to predict the prognosis of liver cancer. And reveals the energy metabolism pattern and microenvironment of liver cancer.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114822688B_ABST
    Figure CN114822688B_ABST
Patent Text Reader

Abstract

The present invention discloses a prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism, clinical phenotype data, expression spectrum data, gene CNV and SNV mutation data of liver cancer samples are obtained; sample scores of four metabolic pathways of tumor samples are calculated to determine several optimal classifications; the differences in expression of metabolic pathways, gene mutations, clinical phenotypes, and immune characteristics of several optimal classifications are analyzed, and then the results are comprehensively analyzed to establish a primary liver cancer prognostic analysis model. The advantage of the present invention is that a primary liver cancer gene classifier is constructed through big data analysis technology, the molecular subtypes of primary liver cancer are determined, and the differences in metabolic pathway expression, gene mutations, clinical manifestations, and immune differences of the molecular subtypes of primary liver cancer are analyzed, thereby establishing a primary liver cancer prognostic analysis model for accurately predicting the prognosis of primary liver cancer, and providing a research basis for revealing the energy metabolism pattern and tumor microenvironment of primary liver cancer.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the research field of liver cancer, and in particular to a prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism. Background Art

[0002] In the past decade, metabolic reprogramming has been considered one of the ten major characteristics of cancer, that is, carcinogenic tumors have a common phenotype that can effectively produce the energy and macromolecules required for their own metabolism, allowing tumor cells to grow uncontrollably. The latest research shows that cancer cell metabolism between and within tumors is plastic and flexible, resulting in poor treatment effects. At the same time, studies have also found that human cells have significant differences in the metabolic characteristics of major energy substances such as glucose, fatty acids, and glutamine. This difference is related to a series of factors such as genes, nutrients, and oxygen acquisition. Therefore, studying the main metabolic characteristics of carcinogenic tumor cells is helpful to reveal the factors that affect the uncontrolled growth of carcinogenic tumor cells, guide cancer treatment, and predict treatment response and prognosis.

[0003] Liver cancer is a refractory malignant tumor with high heterogeneity and the highest mortality rate. Its mortality rate ranks second among tumors, second only to lung cancer. There are many treatments for liver cancer, but the five-year survival rate after treatment is extremely low, only about 12%. This is most likely related to the metabolic heterogeneity of liver cancer. At present, clinically, the prognosis of patients is mainly based on the staging of liver cancer, but this method cannot accurately determine the prognosis of liver cancer patients. Therefore, the present application provides a prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism. Summary of the invention

[0004] The present invention aims to provide a prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism of the present invention comprises the following contents:

[0007] S1, obtain clinical phenotype data, expression profile data, gene CNV mutation data and SNV mutation data of liver cancer samples;

[0008] S2, using the ssGSEA method of the GSVA package, the sample scores of the four metabolic pathways of the tumor samples were calculated;

[0009] S3, using the ConsensusClusterPlus package to perform consensus clustering on the sample scores of the four metabolic pathways;

[0010] S4, calculating the consistency matrix and the consistency cumulative distribution function to determine several optimal classifications;

[0011] S5, analysis of the differences in gene expression among the best classes in the four metabolic pathways;

[0012] S6, analyzing the differences between the CNV mutation data and the SNV mutation data of the gene among several optimal classifications;

[0013] S7, analyzing the differences in clinical phenotype data between several optimal classifications;

[0014] S8, analysis of the differences in immune characteristics among several best classifications;

[0015] S9. Based on the classification and differential analysis results of the above steps, a prognostic analysis model for primary liver cancer is established.

[0016] Furthermore, in the clinical phenotype data of liver cancer, samples lacking survival time and survival status are removed, and the survival time of all patients is ensured to be greater than 0;

[0017] The gene CNV mutation data is of Masked Copy Number Segment type and integrated by gistic2 software, with the confidence level set to 0.9, and hg38 used as the reference genome;

[0018] The SNV mutation data were calculated using the mutect2 software.

[0019] Furthermore, the four metabolic pathways are glycolysis, PPP, FAO and glutaminolysis.

[0020] Furthermore, the consensus clustering described in step S3 used the PAM algorithm, with euclidean as the metric distance, and performed 500 bootstraps, each of which included 80% of the total tumor samples.

[0021] Furthermore, the consistency cumulative distribution function in step S4 is CDF.

[0022] Furthermore, the number of optimal classifications in step S4 is set to any integer between 2 and 10.

[0023] Furthermore, in step S6, the maftools package is used to analyze the SNV mutation data.

[0024] Furthermore, in step S7, the clinical phenotype data are analyzed using a chi-square test.

[0025] Furthermore, in step S8, the cell score of each liver cancer sample was calculated by the MCPcounter.estimate function of the MCPcounter package, and the differences in immune characteristics were analyzed by the wilcox.test test.

[0026] The advantage of the present invention is that a primary liver cancer gene classifier is constructed through big data analysis technology, the molecular subtypes of primary liver cancer are determined, and the metabolic pathway expression differences, gene mutation differences, clinical manifestation differences and immune differences of the molecular subtypes of primary liver cancer are analyzed, thereby establishing a primary liver cancer prognosis analysis model for accurately predicting the prognosis of primary liver cancer and providing a research basis for revealing the energy metabolism pattern and tumor microenvironment of primary liver cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flow chart of the method of the present invention.

[0028] Figure 2 It is a schematic diagram of the clustering result of the method of the present invention.

[0029] Figure 3 Schematic diagram of two subtypes of liver cancer obtained in the method of the present invention.

[0030] Figure 4 This is a PCA analysis result diagram of liver cancer samples using four metabolic pathways in the embodiment of the present invention.

[0031] Figure 5 It is a schematic diagram of comparing the differences in gene expression between several best classifications in four metabolic pathways based on the TCGA dataset in the method described in the present invention.

[0032] Figure 6 It is a schematic diagram of the difference comparison of gene expression between several best classifications in four metabolic pathways based on the GSE data set described in the present invention.

[0033] Figure 7 It is a schematic diagram of comparing gene CNV mutation data based on TCGA dataset in the method of the present invention.

[0034] Figure 8 It is a waterfall plot of the top 15 genes with the most SNV mutations based on the TCGA dataset in the method described in the present invention.

[0035] Fig. 9 It is a schematic diagram of the comparison of the differences in immune characteristics between the best categories based on the TCGA data set in the method of the present invention.

[0036] Fig.10 It is a schematic diagram of comparing the differences in immune features between the best categories based on the GSE data set in the method of the present invention. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] like Figure 1 As shown, the prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism of the present invention includes the following contents:

[0039] S1, obtain clinical phenotype data, expression profile data, gene CNV (copy number variation) mutation data and SNV (single-nucleotide variant) mutation data of liver cancer samples;

[0040] The liver cancer sample data of this embodiment downloads the clinical phenotype data of liver cancer from the TCGA (The Cancer Genome Atlas) database, and removes samples that lack survival time and survival status to ensure that the survival time of all patients is greater than 0, and finally obtains 360 tumor samples.

[0041] At the same time, this embodiment also directly downloads the expression spectrum data in the TCGA database, the gene CNV mutation data about liver cancer in the TCGA database, and the gene SNV mutation data calculated by mutect2 (a commonly used gene mutation detection software) in the TCGA database.

[0042] In order to enrich liver cancer samples, this example also downloaded a set of chip data from GEO (Gene Expression Omnibus, a gene expression database created and maintained by the National Center for Biotechnology Information (NCBI) of the United States), removed samples with no clinical follow-up information and a total survival time of 0, and finally obtained 242 tumor samples and 10,501 genes.

[0043] In this example, the molecular signature database (MSigDB) was used to download pathway genes related to four metabolic pathways: glycolysis, PPP (pentose phosphate pathway), FAO (fatty acid oxidation) and glutaminolysis.

[0044] S2, using the ssGSEA method of the GSVA (Gene Set Variation Analysis, known as gene set variation analysis, a non-parametric unsupervised analysis method, mainly used to evaluate the gene set enrichment results of the microarray nuclear transcriptome) package to calculate the sample scores of the four metabolic pathways of the tumor sample;

[0045] S3, the ConsensusClusterPlus package (for consensus clustering, classifying samples according to gene expression) was used to perform consensus clustering on the sample scores of the four metabolic pathways; the consensus clustering used the PAM algorithm, euclidean as the metric distance, and 500 bootstraps (re-sampling method) processes, each of which included 80% of the total tumor samples.

[0046] S4, calculating the consistency matrix and the consistency cumulative distribution function CDF to determine a number of optimal classifications; the optimal number of classifications is set to any integer between 2 and 10.

[0047] In this embodiment, 360 liver cancer samples obtained from the TCGA database are clustered by consensus clustering (ConsensusClusterPlus), and the best cluster is determined according to the consistency cumulative distribution function CDF. By observing the CDF Deltaarea (relative change of the area under the consistency cumulative distribution function curve) curve, it can be seen that when the best classification number is 2, a relatively stable clustering result is obtained, such as Figure 2 As shown in A and B. Thus, two subtypes of liver cancer can be identified, such as Figure 3 shown.

[0048] In order to verify the accuracy of the classification results, the inventors of the present application conducted PCA (principal component analysis) analysis on liver cancer samples based on four metabolic pathways, such as Figure 4 As shown in A and B in FIG, this analysis further supports the heterogeneity of metabolism among tumors and finds that the classification of liver cancer into two subtypes is the most reliable classification, thereby verifying the accuracy and credibility of the method of liver cancer gene classification based on this application.

[0049] S5, such as Figure 5As shown in A, B, and C, by comparing the differences in the genes of the four metabolic pathways of the tumor samples of the two liver cancer subtypes, and drawing the sample score heat maps of the four metabolic pathways of the tumor samples of the two liver cancer subtypes, the differences in the gene expression between the four best classifications were analyzed; The inventors of the present application found that in the sample scores of the three metabolic pathways of Glutaminolysis, FAO, and PPP, the subtype C2 score was higher than the subtype C1. This conclusion was also confirmed in the GSE dataset, such as Figure 6 As shown in A, B, and C.

[0050] S6, analyzing the differences between the CNV mutation data and the SNV mutation data of the genes among several optimal classifications; integrating the CNV (copy number variation) mutation data of genes related to liver cancer downloaded from the TCGA database using gistic2 (copy number variation integration software), setting the confidence level to 0.9, and using hg38 as the reference genome for analysis. The analysis results are as follows: Figure 7 As shown, there are some differences in gene CNV mutations between the two liver cancer subtypes. At the same time, the SNV mutation data were analyzed using the maftools package (the mutation annotation format (MAF) is widely accepted and used to store detected somatic variants. As long as the data is in MAF format, this package will try to efficiently summarize, analyze, annotate and visualize MAF files from TCGA sources or any internal research) as shown in the figure. Figure 8 As shown, a waterfall plot of the top 15 genes with the most SNV mutations in two liver cancer subtypes is drawn.

[0051] S7, the chi-square test was used to analyze the differences in clinical phenotype data between several optimal classifications; the results showed that in the TCGA dataset, there were significant differences in T stage, stage, and grade between the two subtypes, and in the GSE dataset, there were significant differences in stage and age between the two subtypes.

[0052] S8, analyze the differences in immune characteristics among several optimal classifications; immune characteristics mainly refer to the abundance of 10 major immune cells such as T cells, NK cells, and B cells.

[0053] Specifically, the cell score of each liver cancer sample in the TCGA dataset was calculated by the MCPcounter (microenvironment cell populations-counter, which allows the relative abundance of eight immune and two non-immune stromal cell populations in tumor tissues to be quantified based on transcriptional data) package MCPcounter.estimate function, and the difference in immune characteristics of the two liver cancer subtypes was analyzed by wilcox.test (non-parametric test); it was found that the two liver cancer subtypes had significant differences in T cells, CD8T cells, etc., such as Fig. 9 As shown in the figure. The results showed that T cells, CD8T cells, cytotoxic lymphocytes, B cell lineage, monocyte lineage and myeloid dendritic cells in subtype 1 were more abundant than those in subtype 2, suggesting that there was a higher abundance of immune cell infiltration in subtype 1. This conclusion was also confirmed in the GSE dataset, as shown in the figure. Fig.10 shown.

[0054] S9. Based on the classification and differential analysis results of the above steps, a prognostic analysis model for primary liver cancer is established.

Claims

1. A prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism, characterized in that: Includes the following: S1, obtain clinical phenotype data, expression profile data, gene CNV mutation data and SNV mutation data of liver cancer samples; S2, the ssGSEA method of the GSVA package was used to calculate the sample scores of the four metabolic pathways of tumor samples; S3, using the ConsensusClusterPlus package to perform consensus clustering on the sample scores of the four metabolic pathways; S4, calculating the consistency matrix and the consistency cumulative distribution function to determine several optimal classifications; S5, analysis of the differences in gene expression among the best classes in the four metabolic pathways; S6, analyzing the differences between the CNV mutation data and the SNV mutation data of the gene among several optimal classifications; S7, analyzing the differences in clinical phenotype data between several optimal classifications; S8, analysis of the differences in immune characteristics among several best classifications; S9. Based on the classification and differential analysis results of the above steps, a prognostic analysis model for primary liver cancer is established.

2. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: In the clinical phenotype data of liver cancer, samples lacking survival time and survival status are removed, and the survival time of all patients is ensured to be greater than 0; The gene CNV mutation data were integrated using gistic2 software, with the confidence level set to 0.9, and hg38 used as the reference genome; The SNV mutation data were calculated using the mutect2 software.

3. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: The four metabolic pathways are glycolysis, PPP, FAO and glutaminolysis.

4. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: The consensus clustering described in step S3 used the PAM algorithm, euclidean as the metric distance, and performed 500 bootstraps, each of which included 80% of the total tumor samples.

5. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: The consistent cumulative distribution function described in step S4 is CDF.

6. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: The number of optimal classifications in step S4 is set to any integer between 2 and 10.

7. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: In step S6, the maftools package is used to analyze the SNV mutation data.

8. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: In step S7, the clinical phenotype data were analyzed using the chi-square test.

9. The prognostic analysis method based on primary liver cancer gene classification and liver cancer tissue energy metabolism according to claim 1, characterized in that: In step S8, the cell score of each liver cancer sample was calculated using the MCPcounter.estimate function of the MCPcounter package, and the differences in immune characteristics were analyzed using the wilcox.test test.