Blood diagnosis marker for identifying good prognosis prostate cancer and metastatic castration-resistant prostate cancer and application of blood diagnosis marker

Abstract

The invention provides a blood diagnosis marker for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer and application of the blood diagnosis marker, and relates to the technical field of prostate cancer risk assessment. The blood diagnosis markers are NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7 and ZSCAN2, and the blood diagnosis markers are used for diagnosis of the blood. The marker can be used for preparing a kit and a risk assessment chip for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer. According to the invention, the defects in the prior art are overcome, and a good and reliable basis is provided for identifying the prostate cancer with good prognosis and the metastatic castration-resistant prostate cancer by determining the blood diagnosis marker.

Description

Technical Field

[0001] The present invention relates to the technical field of prostate cancer risk assessment, and particularly relates to a blood diagnostic marker for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer and its application. Background Art

[0002] Prostate cancer is a highly heterogeneous disease, and the conditions of patients vary significantly. Some patients are diagnosed at an early stage and may not require treatment or may be cured through radical treatment; while other patients may relapse after initial treatment and eventually develop into metastatic castration-resistant prostate cancer (mCRPC) and die therefrom. Therefore, identifying and validating multi-purpose blood or urine biomarker tests is crucial for the individualized management of prostate cancer. Such tests are relatively non-invasive, repeatable, and easy to implement in clinical practice.

[0003] Although serum prostate-specific antigen (PSA) has been widely studied in the management of prostate cancer, its ability to serve as a reliable intermediate endpoint for overall survival has been questioned. In recent years, the development of high-throughput technologies has made it possible to identify other useful tissue and body fluid biomarkers. Studies have shown that there are significant differences in the whole blood gene expression profiles between patients with invasive metastatic castration-resistant prostate cancer and patients with low-grade, low-disease-burden prostate cancer, and such expression profiles have clinical application value. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a blood diagnostic marker for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer and its application, and provides a good and reliable basis for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer by determining the blood diagnostic marker.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions:

[0006] A blood diagnostic marker for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer, and the marker is NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7, and ZSCAN2.

[0007] Preferably, among the blood diagnostic markers, NCKAP1, ARMC8, LRRN4, CDKN2D, and CLEC1B are up-regulated genes, and ERCC1, ZBTB46, PAQR7, PAQR8, NOG, LRRC47, STARD7, and ZSCAN2 are down-regulated genes.

[0008] Preferably, the determination of the biomarker includes the following steps:

[0009] S1. Obtain three datasets of transcriptome data of prostate cancer patients;

[0010] S2. After normalizing the datasets, perform differential gene analysis to obtain up-regulated genes and down-regulated genes, which are the differential genes;

[0011] S3. Use the clusterProfiler package to perform GO and KEGG enrichment analysis on the differential genes;

[0012] S4. Use weighted gene co-expression network analysis to obtain key genes;

[0013] S5. Perform intersection analysis on the key genes and the differential genes to obtain key differential genes;

[0014] S6. Use the random forest algorithm to construct a diagnostic model for the key differential genes, and determine that the differential genes in this diagnostic model are NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7, and ZSCAN2.

[0015] Preferably, all three datasets in S1 include transcriptome data of patients with non-metastatic castration-resistant prostate cancer.

[0016] Preferably, the method of differential gene analysis in step S2 is to use the limma package to perform differential gene analysis on the datasets, and the control screening criteria are P < 0.05 and |logFC| > 0.

[0017] The above blood diagnostic biomarker can be used to prepare a kit for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer, as well as a risk assessment chip.

[0018] The present invention provides a blood diagnostic biomarker for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer and its application. Compared with the prior art, the advantages are as follows:

[0019] In the present invention, differential analysis and weighted gene co-expression network analysis are performed on the dataset using the limma package. Subsequently, key differential genes are obtained based on the intersection, and then screened through different machine learning algorithms. The performance of the diagnostic model is optimized through 113 algorithm combinations. The diagnostic effect of the model is evaluated using the cross-validation method, and the performance of each model is compared through the area under the ROC curve (AUC). Finally, the algorithm combination with the best performance is selected to construct the diagnostic model of differential genes, and the most important 13 differential genes in this model are determined to be NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7, ZSCAN2, which have extremely high diagnostic efficacy and high potential diagnostic value in prostate cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Schematic diagram for screening differential genes in peripheral blood of prostate cancer with good prognosis and metastatic castration-resistant prostate cancer (mCRPC), where A is before the normalization process of the dataset; B is after the normalization process of the dataset; on the left of C is the heat map of the screened up-regulated genes and down-regulated genes; on the right of C is the volcano plot of the screened up-regulated genes and down-regulated genes (red dots represent up-regulated genes, and blue dots represent down-regulated genes); D is the schematic diagram of GO enrichment analysis;

[0021] Figure 2 Schematic diagram of the clustering tree of weighted gene co-expression network analysis in Example 1 of the present invention;

[0022] Figure 3 Schematic diagram of the soft threshold selection of weighted gene co-expression network analysis in Example 1 of the present invention, where the left side is scale independence and the right side is average connectivity;

[0023] Figure 4 Schematic diagram of the clustering trend of weighted gene co-expression network analysis in Example 1 of the present invention;

[0024] Figure 5 Schematic diagram of the module correlation of weighted gene co-expression network analysis in Example 1 of the present invention;

[0025] Figure 6 Schematic diagram of the correlation between the module and phenotype of weighted gene co-expression network analysis in Example 1 of the present invention;

[0026] Figure 7 Schematic diagram of the intersection analysis of key genes and differential genes of weighted gene co-expression network analysis in Example 1 of the present invention;

[0027] Figure 8Schematic diagram for screening the optimal diagnostic model in Embodiment 1 of the present invention, where A is a schematic diagram of the combination of 113 algorithms; B-E are schematic diagrams of the AUC curve analysis of the training set and different validation sets; F-I are schematic diagrams of the matrix analysis of the training set and different validation sets; J is a schematic diagram of the AUC curve analysis of 13 most important differential genes; K is a volcano plot of differential genes. Detailed implementation manners

[0028] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0029] Embodiment 1:

[0030] Determination of blood diagnostic markers for identifying prostate cancer with good prognosis and metastatic castration-resistant prostate cancer:

[0031] 1. Data collection:

[0032] Obtain the transcriptome data sets GSE37119, GSE66532, and GSE248619 of prostate cancer patients from the GEO (https: / / www.ncbi.nlm.nih.gov / geo / ) database. All three data sets contain the transcriptome data of patients with metastatic castration-resistant prostate cancer (mCRPC). GSE37119 contains 39 prostate cancer patients with good prognosis and 68 mCRPC patients, GSE66532 contains 43 mCRPC patients, and GSE248619 contains 95 mCRPC patients;

[0033] The data sets GSE248619, GSE37119, and GSE66532 were standardized. See Figure 1 A for details before standardization. However, after standardization, the sample data is basically within the same range, indicating that the data quality is good and the normalization effect is significant ( Figure 1 B);

[0034] The limma package was used to perform differential gene analysis on the data sets; according to the screening criteria (P < 0.05 and |logFC| > 0), a total of 232 up-regulated genes and 143 down-regulated genes were screened out. See Figure 1 C for details. The heat map and volcano plot clearly show the distribution of these differential genes;

[0035] The clusterProfiler package was used to perform GO and KEGG enrichment analyses on the screened differential genes. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were used to perform enrichment analyses on the screened differential genes. GO enrichment analysis evaluates the potential functions of genes from three aspects: biological process (BP), cellular component (CC), and molecular function (MF); KEGG pathway enrichment analysis reveals the signaling pathways and their biological functions involved in differential genes; such as Figure 1 D, where GO enrichment analysis revealed that the differential genes were mainly concentrated in multiple pathways such as synaptic pruning, cell junction disassembly, and endocytic vesicle lumen.

[0036] 2. Weighted gene co-expression network analysis (WGCNA):

[0037] To further screen for key genes related to mCRPC, weighted gene co-expression network analysis (WGCNA) was performed. The genes were divided into 9 modules by the dynamic tree cutting method( Figures 2 - 5 ). Correlation analysis showed that the MEbrown and MEyellow modules were significantly correlated with the clinical characteristics of prostate cancer. Intersection analysis of the key genes in this module with the differential genes yielded 43 key differential genes( Figure 7 ).

[0038] 3. After screening out the key differential genes, referring to Figure 8 A, combined with 10 machine learning algorithms to screen the differential genes identified by WGCNA. By evaluating 113 algorithm combinations (including single algorithms, ensemble methods, and multi-model combinations), the performance of the diagnostic model was optimized, specifically referring to the algorithm combinations in the article "Multi-omics identification of an immunogenic cell death-related signature for clear cell renal cell carcinoma in the context of 3P medicine and based on a 101-combination machine learning computational framework";

[0039] Samples from the GSE16419 and GSE37119 datasets were used as the training set, and GSE16419, GSE37119, and GSE203024 were used as the validation set. 113 prediction models were adopted, and the C-index (CI) of the training set and the validation set was calculated. By evaluating the performance of each algorithm, the random forest (RF) algorithm was finally selected to construct a diagnostic model with the best overall performance ( Figure 8 A);

[0040] This model was based on 13 of the most important differentially expressed genes (NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7, ZSCAN2); the AUC values in the training set and the validation set were 0.997, 1.000, 0.983, and 1.000 respectively), showing extremely high diagnostic efficacy ( Figure 8 B-K);

[0041] The results of the AUC curve analysis showed that the diagnostic efficacy of each gene was above 0.6, further verifying the potential diagnostic value of these genes in mCRPC ( Figure 8 J). To verify the expression levels of the genes in the diagnostic model, a volcano plot was drawn to show the expression differences of 5 genes ( Figure 8 K).

[0042] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A blood diagnostic marker for identifying good prognosis prostate cancer and metastatic castration-resistant prostate cancer, characterized in that: The markers are NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7 and ZSCAN2.

2. The blood diagnostic marker according to claim 1, characterized in that: Among the blood diagnostic markers, NCKAP1, ARMC8, LRRN4, CDKN2D, and CLEC1B are up-regulated genes, and ERCC1, ZBTB46, PAQR7, PAQR8, NOG, LRRC47, STARD7, and ZSCAN2 are down-regulated genes.

3. The blood diagnostic marker according to claim 1, characterized in that The determination of the marker comprises the following steps: S1. Obtain a dataset of transcriptome data from prostate cancer patients; S2. After the data set is standardized, differential gene analysis is performed to obtain up-regulated genes and down-regulated genes as differential genes; S3, GO and KEGG enrichment analysis of differentially expressed genes was performed using the clusterProfiler package; S4. Use weighted gene co-expression network analysis to obtain key genes; S5. Perform intersection analysis on key genes and differential genes to obtain key differential genes; S6. The random forest algorithm was used to construct a diagnostic model for the key differentially expressed genes, and the differentially expressed genes in the diagnostic model were determined to be NCKAP1, ERCC1, ZBTB46, PAQR7, ARMC8, PAQR8, LRRN4, NOG, LRRC47, CDKN2D, CLEC1B, STARD7, and ZSCAN2.

4. The blood diagnostic marker according to claim 3, characterized in that: The three data sets in S1 all include transcriptome data of patients with advanced castration-resistant prostate cancer.

5. The blood diagnostic marker according to claim 3, characterized in that: The differential gene analysis method in step S2 is to use the limma package to perform differential gene analysis on the data set, and the control screening criteria are P<0.05 and |logFC|>0.

6. Use of the blood diagnostic marker as described in any one of claims 1 to 5 in the preparation of a kit for identifying prostate cancer with a good prognosis and metastatic castration-resistant prostate cancer, and a risk assessment chip.

Images