Abstract
Clear cell renal carcinoma has been reported in many research studies as a rather heterogeneous disease. Identification of different subtypes and their molecular characteristics can help in choosing a more effective treatment and predicting a response to it. In this study, using multi-omics clustering of RNA-Seq data of patients with clear cell renal carcinoma from TCGA. Specific genes were identified for the most aggressive ccRCC subtype associated with metastasis and a subtype associated with a more favorable course of the disease. Among them were genes associated with blood clotting (FGA, FGG) and genes associated with changes in the immune characteristics of a tumor (ENAM, IGFBP1, IL6). In addition, an association of hub genes of poor survival ccRCC subtype with the levels of infiltration of endothelial cells, hematopoietic stem cells, T cells NK and mast cells was revealed. It was shown that MFI2, CP, FGA, and FGG expression can predict the response to sunitinib, while the APOB, ENAM, IGFBP1, and MFI2 expression predict the response to nivolumab. The results obtained provide insight into the genetic characteristics underlying the aggressive subtype of ccRCC and may help develop new approaches to the treatment of this disease.
Similar content being viewed by others
Introduction
Clear cell renal carcinoma (ccRCC) is the most common subtype of renal cancer1. The 5-year survival of ccRCC patients is about 60–70%2. The available data show an increase in the number of new cases of ccRCC over the past decades2. Despite the extensive data accumulated recently, the identification of reliable genes that make it possible to predict and select therapy for ccRCC is still an urgent problem.
To date, abundant evidence suggests that ccRCC is a rather heterogeneous type of cancer3. In particular, subtypes of metastatic ccRCC have been identified based on their response to sunitinib4. Recently, different subtypes of ccRCC have been identified based on the activity of transcription factors5. These data show that within the histologic type, ccRCC group of samples can be distinguished by the general unique characteristics.
Common TNM classification of samples is often insufficient to prescribe a suitable targeted treatment6. It was shown that gene expression depends on the stages of ccRCC, and different stages of ccRCC are characterized by different aberrant regulated pathways, including the coagulation cascade7.
Early identification of the most aggressive tumors could significantly improve the treatment of ccRCC. The identification of key molecular changes specific to such malignant subtypes can help in the development of effective personalized therapies. Thus, the aim of this work is to reveal major subtypes of ccRCC tumors and to identify key genetic characteristics of these subtypes.
Results
Clustering of ccRCC samples revealed three subtypes with different survival rates
Clustering of ccRCC samples from the TCGA database was carried out by the k-means method with 100 replicates. For the clustering of samples, k = 3 was chosen, since it was previously reported that clustering stability increases from k = 2 to k = 3, but not at k > 3 for the ccRCC samples3. For analysis, I selected tumor samples for which no radiotherapy or additional pharmacotherapy was performed. As a result, 456 samples were divided into two large clusters (n = 299 and n = 126) and one smaller (n = 31).
Survival analysis revealed that Cluster 2 was associated with worse overall survival, while Cluster 1 was associated with better overall patient survival (Fig. 1b). At the same time, in Cluster 2, samples with late stages (III and IV) as well as samples with metastases prevail (Fig. 1c). While in Cluster 1, samples with stage I prevail (Fig. S1b). In addition, the available data suggests patients from Cluster 1 respond better to treatment than patients from Cluster 2 (Fig. S1c,d).
For gene clustering, I selected 2000 genes with the most highly variable expression. As a result, the genes were divided into 5 clusters (Fig. 1a). At the same time, the analysis of gene ontology shows a pronounced cluster associated with ion transport and pH maintenance, an immune response cluster, a cluster associated with an acute-phase response, a cluster associated with transmembrane organic anion transport, and a cluster associated with cell adhesion and organization of the extracellular matrix (Fig. S2). Thus, ccRCC subtypes with different survival rates and different responses to treatment were identified.
Identification of hub genes specific for the poor prognostic subtype
For the detected clusters, I identified differentially expressed genes using the rank sum test (Fig. 2a,b). The difference was considered significant at P < 0.05, and identified 310 genes for Cluster 2 and 315 genes for Cluster 1. Then, a network of protein–protein interactions (PPI) of Cluster 1 and Cluster 2 was built, taking into account changes in gene expression in ccRCC (Fig. 2c,d).
For Clusters 1 and 2, I calculated the top 10 hub genes (Fig. 3a,b). The Maximal Clique Centrality (MCC) algorithm was used, which was reported as the most effective method to find hub genes8. Clustering with k = 4 for the cluster with poorest survival revealed the same 10 hub genes: APOB, IGFBP1, CP, ENAM, MFI2, CHRDL1, IL6, TF, FGA, FGG. Interestingly, among the hub genes of Cluster 1 was the SERPINA5 gene, one of the most decreased in ccRCC (Fig. 2d). CP was also revealed as a hub gene for Cluster 2, which had one of the most increased expressions in ccRCC (Fig. 2c). The genes encoding the fibrinogen alpha chain (FGA) and fibrinogen gamma chain (FGG) have been identified as key for both clusters. Moreover, Cluster 2 is characterized by higher expression of these genes (Fig. 3). Gene set enrichment analysis revealed that Cluster 2 genes are associated with regulation of insulin-like growth factor, endoplasmic reticulum, and protein post-translational modifications (Fig. S3b). For Cluster 1, there is an association with fibrinolysis, hemopexin, and negative regulation of coagulation (Fig. S3a). Interestingly, for the expression of most of the detected genes, a clear bimodal or trimodal distribution was observed, which also indicates the presence of subpopulations in the sample under study (Fig. S4)9.
Hub genes of the PPI network of the aggressive subtype are associated with a poor prognosis of ccRCC
Survival analysis for hub genes revealed increased expression of 5 out of 10 genes in Cluster 2 is associated with poor survival (Fig. 4a). The same cannot be said about the Cluster 1 hub genes. Gene expression varies in Cluster 1 and Cluster 2 (Fig. 4b). If increased gene expression is associated with a poor prognosis, then its expression is increased in Cluster 2 and decreased in Cluster 1 (Fig. 4a). Conversely, if decreased expression is associated with poor prognosis, its expression is decreased in Cluster 2 and increased in Cluster 1 (Fig. 4b).
Hub genes of the ccRCC cluster with the worst survival predict the response to TKIs
To further check the association of Cluster 2 hub genes with poor survival in ccRCC in an independent sample, the GSE3538 dataset was used. Using multivariate Cox analysis, expression of a gene signature including MFI2, APOB, IGFBP1, CP, FGA, FGG, IL6, and APOB was shown to be associated with poor survival (P < 0.05, Fig. 5a). A statistically significant increase in expression in samples with a high risk was observed for the genes APOB, CP (P < 0.001), MFI2 (P < 0.01) and a decrease in expression in samples with a high risk for the FGG gene (P < 0.001) (Fig. 5b). In addition, the datasets E-MTAB-3218 and E-MTAB-3267 showed that hub genes expression of Cluster 2 are associated with a response to TKIs nivolumab and sunitinib (P < 0.05) (Fig. 5c,d). The highest AUC and the most significant ROC P-value was found for MFI2 with sunitinib (0.71, P < 0.001), and for CP (0.73, P < 0.001) with nivolumab.
Association of the aggressive subtype of ccRCC with the TME
According to the TIMER2.0 database, the expression of the Cluster 2 hub genes correlates with the infiltration level of a number of cells in the TME in KIRC. In particular, with endothelial cells, HSCs, T cells NK, and mast cells. For the ENAM gene, there is an opposite correlation in comparison with other genes; in particular, a negative correlation with T cells NK and a positive correlation with Neutrophils (P < 0.05) (Fig. 6A). For MFI2, there is the strongest inverse correlation with endothelial cells and HSCs (P < 0.05) (Fig. 6A). In addition, according to the survival analysis, there is a negative correlation of 10 hub genes of Cluster 2 with endothelial cells and HSCs (Fig. 6B,C). At the same time, I found a positive correlation in survival time for natural killer T cells (T cells NK) and mast cells (Fig. 6C,D). We also found a significant positive correlation of mutations in the SETD2 gene with the expression of the FGG (rs = 0.55, P < 0.001), FGA (rs = 0.78, P < 0.001), and IL6 (rs = 0.56, P < 0.001), and a negative correlation with the expression of ENAM (rs = -0.67, P < 0.001, Fig. 6F–I). The frequency of the SETD2 mutation in KIRC samples is about 12% (45/370).
Discussion
Using data from patients with ccRCC from TCGA, this study identified specific genes for subtypes of ccRCC with the most and least favorable survival rates. The data obtained indicate the heterogeneity of clear cell renal carcinoma and are consistent with previous studies indicating two distinct molecular subtypes of ccRCC with different survival rates—ccA and ccB10. Ten hub genes have been identified for genes that are specifically expressed in this aggressive ccRCC subtype. According to gene set enrichment analysis, these genes are involved in vital cellular functions, such as post-translational protein modifications, endoplasmic reticulum, and insulin growth factor receptor signaling (Fig. S3b).
The revealed 10 hub genes associated with Cluster 1 and Cluster 2 represent targets for the development of targeted therapy for the corresponding ccRCC subtype. The involvement of some of the identified hub genes in the development and progression of ccRCC has already been reported earlier. In particular, increased TF expression and its effect on metastasis in ccRCC have been reported before11. As well as IL6 has previously been reported as an early-stage immunologic prognostic factor for organ-confined ccRCC patients12. Recently, IGFBP1 has been shown to regulate monocytes and play a vital role in the immune system of ccRCC tumors and enhances their progression13. According to the received data, a high level of T cells NK at a high expression of the Cluster 2 hub genes corresponds to poor survival, which indicates enhanced immunosuppression of the TME in this subtype of ccRCC (Fig. 6D).
It has also recently been shown that ENAM can suppress cell proliferation in ccRCC and significantly increase cytotoxic cells, CD56 NK cells, NK cells, and CD8 + T cells in the immune TME14. In this work, a correlation was revealed between the hub genes of the ccRCC subtype with poor prognosis and the immune cell infiltration cells. Negative correlation of MFI2, FGA, and FGG with endothelial cells and hematopoietic cells, which contribute to angiogenesis15,16. Apparently, this explains the poorer response to TKIs in samples with higher expression of MFI2, CP, FGA, FGG, and APOB (Fig. 5c,d). The negative correlation between the expression of the hub genes of Cluster 2 with mast cells and better survival at a high level of mast cells (Fig. 6E) is probably explained by the ability of mast cells to release heparin17.
Interestingly, FGA and FGG have been identified as specific markers for both clusters. An increase in the expression of these two genes, as well as F2, a coding prothrombin, is specific to poor prognostic Cluster 2 (Fig. 4b). The role of fibrinogen as a marker of poor prognosis and metastasis for kidney cancer has been shown previously18. Fibrinogen has an advantage as a prognostic marker due to its easier availability19,20. The key genes for Cluster 1 are associated with fibrinolysis and regulation of coagulation.
It is known that accumulations of lipids are observed in the cytoplasm of ccRCC tumor cells21, and the importance of changes in lipid metabolism in the development of ccRCC was studied previously22. Therefore, it is interesting that among the hub genes of the malignant subtype ccRCC, the APOB gene was also identified, encoding apolipoprotein B, which is involved in lipid metabolism and the main protein component of lipoproteins23.
In this work, it was revealed that mutations in the SETD2 gene in ccRCC correlate with the expression of FGG, FGA, IL6, and ENAM, the hub-genes of Cluster 2 (Fig. 6F–I). Histone methyltransferase gene SETD2 is frequently mutated and is a suppressor gene in ccRCC24. Recently, SETD2 mutations have been shown to suppress autophagy in ccRCC25. In addition, mutations in SETD2 are known to be one cause of sunitinib resistance in renal cell cancer26.
From the data obtained, it follows that increased coagulation is one of the key features of the malignant subtype of ccRCC. Apparently, in ccRCC, the hematogenous route of metastasis predominates in the lungs, bones, and liver27. In addition, according to the NDEx Biological Network Repository, FGA, FGB, and FGG are associated with positive regulation of heterotypic cell-cell adhesion (P < 0.001)28. Previous studies reported changes in the expression of CP, FGA, and FGG in metastatic ccRCC samples29. Therefore, the combined use of anticoagulants with conventional treatment can be useful in therapy of malignant forms of ccRCC. In particular, the safe use of sunitinib and dalteparin to treat metastatic forms of ccRCC has already been described30. There is also evidence that heparin, commonly used to treat thromboembolic events in cancer patients, contributes to their survival and has anti-metastatic activity31.
Thus, the subtypes of ccRCC associated with the least and most favorable prognosis and with better response to treatment were studied (Fig. S1). Hub genes and the processes in which they are involved were identified for each of the subtypes. The identified genes specific to both subtypes may be of interest for further study and development of new approaches to the treatment of different ccRCC subtypes.
Methods
Public data sources
Normalized data from 606 samples, their phenotypes including and clinical pathological characteristics and curated survival data, were downloaded from The Cancer Genome Atlas (TCGA) using the UCSC Xena Functional Genomics Browser32. The downloaded data included the RNA-Seq dataset (Illumina HiSeq platform, gene-level RSEM-normalized, log2 transformed). The GSE3538 database was taken as an independent verification sample, which includes microarray data from 177 conventional renal cell carcinoma samples. To analyze the response to tyrosine kinase inhibitors, I downloaded the E-MTAB-3218 database, including sequencing data of 114 biopsies from metastatic clear cell renal carcinoma before or after treatment with nivolumab, and E-MTAB-3267, including sequencing data from 59 patients with metastatic ccRCC treated with sunitinib.
Clustering of ccRCC samples
Analysis of TCGA (The Cancer Genome Atlas) data performed using the Profiler of Multi-Omics data (PROMO) software33. This software, developed by Ron Shamir's lab, allows analyzing large genomic cancer datasets and their associated clinical information. The Cancer Genome Atlas Kidney Renal Clear Cell Carcinoma (TCGA-KIRC) data and their clinical characteristics were loaded into PROMO with further pre-processing and preparation for subsequent analysis. The integrity of the data was checked, and their distribution was characterized. Unsupervised clustering was carried out using the k-means method at k = 3 for samples and k = 5 for genes. The program algorithm was run 100 times. Identification of differentially expressed genes for individual clusters was carried out using the rank sum test. The change in expression at P < 0.05 was considered statistically significant. The predictive value of the identified sample clusters was studied by survival analysis using the Kaplan–Meier method and a log-rank test.
Survival and ROC analysis
Survival analysis was performed using the Gene Expression Profile Interactive Analysis GEPIA tool34. Survival plots were constructed using the Kaplan–Meier method with cut-off by medians. The difference in the samples of low and high expression is considered through the logrank test. P-values < 0.05 were considered statistically significant. To visualize differentially expressed genes in ccRCC clusters and construct the histogram of gene expression distribution, I used the ImageGP platform (http://www.ehbio.com/ImageGP). The construction of survival curves by the Kaplan–Meier method for the control dataset GSE3538, including RNA-Seq data of 177 ccRCC patients, was carried out using the SurvExpress tool, which allows analyzing the prognostic value of multiple genes as a signature35. The ROC plotter tool36 was used to analyze the predictive ability of genes for response to TKIs. Multivariate Cox regression was used to construct the prognostic risk models.
Building networks of protein–protein interactions
The construction of networks of protein–protein interactions carried out using the Cytoscape program. The interaction data was downloaded from the STRING database of known and predicted protein–protein interactions, version 1137. Differentially expressed genes in TCGA samples were identified using the GEPIA tool with ANOVA test and q-value cutoff = 0.05. The identification of hub genes was carried out using the CytoHubba plugin with the topological analysis method Maximal Clique Centrality (MCC)8. Gene set enrichment analysis (GSEA) was performed for identification of pathways associated with hub genes38. Visualization of networks of interactions was carried out using Cytoscape 3.7.1 software39.
Analysis of the correlation of gene expression with the TME
The Tumor Immune Estimation Resource (TIMER2.0) resource was used to analyze the relationship between gene expression and the level of immune-cell infiltration in the ccRCC TME40. TIMER2.0 allows us to estimate the immune infiltration levels for TCGA using six state-of-the-art algorithms, including TIMER, xCell, MCP-counter, CIBERSORT, EPIC, and quanTIseq. In addition, using TIMER2.0, I searched for the correlation of gene expression levels with the status of non-synonymous somatic mutations in the most frequently mutated genes in ccRCC.
Statement
TCGA and GEO belong to public databases. The patients involved in the database provided ethical approval. Users can download relevant data for free for research and publish relevant articles. This study is based on open-source data, so there are no ethical issues.
Data availability
The datasets analyzed during the current study are available from the corresponding author on reasonable request. All methods were carried out in accordance with relevant guidelines and regulations.
Abbreviations
- ccRCC:
-
Clear cell renal carcinoma
- PPI:
-
Protein–protein interactions
- DEGs:
-
Differentially expressed genes
- KIRC:
-
Kidney renal clear cell carcinoma
- HSCs:
-
Hematopoietic stem cells
- TKIs:
-
Tyrosine kinase inhibitors
- TME:
-
Tumor microenvironment
References
Moch, H., Cubilla, A. L., Humphrey, P. A., Reuter, V. E. & Ulbright, T. M. The 2016 WHO classification of tumours of the urinary system and male genital organs-part a: Renal, penile, and testicular tumours. Eur. Urol. 70, 93–105. https://doi.org/10.1016/j.eururo.2016.02.029 (2016).
Feng, X., Zhang, L., Tu, W. & Cang, S. Frequency, incidence and survival outcomes of clear cell renal cell carcinoma in the United States from 1973 to 2014: A SEER-based analysis. Medicine 98, e16684. https://doi.org/10.1097/md.0000000000016684 (2019).
Wu, P. et al. Integrated genomic analysis identifies clinically relevant subtypes of renal clear cell carcinoma. BMC Cancer 18, 287. https://doi.org/10.1186/s12885-018-4176-1 (2018).
Beuselinck, B. et al. Molecular subtypes of clear cell renal cell carcinoma are associated with sunitinib response in the metastatic setting. Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 21, 1329–1339. https://doi.org/10.1158/1078-0432.ccr-14-1128 (2015).
Zhu, Y. et al. Patient stratification of clear cell renal cell carcinoma using the global transcription factor activity landscape derived from RNA-seq data. Front. Oncol. 10, 526577. https://doi.org/10.3389/fonc.2020.526577 (2020).
Ficarra, V., Galfano, A., Mancini, M., Martignoni, G. & Artibani, W. TNM staging system for renal-cell carcinoma: Current status and future perspectives. Lancet Oncol. 8, 554–558. https://doi.org/10.1016/s1470-2045(07)70173-0 (2007).
Qi, Y. et al. New mechanistic insights of clear cell renal cell carcinoma from integrated miRNA and mRNA expression profiling studies. Biomed. Pharmacother. Biomedecine pharmacotherapie 111, 821–834. https://doi.org/10.1016/j.biopha.2018.12.099 (2019).
Chin, C. H. et al. cytoHubba: Identifying hub objects and sub-networks from complex interactome. BMC Syst. Biol. 8 Suppl 4, S11. https://doi.org/10.1186/1752-0509-8-s4-s11 (2014).
Mar, J. C. The rise of the distributions: Why non-normality is important for understanding the transcriptome and beyond. Biophys. Rev. 11, 89–94. https://doi.org/10.1007/s12551-018-0494-4 (2019).
Brannon, A. R. et al. Molecular stratification of clear cell renal cell carcinoma by consensus clustering reveals distinct subtypes and survival patterns. Genes Cancer 1, 152–163. https://doi.org/10.1177/1947601909359929 (2010).
Yang, H. et al. Exploring the mechanism of clear cell renal cell carcinoma metastasis and key genes based on multi-tool joint analysis. Gene 720, 144103. https://doi.org/10.1016/j.gene.2019.144103 (2019).
Fu, Q. et al. Prognostic value of interleukin-6 and interleukin-6 receptor in organ-confined clear-cell renal cell carcinoma: A 5-year conditional cancer-specific survival analysis. Br. J. Cancer 113, 1581–1589. https://doi.org/10.1038/bjc.2015.379 (2015).
Xu, T. et al. MMP9 and IGFBP1 regulate tumor immune and drive tumor progression in clear cell renal cell carcinoma. J. Cancer 12, 2243–2257. https://doi.org/10.7150/jca.48664 (2021).
Ren, X. et al. ENAM gene associated with T classification and inhibits proliferation in renal clear cell carcinoma. Aging 13, 7035–7051. https://doi.org/10.18632/aging.202558 (2021).
Goka, E. T., Chaturvedi, P., Lopez, D. T. M. & Lippman, M. E. Rac signaling drives clear cell renal carcinoma tumor growth by priming the tumor microenvironment for an angiogenic switch. Mol. Cancer Ther. 19, 1462–1473. https://doi.org/10.1158/1535-7163.MCT-19-0762 (2020).
Takakura, N. et al. A role for hematopoietic stem cells in promoting angiogenesis. Cell 102, 199–209. https://doi.org/10.1016/s0092-8674(00)00025-8 (2000).
Krystel-Whittemore, M., Dileepan, K. N. & Wood, J. G. Mast cell: A multi-functional master cell. Front Immunol. 6, 620. https://doi.org/10.3389/fimmu.2015.00620 (2015).
Obata, J. et al. Plasma fibrinogen level: An independent prognostic factor for disease-free survival and cancer-specific survival in patients with localised renal cell carcinoma. BJU Int. 118, 598–603. https://doi.org/10.1111/bju.13414 (2016).
Kang, X. & Shi, H. Combination of hematology indicators and oncological characteristics as a new promising prognostic factor in localized clear cell renal cell carcinoma. Cancer Manag. Res. 12, 10023–10033. https://doi.org/10.2147/cmar.s264400 (2020).
Klim, S. M. et al. Author correction: Fibrinogen—a practical and cost efficient biomarker for detecting periprosthetic joint infection. Sci. Rep. 10, 13795. https://doi.org/10.1038/s41598-020-70336-z (2020).
Ericsson, J. L., Seljelid, R. & Orrenius, S. Comparative light and electron microscopic observations of the cytoplasmic matrix in renal carcinomas. Virchows Archiv fur pathologische Anatomie und Physiologie und fur klinische Medizin 341, 204–223. https://doi.org/10.1007/bf00961071 (1966).
Du, W. et al. HIF drives lipid deposition and cancer in ccRCC via repression of fatty acid metabolism. Nat. Commun. 8, 1769. https://doi.org/10.1038/s41467-017-01965-8 (2017).
Sirwi, A. & Hussain, M. M. Lipid transfer proteins in the assembly of apoB-containing lipoproteins. J. Lipid Res. 59, 1094–1102. https://doi.org/10.1194/jlr.R083451 (2018).
Duns, G. et al. Histone methyltransferase gene SETD2 is a novel tumor suppressor gene in clear cell renal cell carcinoma. Cancer Res 70, 4287–4291. https://doi.org/10.1158/0008-5472.CAN-10-0120 (2010).
Gonzalez-Rodriguez, P. et al. SETD2 mutation in renal clear cell carcinoma suppress autophagy via regulation of ATG12. Cell Death Dis. 11, 69. https://doi.org/10.1038/s41419-020-2266-x (2020).
Elgendy, M. et al. Identification of mutations associated with acquired resistance to sunitinib in renal cell cancer. Int. J. Cancer 145, 1991–2001. https://doi.org/10.1002/ijc.32256 (2019).
Kovačić, M., Krvavica, A. & Rudić, M. Renal cell carcinoma metastasis to the sinonasal cavity: Case report. Acta Clin. Croat. 54, 223–226 (2015).
Pratt, D. et al. NDEx, the network data exchange. Cell Syst. 1, 302–305. https://doi.org/10.1016/j.cels.2015.10.001 (2015).
Ferronika, P. & Hof, J. Comprehensive profiling of primary and metastatic ccRCC reveals a high homology of the metastases to a subregion of the primary tumour. Cancers https://doi.org/10.3390/cancers11060812 (2019).
Wentink, M. Q. et al. Phase I study of dalteparin in combination with sunitinib in patients with metastatic clear cell renal carcinoma. Clin. Genitourin. Cancer https://doi.org/10.1016/j.clgc.2017.07.004 (2017).
Bendas, G. & Borsig, L. Cancer cell adhesion and metastasis: Selectins, integrins, and the inhibitory potential of heparins. Int. J. Cell Biol. 2012, 676731. https://doi.org/10.1155/2012/676731 (2012).
Goldman, M. J. et al. Visualizing and interpreting cancer genomics data via the Xena platform. Nat. Biotechnol. 38, 675–678. https://doi.org/10.1038/s41587-020-0546-8 (2020).
Netanely, D., Stern, N., Laufer, I. & Shamir, R. PROMO: An interactive tool for analyzing clinically-labeled multi-omic cancer datasets. BMC Bioinform. 20, 732. https://doi.org/10.1186/s12859-019-3142-5 (2019).
Tang, Z. et al. GEPIA: A web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 45, W98-w102. https://doi.org/10.1093/nar/gkx247 (2017).
Aguirre-Gamboa, R. et al. SurvExpress: An online biomarker validation tool and database for cancer gene expression data using survival analysis. PLoS ONE 8, e74250. https://doi.org/10.1371/journal.pone.0074250 (2013).
Fekete, J. T. & Gyorffy, B. ROCplot.org: Validating predictive biomarkers of chemotherapy/hormonal therapy/anti-HER2 therapy using transcriptomic data of 3104 breast cancer patients. Int. J. Cancer 145, 3140–3151. https://doi.org/10.1002/ijc.32369 (2019).
Szklarczyk, D. et al. STRING v11: Protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 47, D607-d613. https://doi.org/10.1093/nar/gky1131 (2019).
Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U.S.A. 102, 15545–15550. https://doi.org/10.1073/pnas.0506580102 (2005).
Shannon, P. et al. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504. https://doi.org/10.1101/gr.1239303 (2003).
Li, T. et al. TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res. 48, W509–W514. https://doi.org/10.1093/nar/gkaa407 (2020).
Acknowledgements
The results published here are based on the data generated by the TCGA Research Network: https://www.cancer.gov/tcga. The author greatly appreciate the patients and investigators who participated in TCGA and GEO for providing data.
Author information
Authors and Affiliations
Contributions
G.P. did all the investigation and computations, analyzed the data, prepared figures, and write the main manuscript text.
Corresponding author
Ethics declarations
Competing interests
The author declares no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Puzanov, G.A. Identification of key genes of the ccRCC subtype with poor prognosis. Sci Rep 12, 14588 (2022). https://doi.org/10.1038/s41598-022-18620-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-022-18620-y
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.