SPAG9 Expression Predicts Good Prognosis in Patients with Clear-Cell Renal Cell Carcinoma: A Bioinformatics Analysis with Experimental Validation

Clear-cell renal cell carcinoma (ccRCC) is the most common and aggressive type of renal-cell carcinoma (RCC). Sperm-associated antigen 9 (SPAG9) has been reported to promote the progression of a variety of tumors and is thus a potential prognostic marker. This study combined a bioinformatics analysis with an experimental validation, exploring the prognostic value of SPAG9 expression in ccRCC patients and the possible underlying mechanisms. The SPAG9 expression was associated with a poor prognosis in pan-cancer patients, but with a good prognosis and slow tumor progression in ccRCC patients. To explore the underlying mechanism, we investigated the roles of SPAG9 in ccRCC and bladder urothelial carcinoma (BLCA). The latter was chosen for comparison with ccRCC to represent the tumor types in which SPAG9 expression suggests a poor prognosis. The overexpression of SPAG9 increased the expression of autophagy-related genes in 786-O cells but not in HTB-9 cells, and SPAG9 expression was significantly correlated with a weaker inflammatory response in ccRCC but not in BLCA. Through an integrated bioinformatics analysis, we screened out seven key genes (AKT3, MAPK8, PIK3CA, PIK3R3, SOS1, SOS2, and STAT5B) in this study. The correlation between SPAG9 expression and ccRCC prognosis depends on the expression of key genes. Since most of the key genes were PI3K-AKT-pathway members, we used the PI3K agonist 740Y-P to stimulate the 786-O cells, to mimic the effect of key-gene overexpression. Compared with the Ov-SPAG9 786-O cells, the 740Y-P further increased the expression of autophagy-related genes by more than twofold. Moreover, we constructed a nomogram based on SPAG9/key genes and other clinical features, which was proven to have some predictive value. Our study found that SPAG9 expression predicted opposite clinical outcomes in pan-cancer and ccRCC patients, and we speculated that SPAG9 suppresses tumor progression by promoting autophagy and inhibiting inflammatory responses in ccRCC. We further found that some genes might cooperate with SPAG9 to promote autophagy, and that these were highly expressed in the tumor stroma and could be represented by key genes. The SPAG9-based nomogram can help to estimate the long-term prognosis of ccRCC patients, indicating that SPAG9 is a potential prognostic marker for ccRCC.


Introduction
Renal cell carcinoma (RCC) is the most common adult renal malignancy, of which the most common and aggressive histologic type is clear-cell renal cell carcinoma (ccRCC). Clear-cell renal cell carcinoma is characterized by abundant glycogen and lipids in the cytoplasm, accounting for 80-90% of RCC cases [1]. The incidence of ccRCC has increased steadily in recent decades, and due to its lack of specific clinical manifestations, nearly onethird of ccRCC patients have already developed distant metastasis at the time of diagnosis and, thus, missed the best opportunity for surgery [2]. Clear-cell renal cell carcinoma is insensitive to chemotherapy or radiotherapy, and although molecular targeted therapies have been widely used in patients with advanced ccRCC, most patients develop drug resistance after 5-11 months [3]. Therefore, it is necessary to investigate the molecular mechanisms underlying the occurrence and development of ccRCC and to develop potential therapeutic targets and prognostic markers.
This study combined a bioinformatics analysis and experimental validation, exploring the prognostic value of SPAG9 expression in ccRCC patients and the possible underlying mechanisms, and constructed a SPAG9-based ccRCC prognostic model to help with risk stratification.

Data Acquisition
The RNA-seq data and clinical data of ccRCC patients were downloaded from the TCGA database (https://portal.gdc.cancer.gov/ accessed on 15 June 2022). Gene expression was shown as FPKM (fragments per kilobase of transcript per million mapped fragments). The clinical data included age, gender, pathological grade (G1~G4), clinical stage (Stage I~Stage IV), T stage (tumor size, T1~T4), N stage (tumor lymph-node metastasis, N0~N1), and M stage (tumor distant metastasis, M0~M1). The clinical data of ccRCC included 537 patients, 526 of whom had complete and valid survival information (survival time and vital status), for Kaplan-Meier survival analysis, and 515 of whom had complete and valid survival information and clinical information (age, gender, pathological grade and clinical stage), for Cox regression analysis.
We also obtained RNA-seq data and clinical data of ccRCC from the ICGC database (https://dcc.icgc.org/ accessed on 15 June 2022): RECA-EU, including 91 patients with complete and valid survival information (survival time and vital status) and clinical information (age, gender). Basic information of the ccRCC patients from the TCGA database and the ICGC database is shown in Tables S1 and S2.
The scRNA-seq data were obtained from the GEO database (https://www.ncbi. nlm.nih.gov/geo/ accessed on 1 June 2022). Two ccRCC samples (GSM4630028 and GSM4630028) and one normal kidney-tissue sample (GSM4630031) were derived from the GSE152938 dataset, with a reading depth of 10× Genomics based on HiSeq X Ten (Illumina, San Diego, CA, USA) [20]. All 3 scRNA-seq samples were from human organisms.

Analysis with the Online Tool
The GEPIA2 is an online analysis tool developed by Tang et al., based on the data from the TCGA database (http://gepia2.cancer-pku.cn/ accessed on 15 June 2022) [21]. Some of the survival curves were established by GEPIA2: after entering gene names, median gene expression was selected as the group cutoff, and the overall survival (OS) rate of the 2 groups was compared by log-rank test. The online tool calculated the hazard ratios (HRs) based on the Cox PH Model, and 95% confidence interval (95% CI) was shown as dotted lines in the pictures. Axis units were set to months. Some correlation analyses were also performed by GEPIA2: after entering gene names, and Pearson's correlation coefficient between the expressions of genes was calculated.

Cell Culture and RNA Transfection
The human ccRCC cell line 786-O and BLCA cell line HTB-9 were provided by the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were cultured in DMEM-F12 medium containing 10% fetal bovine serum (FBS) and kept at 37 • C in a cell incubator with 5% CO 2 and 95% air. Overexpression of SPAG9 (Ov-SPAG9) and overex-

RNA Extraction and RT-qPCR Assay
Total RNA from cells was extracted using TRIzol reagent (Invitrogen) and reversetranscribed into cDNA using a RevertAid cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA). The RT-qPCR was performed with SYBR Green PCR Master Mix (Illumina), according to the manufacturer's instructions. Relative mRNA levels were measured with the aid of the 2 −∆∆Ct method and standardized to GAPDH. The primer sequences and the qPCR data are shown in Tables S3 and S4.
Gene set enrichment analysis (GSEA) software (version 4.0.3) was used to identify pathways related to SPAG9 [24,25]. The target set "C2.cp.kegg.v7.5.1.symbols.gmt" was downloaded from the Molecular Signatures database. The false discovery rate p value (FDR p value) and normalized enrichment score (NES) were used as screening criteria, and pathways with |NES| > 2 and FDR p values ≤ 0.05 were considered to be significantly enriched. Genes that contribute to the enrichment score (ES) of a particular pathway are called core enrichment genes, i.e., the leading-edge subset. According to the official description of GSEA, these genes can be interpreted as the core of a pathway and are, therefore, biologically important [25]. Based on the core enrichment genes of each pathway, we used the single-sample gene set-enrichment analysis (ssGSEA) algorithm to evaluate the pathway activity in samples.

Analysis with R Software
The R software (version 3.5.1) with the packages survival and survminer was used for the survival analysis. The Kaplan-Meier method was used to plot the survival curve, and log-rank was set as the statistical significance test. In Cox regression analysis, HRs were calculated based on the Cox PH Model. In the process of digitizing gender data, "female" was set as 0 and "male" was set as 1. Packages ggplot2 and ggpubr were used for the boxplots; Wilcoxon rank sum was set as the statistical significance test. The ratio of immune and stromal components in tumor samples was calculated by the ESTIMATE algorithm [26]. The heatmaps were plotted with the package Pheatmap. Correlation analysis was performed with the cor() function; Pearson's correlation test served as the statistical significance test.

ScRNA-seq Data Processing
The ScRNA-seq data were initially processed by the package Seurat in R software. The percentage of mitochondrial genes was calculated by the PercentageFeatureSet() function. The correlations between sequencing depth and mitochondrial gene sequences and between sequencing depth and total intracellular sequences were calculated. Genes detected in <3 cells, cells with <200 total detected genes, cells with <50 sequencing numbers, and cells with ≥5% mitochondrial gene sequences were excluded. After filtering, LogNormalize() was used to normalize the data; FindVariableFeatures() was used to identify the top 1500 hypervariable genes. Principal component analysis (PCA) was performed, and under the condition of p value < 0.05, dimensions with significant separation were screened out [27]. The 20 principal components (PCs) were selected for secondary dimensionality reduction through the tSNE algorithm [28]. Marker genes in each cell population were identified with the criteria of |log2[fold change(FC)]| > 1 and p value < 0.05. Cell populations were annotated by the package SingleR and manually corrected with the CellMarker database (http://biocc.hrbmu.edu.cn/CellMarker/index.jsp/ accessed on 15 June 2022) based on the marker genes [29,30]. Pseudotime trajectories were constructed by the package Monocle2; annotated cell populations were positioned to specific locations [31].

Prognostic Model Construction and Validation
The TCGA cohort (515 patients) was set as the training cohort for constructing the prognostic model, and the ICGC cohort (91 patients) was set as the validation cohort for external validation. The RiskScore was constructed with lasso regression using the R package glmnet [32]. RiskScore = ExpGENE1 × β1 + ExpGENE2 × β2 + . . . + ExpGENEn × βn, in which "Exp" represents the expression level of the corresponding gene and "β" represents the regression coefficient calculated by multivariate Cox analysis [33]. Age, gender, and RiskScore were utilized to construct the nomogram by using the R package rms. Calibration plots were constructed to evaluate the predictive abilities of the nomograms; the number of resampling was 1000 [34]. The receiver operating characteristic (ROC) curves were plotted using the R package timeROC. Time-dependent C-indexes were calculated and plotted using the pec package.

Statistical Significance
A p value < 0.05 was considered statistically significant.
In pan-cancer, SPAG9 expression was significantly associated with poor prognosis (Figure 2A). In all the tumor types except for ccRCC, SPAG9 was significantly correlated with OS in adrenocortical carcinoma (ACC), BLCA, and kidney chromophobe carcinoma (KICH), and it was a risk factor ( Figure 2B). with OS in adrenocortical carcinoma (ACC), BLCA, and kidney chromophobe carcinoma (KICH), and it was a risk factor ( Figure 2B).   with OS in adrenocortical carcinoma (ACC), BLCA, and kidney chromophobe carcinoma (KICH), and it was a risk factor ( Figure 2B).

SPAG9 Increases the Expression of Autophagy-Related Genes in 786-O Cells, but Not in HTB-9 Cells
Our study found that SPAG9 expression predicted opposite clinical outcomes in the pan-cancer and ccRCC patients. To explore the underlying mechanism, we investigated the roles of SPAG9 in ccRCC and bladder urothelial carcinoma (BLCA). The latter was chosen for comparison with ccRCC to represent the tumor types in which SPAG9 expression suggests poor prognosis. The SPAG9 expression was significantly positively correlated with OS in the ccRCC patients and significantly negatively correlated with OS in the BLCA patients. Previous studies showed that SPAG9 promotes proliferation and migration in both ccRCC and BLCA cell lines [4,6,14], while SPAG9 plays opposite prognostic roles in these two cancers. Therefore, we speculated that SPAG9 participates in other physiological processes in ccRCC.
The molecules LC3B, Beclin1, and p62 are landmarks in the autophagic process [37], and the intensity of autophagy was assessed by examining the expression levels of these genes. We overexpressed SPAG9 in the 786-O cells and HTB-9 cells and detected the expression levels of MAP1LC3B (LC3B), BECN1 (Beclin1), and SQSTM1 (p62) by RT-qPCR. The overexpression of SPAG9 significantly increased the MAP1LC3B, BECN1, and SQSTM1 mRNA levels in the 786-O cells (Ov-NC vs. Ov-SPAG9, p value < 0.05, Figure 3A), but not in the HTB-9 cells ( Figure 3B).

SPAG9 Increases the Expression of Autophagy-Related Genes in 786-O Cells, but Not in HTB-9 Cells
Our study found that SPAG9 expression predicted opposite clinical outcomes in the pancancer and ccRCC patients. To explore the underlying mechanism, we investigated the roles of SPAG9 in ccRCC and bladder urothelial carcinoma (BLCA). The latter was chosen for comparison with ccRCC to represent the tumor types in which SPAG9 expression suggests poor prognosis. The SPAG9 expression was significantly positively correlated with OS in the ccRCC patients and significantly negatively correlated with OS in the BLCA patients. Previous studies showed that SPAG9 promotes proliferation and migration in both ccRCC and BLCA cell lines [4,6,14], while SPAG9 plays opposite prognostic roles in these two cancers. Therefore, we speculated that SPAG9 participates in other physiological processes in ccRCC.
The molecules LC3B, Beclin1, and p62 are landmarks in the autophagic process [37], and the intensity of autophagy was assessed by examining the expression levels of these genes. We overexpressed SPAG9 in the 786-O cells and HTB-9 cells and detected the expression levels of MAP1LC3B (LC3B), BECN1 (Beclin1), and SQSTM1 (p62) by RT-qPCR. The overexpression of SPAG9 significantly increased the MAP1LC3B, BECN1, and SQSTM1 mRNA levels in the 786-O cells (Ov-NC vs. Ov-SPAG9, p value < 0.05, Figure 3A), but not in the HTB-9 cells ( Figure  3B).

SPAG9 Expression Was Significantly Correlated with a Weaker Inflammatory Response in ccRCC but Not in BLCA
Previous studies showed that autophagy activation inhibited tumor progression by suppressing the inflammatory response [38], so we investigated the relationship between SPAG9 and the inflammatory response in ccRCC and BLCA. We used the ssGSEA method to evaluate the inflammatory response, inflammatory factor production, and T-cell-exhaustion levels in the ccRCC and BLCA samples. As shown in Figure 4, SPAG9 expression was significantly correlated with a weaker inflammatory response, fewer inflammatory factors, and a lower degree of T-cell exhaustion in ccRCC (lower SPAG9 expression vs. higher SPAG9 expression, p value < 0.05, Figure 4A), but not in BLCA ( Figure 4B).

SPAG9 Expression Was Significantly Correlated with a Weaker Inflammatory Response in ccRCC but Not in BLCA
Previous studies showed that autophagy activation inhibited tumor progression by suppressing the inflammatory response [38], so we investigated the relationship between SPAG9 and the inflammatory response in ccRCC and BLCA. We used the ssGSEA method to evaluate the inflammatory response, inflammatory factor production, and T-cell-exhaustion levels in the ccRCC and BLCA samples. As shown in Figure 4, SPAG9 expression was significantly correlated with a weaker inflammatory response, fewer inflammatory factors, and a lower degree of T-cell exhaustion in ccRCC (lower SPAG9 expression vs. higher SPAG9 expression, p value < 0.05, Figure 4A), but not in BLCA ( Figure 4B).

The Correlation between SPAG9 Expression and ccRCC Prognosis Depends on the Expression of Key Genes
To further explore the role of SPAG9 in ccRCC, we calculated the ratio of the immune to the stromal components in the ccRCC samples by using the ESTIMATE algorithm. The ccRCC samples were divided into three groups, based on the median immune and stromal scores: a stroma-rich group (samples with StromalScore > median StromalScore and Im-muneScore < median ImmuneScore), an immunocyte-rich group (samples with Stro-malScore < median StromalScore and ImmuneScore > median ImmuneScore), and a hightumor-purity group (samples with StromalScore < median StromalScore and Im-muneScore < median ImmuneScore). A Kaplan-Meier survival analysis was conducted on the three groups, and the SPAG9 expression was significantly associated with OS only in the stroma-rich group ( Figure S1). Therefore, we performed the following analyses on the stroma-rich group. (1) The samples were divided into two groups based on the median SPAG9 expression, and GSEA was

The Correlation between SPAG9 Expression and ccRCC Prognosis Depends on the Expression of Key Genes
To further explore the role of SPAG9 in ccRCC, we calculated the ratio of the immune to the stromal components in the ccRCC samples by using the ESTIMATE algorithm. The ccRCC samples were divided into three groups, based on the median immune and stromal scores: a stroma-rich group (samples with StromalScore > median StromalScore and ImmuneScore < median ImmuneScore), an immunocyte-rich group (samples with StromalScore < median StromalScore and ImmuneScore > median ImmuneScore), and a high-tumor-purity group (samples with StromalScore < median StromalScore and Im-muneScore < median ImmuneScore). A Kaplan-Meier survival analysis was conducted on the three groups, and the SPAG9 expression was significantly associated with OS only in the stroma-rich group ( Figure S1). Therefore, we performed the following analyses on the stroma-rich group. (1) The samples were divided into two groups based on the median SPAG9 expression, and GSEA was conducted. As shown in Figure 5A, the genes in the high-SPAG9-expression group were significantly enriched in transmembrane signal transduction and phosphorylation regulatory pathways (adipocytokine-signaling pathway, ERBB-signaling pathway, JAK-STAT-signaling pathway, mTOR-signaling pathway, natural-killer-cell-mediated cytotoxicity, neurotrophin-signaling pathway, phosphatidylinositol-signaling system, and T-cell-receptor-signaling pathway), and the genes in the low-SPAG9-expression group were significantly enriched in mitochondrial dysfunction and abnormal calcium signaling pathways (Alzheimer's disease, cardiac muscle contraction, Huntington's disease, oxidative phosphorylation, and Parkinson's disease). (2) Based on the core enrichment genes of each pathway, we used the ssGSEA algorithm to evaluate the pathway activity, and the correlation between the pathway activity and the OS was then calculated. Among the pathways listed above, the activities of the adipocytokine-signaling pathway, ERBB-signaling pathway, and JAK-STATsignaling pathway were the most strongly correlated with the OS in the ccRCC patients, and these three pathways shared many common core enrichment genes ( Figure 5B,C, Table S6).
(3) From the common core enrichment genes of these three pathways, seven key genes (AKT3, MAPK8, PIK3CA, PIK3R3, SOS1, SOS2, and STAT5B) were screened out, and the expression of these genes had a high correlation with both the SPAG9 expression and the OS in ccRCC ( Figure 5D). The workflow for the screening of these key genes is shown in Table S7. In ccRCC, only the patients with high SPAG9 expression/high key-gene expression had a better prognosis ( Figure 6).   By analyzing the scRNA-seq data of the ccRCC patients, we found that the key genes were highly expressed in the tumor stromal cells (endothelial cells) ( Figure S2), which was consistent with the previous finding that SPAG9 expression was significantly associated with OS only in the stroma-rich group ( Figure S1).

The Key Genes May Have a Synergistic Effect with SPAG9 in Terms of Promoting Autophagy
Since most of the key genes belonged to the PI3K-AKT pathway ( Figure 6A), to demonstrate the synergistic relationship between SPAG9 and key genes, the PI3K agonist 740Y-P was used to stimulate the 786-O cells to mimic the effect of key-gene overexpression. The 740Y-P further increased the MAP1LC3B, BECN1, and SQSTM1 mRNA levels by more than twofold compared with Ov-SPAG9 cells (Ov-SPAG9 + 740Y-P vs. Ov-SPAG9, p value < 0.05, Figure 7)
The predictive power of the nomogram was validated using C-indexes and ROC curves. The C-indexes of the nomogram were higher than 0.6 in both the training cohort and the validation cohort, and they were higher for the patients with shorter disease courses ( Figure 9A). The ROC analysis indicated that the nomogram had some predictive value in both the training cohort and the validation cohort, especially in the patients with a 1-year disease course (in the training cohort: AUC at 1 years = 0.708, AUC at 3 years = 0.675, AUC at 5 years = 0.686; in the validation cohort: AUC at 1 years = 0.672, AUC at 3 years = 0.657, AUC at 5 years = 0.625, Figure 9B).

Discussion
It was reported that SPAG9 promotes proliferation and migration in many tumor-cell lines, including the ccRCC cell lines Caki-1 and NII-AKS413 [4]. However, we found that SPAG9 was associated with good prognosis in ccRCC patients, which contrasted with its widely recognized cancer-promoting role and was not found in other tumors. Thus, we speculated that SPAG9 participates in other physiological processes in ccRCC.
In ccRCC, Radovanovic et al. found that an increase in autophagic flux was associated with a lower tumor stage, reduced metastasis, and an improved 5-year survival rate [36]. Xu also found that the autophagy-promoting gene MAP1S was associated with better prognoses in ccRCC patients [35]. In ccRCC, studies reported that the activation of autophagy effectively inhibits tumor progression [41,42]. Moreover, prognostic models based on autophagy have also been widely developed and proven to have good predictive value [43][44][45][46][47]. Thus, autophagy may play an important role in ccRCC and, overall, it has a protective effect.
Autophagy inhibits inflammatory responses by inhibiting oxidative stress and lysosomal rupture [38]. Inflammation plays an important role in tumor progression: inflammatory responses stimulate tumor-associated macrophages to secrete cytokines, such as TGF-β and TNF-α, and promote tumor-cell metastasis; a persistent inflammatory response also leads to T-cell exhaustion and further tumor progression [22]. Clear-cell renal cell carcinoma is a tumor type with a high degree of immune infiltration [1]. The activation of autophagy in ccRCC may reduce the pro-inflammatory stress from tumor cells to the surrounding immune cells, inhibit tumor progression, and prolong the survival times of patients.
Therefore, we hypothesized that SPAG9 suppresses tumor progression by promoting autophagy in ccRCC. We found that the overexpression of SPAG9 significantly increased the mRNA levels of autophagy-associated genes in the 786-O cells but not in the HTB-9 cells. Moreover, the SPAG9 expression was significantly correlated with weaker inflammatory responses, fewer inflammatory factors, and a lower degree of T-cell exhaustion in ccRCC, but not in BLCA. Based on this information, we established a possible mechanism through which SPAG9 affects the prognosis of patients with ccRCC and BLCA: in ccRCC, SPAG9 promotes tumor growth but also promotes autophagy and inhibits the inflammatory response, and it has a protective effect on the whole; in BLCA, SPAG9 promotes tumor growth and leads to poor prognoses.
Through further analyses, we found that the correlation between SPAG9 expression and ccRCC prognosis was dependent on the tumor stroma. By analyzing stroma-rich samples, we identified seven key genes (AKT3, MAPK8, PIK3CA, PIK3R3, SOS1, SOS2, and STAT5B) in this study. In ccRCC, SPAG9 expression was significantly associated with OS only when key genes were highly expressed. The scRNA-seq data further proved that the key genes were highly expressed in the tumor stroma of ccRCC. Since most of the key genes were PI3K-AKT-pathway members, we used the PI3K agonist 740Y-P to stimulate the 786-O cells to mimic the effect of key-gene overexpression. Furthermore, the addition of 740Y-P significantly increased the expression of the autophagy-related genes by more than twofold compared with the Ov-SPAG9 cells. In conclusion, we speculated that in the tumor stroma, some genes cooperate with SPAG9 to promote autophagy, and that these may be key genes, which can be represented by the key genes.
Finally, we constructed a nomogram based on the SPAG9/key genes and other clinical features to predict survival in the ccRCC patients. Combined with external data validation, the nomogram proved to have some predictive value. The estimation of the long-term prognosis of ccRCC patients can help physicians develop individualized treatment plans [48]. With this new nomogram, we look forward to providing more helpful guidance for clinical work.

Conclusions
In recent years, research on the role of SPAG9 in tumors has mainly focused on its role in promoting tumor progression [49], but our study found that SPAG9 expression was significantly associated with good prognoses in ccRCC, a result that has not been found in other cancer types and contradicts the widely recognized cancer-promoting role of SPAG9. Combined with a bioinformatics analysis and an experimental validation, we speculated that SPAG9 suppresses tumor progression by promoting autophagy and inhibiting inflammatory responses in ccRCC. We further found that some genes might cooperate with SPAG9 to promote autophagy, and that these were highly expressed in the tumor stroma and may be key genes. Finally we constructed a nomogram based on SPAG9/key genes and other clinical features, which was proven to have some predictive value.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/genes14040944/s1. Figure S1: SPAG9 expression is significantly associated with OS only in the stroma-rich group. Figure S2: The key genes are highly expressed in ccRCC stromal cells. Table S1: Baseline clinical characteristics of ccRCC patients from the TCGA database. Table S2: Baseline clinical characteristics of ccRCC patients from the ICGC database. Table S3: Primers designed for this study. Table S4: The qPCR data after analysis with 2 −∆∆Ct method. Table S5: The role of SPAG9 in different tumor-cell lines; Table S6: The correlation between the activity scores of pathways and OS in ccRCC. Table S7: Workflow for screening of key genes.
Author Contributions: Conceptualization, writing-review and editing, validation, and funding acquisition: H.W.; methodology, software, formal analysis, investigation, resources, data curation, writing-original draft preparation, and visualization: L.Q. Funding acquisition and project administration: L.Z. All authors have read and agreed to the published version of the manuscript.