Multi-omics profiling identifies ectopic olfactory receptors as putative drivers of tumor progression and prognostic indicators in clear cell renal cell carcinoma

Dong Jun Yeo; Hee Jin Cho

doi:10.1007/s10238-026-02118-2

. 2026 Mar 21;26(1):198. doi: 10.1007/s10238-026-02118-2

Multi-omics profiling identifies ectopic olfactory receptors as putative drivers of tumor progression and prognostic indicators in clear cell renal cell carcinoma

Dong Jun Yeo ^1,², Hee Jin Cho ^1,^2,^3,^✉

PMCID: PMC13013169 PMID: 41863682

Abstract

Olfactory receptors (ORs) are a subclass of G-protein-coupled receptors (GPCRs) that are primarily expressed in olfactory sensory neurons. Furthermore, ORs have recently been identified in the tumor microenvironment (TME) of various cancers, suggesting potential involvement for ORs in tumor progression. However, the roles of ORs in clear cell renal cell carcinoma (ccRCC), the most prevalent and aggressive subtype of kidney cancer, characterized by limited therapeutic response and a 5-year survival rate of only 10% in advanced stages, have yet to be elucidated. In this study, an integrative multi-omics analysis combining bulk transcriptomic profiles from The Cancer Genome Atlas (TCGA) and single-cell RNA- and ATAC-sequencing datasets from ccRCC patients to characterize the context- and cell-type-specific functions of ORs within the TME. Notably, ORs exhibited distinct, cell-type-specific expression profiles within ccRCC TME. OR51E1 was predominantly expressed in pericytes and correlated with vascular remodeling and angiogenic activity, whereas OR2T10 was enriched in malignant epithelial cells and associated with invasive and metastatic potential, with each OR being associated with resistance to therapeutic agents. In addition, OR-based prognostic modeling and tumor clustering both identified unfavorable prognostic signatures associated with poor patient outcomes and TME-related immunosuppression. These findings highlight ectopic OR networks as clinically relevant molecular features of ccRCC progression and position ORs as potential actionable biomarkers and potential therapeutic targets, offering new avenues for precision oncology in ccRCC.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10238-026-02118-2.

Keywords: Clear cell renal cell carcinoma (ccRCC), Olfactory receptors (ORs), Tumor microenvironment (TME), Single-cell RNA sequencing (scRNA-seq), Single-cell ATAC sequencing (scATAC-seq)

Introduction

Clear cell renal cell carcinoma (ccRCC) represents the predominant histological subtype of renal cell carcinoma (RCC), accounting for approximately 80% of all RCC cases [1]. ccRCC originates from epithelial cells in the proximal nephron and tubular epithelium, characterized by a clear cytoplasm due to the accumulation of intracellular lipid and glycogen and a high mutation rate in the VHL, PBRM1, and SETD2 genes [2, 3]. A total of 434,800 new RCC cases were reported globally in 2022, while 71,759 patients were diagnosed with RCC, and 14,295 RCC-related deaths were estimated in the USA [4]. The following year, approximately 81,800 cases of RCC were anticipated, and 14,890 mortality cases were estimated in the USA [5]. Regarding RCC statistics, the incidence of RCC has been consistently increasing. While patients with localized ccRCC exhibit relatively favorable 5-year survival rates ranging from 50 to 69%, those with highly advanced or metastatic ccRCC face a significantly worse prognosis, with a 5-year survival rate of only 10% [1]. This stark clinical disparity highlights the pressing need to identify novel targets for diagnosis and therapy to improve outcomes in ccRCC.

Olfactory receptors (ORs) are a subclass of G-protein-coupled receptors (GPCRs) predominantly expressed on the membrane of olfactory sensory neurons (OSNs), where the ORs mediate the detection of odorant molecules. Representing the largest gene superfamily in the human genome, OR genes account for approximately 2% of all protein-coding genes, reflecting the remarkable diversity and abundance of these receptor genes [6]. In humans, the repertoire of OR genes comprises roughly 389 functional protein-coding genes, along with an estimated 485 pseudogenes [7]. Despite the substantial number of OR genes, OR expression follows a highly specific regulatory mechanism in which each OSN expresses a single allele from a single OR gene [8]. Nevertheless, recent studies have demonstrated that this regulatory system can be disrupted, leading to aberrant overexpression or ectopic expression of ORs in non-olfactory tissues beyond OSNs. Under such conditions, ORs are no longer restricted to odorant detection but may also perform diverse cellular functions, including regulating cell growth, migration, and secretion [9]. This ectopic expression often confers non-canonical roles, which have been repeatedly linked to pathological conditions.

For example, OR11H1 is reported to be upregulated in Alzheimer’s disease, whereas the expression levels of OR4F4, OR10G8, and OR52L1 are decreased in the affected cortical regions [10]. Similarly, several ORs, including OR2L13, OR1E1, OR2J3, OR52L1, and OR11H1, were found to be downregulated in the early stages of Parkinson’s disease, particularly in the cerebral cortex and substantia nigra, implicating the expression dynamics of ORs in the progression of neurodegenerative disorders [11].

In addition to neurological disorders, the ectopic expression of ORs has also been associated with cancer biology. Prior studies have demonstrated that the activation of specific GPCR-mediated intracellular signaling pathways are often implicated in cancer progression. For example, activation of CXCR4 by its ligand CXCL12 engages MAPK and PI3K/AKT cascades associated with tumor cell proliferation, survival, and epithelial-to-mesenchymal transition (EMT) [12]. Likewise, stimulation of lysophosphatidic acid (LPA) receptors (LPAR1–LPAR3) by LPA forms β-arrestin–dependent signaling complexes that scaffold MAPK, Src, and NF-kB pathways, thereby supporting invasive cancer behavior [13]. Given that ORs belong to the GPCR superfamily, their ectopic activation within tumors similarly raises the possibility that these receptors may be involved in oncogenic signaling and cancer progression. Consistent with this notion, several studies have reported that ORs aberrantly expressed across diverse cancer types are associated with tumor malignant behaviors [14]. For instance, OR7C1 has been shown to promote tumorigenesis in colon cancer [15]. Likewise, OR2W3 has been implicated in promoting cancer cell invasiveness and is associated with triple-negative breast cancer (TNBC) [16].

Building upon these findings, the present study aimed to investigate the clinical and biological significance of ectopically expressed OR genes in ccRCC, with a specific focus on the biological and clinical relevance of these receptor genes to tumor progression. While previous reports have documented OR dysregulation across various cancer types, the biological relevance of this dysregulation in ccRCC remain unexplored. To address this gap, an integrative bioinformatics framework was employed, encompassing bulk RNA-seq, single-cell RNA-seq (scRNA-seq), and single-cell ATAC-seq (scATAC-seq) datasets. This multi-dimensional approach enabled us to delineate the cellular contexts, regulatory dynamics, and intercellular communication patterns associated with OR expression, potentially providing insights into the potential utility of these receptors as diagnostic biomarkers or therapeutic targets in ccRCC.

Methods

Data collection

This study utilized multiple publicly available datasets, including bulk RNA-seq, scRNA-seq, scATAC-seq, and Hi-C data. Bulk RNA-seq data of ccRCC for this study were acquired from TCGA. Both raw count and fragment per kilobase of transcript per million mapped reads (FPKM) values were retrieved using the ‘GDCquery’ function in the R package ‘TCGAbiolinks’ (accessed on 22 November 2022). To ensure accurate clinical stratification, samples from two patients lacking American Joint Committee on Cancer (AJCC) pathological stage information in the datasets were excluded from subsequent analyses. The ccRCC validation cohort (E-MTAB-1980), used for evaluating the prognostic model, was retrieved from the ArrayExpress database (https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-1980, accessed March 29, 2025). The scRNA-seq datasets, including tumor and normal kidney samples, were obtained from the Gene Expression Omnibus (GEO) under accession numbers GSE152938 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE152938, accessed on 24 January 2024), GSE222703 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE222703, accessed on 5 July 2024), and GSE242299 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE242299, accessed on 10 July 2024) [17–19]. The scATAC-seq dataset was downloaded from GSE207493 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE207493, accessed on 15 November 2024), and Hi-C data for the 786-O ccRCC cell line were sourced from GSE99051 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE99051, accessed on 24 January 2025) [20, 21].

Single-cell RNA-seq data processing

Single-cell RNA-seq analysis was performed using the R program (version 4.4.1) and the R package ‘Seurat’ (version 5.1.0). Quality control (QC) was established by filtering cells based on the following criteria: number of unique molecular identifiers (UMIs) ≥ 500, number of detected genes ≥ 200, log-transformed genes per UMI (log10GenesPerUMI) > 0.80, and mitochondrial gene content < 20%. Data normalization, integration (‘IntegrateLayers’, CCA), dimensionality reduction (PCA, UMAP), and clustering (‘FindNeighbors’, ‘FindClusters’) were performed using standard Seurat workflows.

Cell-type annotation was performed using three complementary approaches: (1) differential expression analysis using ‘FindMarkers’ function with the Wilcoxon rank sum test; (2) ‘SingleR’ (version 1.8.1), with ‘celldex’ reference atlas (version 1.4.0); (3) the Azimuth kidney reference mapping (L1/L2 atlases) [22–24].

Single-cell ATAC-seq data processing

Single-cell ATAC-seq analysis was conducted using the R package ‘Signac’ (version 1.14.0), which is fully compatible with the Seurat package for the integrated single-cell analysis. A unified peak set was generated across all 19 samples. Fragment files were converted into fragment objects using the ‘CreateFragmentObject’ function. A sparse matrix of peak accessibility was then created via the ‘FeatureMatrix()’ function.

Next, QC was applied based on following criteria: (1) 1000 < peak region fragments < 20,000; (2) the proportion of reads mapping to ENCODE blacklist regions (blacklist ratio) < 0.05; (3) nucleosome signal < 4; (4) transcription start site (TSS) enrichment score > 1. After filtering, term frequency-inverse document frequency (TF-IDF) normalization, singular value decomposition (SVD) dimensionality reduction (PC2-40), and graph-based clustering (resolution = 0.6) were performed.

To assign cell-type labels, gene activity scores were computed using the ‘GeneActivity’ function, which quantifies accessibility over gene bodies extended 2 kb upstream of the TSS. The resulting gene activity matrix was subjected to the same annotation strategies as those employed for the scRNA-seq data.

Motif and footprint analyses

The motif analysis was performed using the ‘chromVAR’ R package (version 1.28.0). Motif annotations were incorporated by applying the ‘AddMotif’ function with the JASPAR 2020 motif database as a reference [25]. Per-cell motif accessibility scores were computed via ‘chromVAR’, enabling quantification of motif activity across individual cells. These scores were subsequently normalized using z-score transformation, and differential motif activity between cell types was assessed by computing the average difference (avg_diff) in z-scores. A footprinting analysis was conducted for transcription factors (TFs) using ‘Footprint’, and the results were visualized using the ‘PlotFootprint’ function.

Hi-C data analysis

Hi-C data from the 786-O ccRCC cell line (GSE99051) were preprocessed using the ‘cooler’ package (version 0.8.11). Downstream analysis was performed with the ‘HiCExperiment’ (version 1.6.0) and ‘HiContacts’ (version 1.8.0) R packages. For TAD identification, both the ‘TopDom’ R package (version 0.10.1) and the ‘getDiamondInsulation’ function from HiContacts were employed. Visualization of the contact matrices and insulation profiles was conducted using the ‘plotMatrix’ function from HiContacts and the ‘HiCPlotter’ Python tool [26]. Chromosomal compartment analysis was performed using the ‘FAN-C’ Python toolkit (version 0.9.28) [27]. A/B compartment matrices were generated via eigenvector decomposition of the Hi-C correlation matrix using the ‘fanc compartments’ function. The resulting compartments were visualized with the ‘fancplot’ function within the same toolkit.

Cell-to-cell communication analysis

The cell–cell communication analysis was performed using the ‘CellChat’ (version 2.1.2) with human ligand-receptor database ‘CellChatDB.human’. Seurat object was converted to CellChat object, and communication probabilities were computed with standard workflow (‘identifyOverExpressedGenes’, ‘identifyOverExpressedInteractions’, ‘computeCommunProb’, ‘filterCommunication’, ‘aggregateNet’). Communication patterns were identified separately for outgoing/incoming signaling (‘identifyCommunicationPatterns’, elbow method for k), clustering cell types into signaling modules.

Consensus clustering

Consensus clustering was conducted using kidney renal cell carcinoma (KIRC) tumor samples from TCGA and the ‘ConsensusClusterPlus’ R package (version 1.70.0). The clustering was performed using the Manhattan distance and Ward’s method (ward.D) for both inner and outer linkage functions, while the default values were selected for all other parameters. The optimal number of clusters (k) was determined based on the proportion of ambiguous clustering (PAC) score [28].

Weighted gene co-expression network analysis (WGCNA)

Bulk TCGA-KIRC (FPKM), and scRNA-seq cancer cell co-expression networks were constructed using ‘WGCNA’ (version 1.73) and ‘hdWGCNA’ (version 0.4.03), respectively. At the bulk level, hierarchical clustering was applied to identify and excluded outlier samples. To account for the mean–variance relationship in the counts, expression data were transformed using variance stabilizing transformation (VST). To focus on biologically informative signals, we selected the top 10,000 genes based on median absolute deviation (MAD) for downstream network construction. The ‘goodSamplesGene’ function was further utilized to ensure data quality. At single-cell level, only genes expressed in at least 5% of cancer cells were included. Among the OR2 gene family, OR2T10 and OR2G6 satisfied this criterion and were thus selected for downstream analysis. The nearest-neighbor parameter was set to k = 15, and max_shared was set to 10, while all remaining parameters were set to their default configurations. In both bulk and single-cell settings, the appropriate soft-thresholding power was determined by evaluating the scale-free topology fit index and mean connectivity.

Survival analysis

Survival analysis was conducted using RNA-Seq and clinical data from the TCGA-KIRC cohort. Kaplan–Meier survival curves were generated using the ‘survival’ R package (version 3.8–3) and visualized with the ‘survminer’ package (version 0.5.0). Two tumor samples with missing AJCC pathological stage information were excluded from the analysis. For risk-based stratification, tumor samples were divided into a high-risk group (top 30%) and a low-risk group (bottom 70%) based on their risk scores. In addition, survival comparisons across consensus clusters were performed by grouping samples according to their assigned cluster. The concordance index (C-index) was calculated using the ‘survival’ R package. Statistical differences between survival curves were assessed using the log-rank test.

Construction of and validation of a prognostic model

To identify potential prognostic genes among tumor-upregulated OR genes identified by DESeq2, a univariate Cox regression was performed. OR genes with p-value < 0.1 were retained for further analysis. To avoid model overfitting and select the most relevant predictors, the Least Absolute Shrinkage and Selection Operator (LASSO) regression was applied. Subsequently, multivariate Cox regression was conducted, incorporating clinical covariates such as AJCC pathologic stage, gender, and age. Genes with a hazard ratio > 1 and p-value < 0.05 were considered significant and included in the final prognostic model. The prognostic risk score for each patient was computed using the following formula:

where Inline graphic represents the regression coefficient from the multivariate Cox model and denotes the expression level of the OR gene . Based on the risk score, patients were stratified into high-risk (top 30%) and low-risk (bottom 70%) groups. Kaplan–Meier survival curves were generated, and risk score distributions were visualized using the ‘ggrisk’ R package (version 1.3). Predicted performance of the risk model for 1-, 3-, and 5-year overall survival (OS) was evaluated using time-dependent receiver operating characteristic (ROC) analysis with the ‘timeROC’ R package (version 0.4). For external validation, the E-MTAB-1980 ccRCC cohort was utilized.

Differentially expressed genes analysis

Differential expression analysis was performed to identify upregulated and downregulated OR genes in tumors compared to normal samples at the bulk transcriptomic level using three independent tools: DESeq2, edgeR, and limma [29–31]. Genes were considered upregulated or downregulated based on a log2 fold change > 1 and an adjusted p-value < 0.05. For differential expression comparisons between tumor subgroups, DESeq2 was applied. At single-cell level, differentially expressed genes (DEGs) were identified using the ‘FindMarkers’ function from the ‘Seurat’ package.

Gene set enrichment analysis (GSEA)

GSEA was performed using gene sets from the Molecular Signatures Database (MSigDB, version 7.5.1), specifically the C2 (curated gene sets), C5 (GO gene sets), and Hallmark collections [32]. To assess the functional enrichment of gene sets derived from DEGs, the ‘enricher’ function from the ‘clusterProfiler’ R package (version 4.14.6) was employed. To evaluate the normalized enrichment scores (NES) of pathways based on ranked DEG results, both at the bulk and single-cell levels, the ‘fgsea’ R package (version 1.32.4) was utilized. All p-values were adjusted for multiple testing using the false discovery rate (FDR) method.

Gene ontology analysis

Gene ontology (GO) analysis was conducted using the g:Profiler web tool (https://biit.cs.ut.ee/gprofiler/), based on the list of DEGs [33]. Gene symbols were mapped using the most specific and functionally annotated identifiers available, resolving ambiguous mapping by selecting the most informative annotations.

Quantification of pathway activity

To investigate functional pathway activity, gene set enrichment scoring was performed at both the bulk and single-cell levels. At the bulk transcriptomic level, pathway enrichment scores were calculated using the single-sample gene set enrichment (ssGSEA) method implemented in the ‘GSVA’ R package (version 1.42.0), based on the FPKM expression data for the TCGA-KIRC dataset. For scRNA-seq data, the ‘AUCell’ R package (version 1.28.0) was used in conjunction with ssGSEA. Tumor purity was calculated with ‘estimate’ R package (version 1.0.13).

Analysis of drug response

Drug sensitivity data, represented as log-transformed (LN_IC₅₀) values, and transcriptomic profiles of KIRC cell lines were retrieved from the Genomics of Drug Sensitivity in Cancer (GDSC) database [34]. Pearson correlation analysis was performed to assess the association between LN_IC₅₀ values of each drug and the expression levels of the OR genes. Drugs exhibiting statistically significant correlations were further classified and visualized in relation to the associated target pathways.

Statistical analysis

Statistical analysis was conducted using R software (version 4.4.1). The significance of differences between groups was assessed using Student’s t-test. Pearson correlation coefficients were computed using the ‘Hmisc’ R package (version 5.2–3).

Results

Overall expression profiles of OR genes in kidney renal cell carcinoma

In this study, bulk RNA-seq, scRNA-seq, and scATAC-seq data were utilized to investigate the potential contribution of ORs to the progression of ccRCC (Fig. 1a). Bulk RNA-seq analysis was performed using the TCGA-KIRC (kidney renal clear cell carcinoma) dataset, which included transcriptomic profiles from 537 patients [35]. Both the scRNA-seq and scATAC-seq datasets were obtained from the GEO database. The scRNA-seq data were derived from tumor tissues of 13 ccRCC patients, while the scATAC-seq data were obtained from tumor tissues of 19 distinct ccRCC patients.

Fig. 1 — Overall expression profiles of OR genes in KIRC. a Schematic overview of the analytical workflow integrating bulk RNA-seq, scRNA-seq, and scATAC-seq data to explore OR gene functions and prognostic significance in ccRCC. Created with BioRender.com b Box plot comparing OR gene family scores between tumor and normal samples from TCGA-KIRC data. Statistical significance was calculated using Student’s t-test. c Volcano plot of the differentially expressed OR genes from the DEG analysis for KIRC (tumor vs normal samples). Pink, upregulated in tumor; blue, downregulated in tumor; gray, not significant. d Venn diagrams of the overlap of upregulated (top) and downregulated (bottom) OR genes between tumor and normal samples in KIRC. e Inclusion of OR genes within cancer hallmark gene sets, based on Menyhart et al. [36]. The x-axis displays upregulated ORs and other ORs in tumors. The y-axis indicates cancer hallmarks. Colored cells represent cancer hallmarks that encompass at least one OR gene

To explore the potential association between the OR genes and the progression of ccRCC, the overall expression patterns of the OR genes were first examined using the TCGA-KIRC data. Initially, the OR gene family scores were evaluated based on this dataset. A total of 843 OR genes were grouped into 18 gene families, and the associated gene scores were computed, which was subsequently compared between tumor and normal samples. The scores of OR families 1, 6, 8, 9, 10, 11, 51, 52, and 56 were significantly higher in the tumor samples compared to normal samples, whereas OR families 3, 13, and 14 showed higher scores in the normal samples (Fig. 1b).

Therefore, to identify specific OR genes that are significantly dysregulated in tumors compared to normal tissues further, a differential expression (DE) analysis was performed using the KIRC data to identify OR genes with distinct expression patterns between tumor and normal samples. As a result, a total of ten OR genes (OR51E1, OR51E2, OR52K3P, OR52N4, OR2A4, OR2I1P, OR5BA1P, OR7E47P, OR9Q1, and OR10Q1) were consistently identified as tumor-upregulated ORs across all three methods using this multi-tool approach (Fig. 1c–d).

To assess the potential relevance of OR genes in cancer progression, we examined the inclusion of these genes in the cancer hallmark gene sets provided by Menyhart et al. [36]. Among the upregulated OR genes in tumors, protein-coding ORs were found to be enriched in gene sets for sustaining proliferative signaling and evading growth suppression (Fig. 1e). In addition to the tumor-upregulated ORs, OR10J5 was also included in the gene set for sustained angiogenesis, and OR51E2 was involved in the gene set for tissue invasion and metastasis. Although OR genes have yet to be extensively mapped to other cancer hallmarks, the presence of 397 OR genes within the sustaining proliferative signaling and evading growth suppression gene sets suggests broader potential involvement of OR genes in various aspects of ccRCC progression.

Construction of a prognostic model with the OR gene expression

OR genes have been implicated in contributing to clinical outcomes across multiple cancer types. For instance, the elevated expression of OR2G2 has been associated with adverse prognoses in acute myeloid leukemia [37]. Based on these findings, OR genes exhibiting significant prognostic associations in KIRC were identified, and a prognostic model was subsequently constructed based on these prognostic OR genes to predict clinical outcomes in patients with ccRCC.

To assess the prognostic relevance of OR genes, a stepwise analysis framework using 125 tumor-upregulated OR genes was applied, consisting of univariate Cox regression, LASSO-based variable selection, and subsequent multivariate Cox regression (Additional file 2: Table S1, Fig. 2a–b and Additional file 1: Fig. S1a). As a result, three OR genes (OR2B6, OR13A1, and OR11H7) were selected for inclusion in the prognostic model, all of which remained significantly associated with poor prognosis and exhibited hazard ratios greater than one (Fig. 2c). Notably, these genes consistently demonstrated high hazard ratios and maintained their prognostic significance after adjusting for tumor purity as a clinical covariate (Fig. S1b). Then, risk scores were calculated as a weighted sum of gene expression levels: 0.2706 × OR2B6 expression + 0.1483 × OR13A1 expression + 0.1950 × OR11H7 expression.

Fig. 2 — Construction of a prognostic model using the OR gene expression. a An evaluation of LASSO coefficients distribution, based on 33 OR genes with p-values < 0.1 selected from a univariate regression analysis of 125 upregulated genes in tumors identified by DESeq2. b Selection of the most optimal penalty parameter in the LASSO model using cross-validation to ensure model stability. c Forest plot of multivariate Cox regression including three OR genes filtered by LASSO regression, alongside age, AJCC pathological stage, and gender as covariates. d Risk plot of the distribution of survival status of individual tumor samples. e Plot of C-index of risk scores and clinical parameters for prognostic performance evaluation. f Kaplan–Meier survival curves of TCGA-KIRC cohort, stratified by risk score into high-risk (top 30%, n = 151) and low-risk (bottom 70%, n = 354) groups. g Kaplan–Meier survival curves of E-MTAB-1980 cohort, classified by risk score as high-risk (top 30%, n = 30) and low-risk (bottom 70%, n = 69) groups. h The ROC curve of the risk stratification capability and predictive power of the established risk model. i A heatmap of the correlation between gene modules and traits, including three prognostic OR genes. j Dot plot of a summary of the top 30 enriched pathways identified by GSEA on genes from the *OR13A1*-correlated module. The analysis was conducted with the C2, C5, and Hallmark gene sets from MSigDB, and highlights the highest gene ratios and adjusted p-values below 0.05

Based on the calculated risk scores, patients were stratified into a high-risk group (top 30%) and a low-risk group (bottom 70%). We observed that the expression levels of three prognostic OR genes were markedly enriched in the high-risk group (Fig. 2d). Notably, the risk score achieved a superior C-index of 0.7727, outperforming established factors such as AJCC stage, age, gender, and tumor purity (Fig. 2e, Additional file 3: Table S2). Consistent with this observation, survival analysis demonstrated that the high-risk group had a significantly worse prognosis than the low-risk group (Fig. 2f). Furthermore, time-dependent ROC analysis confirmed the favorable predictive accuracy of the model, with area under the curve (AUC) values of 0.659, 0.656, and 0.703 at 1, 3, and 5 years, respectively (Fig. 2h). The prognostic model was further validated in an independent cohort (E-MTAB-1980) comprising 100 distinct patients [38]. Upon applying the same analytical pipeline, the high-risk group consistently exhibited significantly poor survival outcomes, thereby supporting the applicability of the constructed model for prognostic prediction in patients with ccRCC (Fig. 2g).

Next, a WGCNA was performed to investigate the potential biological mechanisms underlying the association of these three OR genes with poor prognosis (Fig. S1c–d) [39]. Subsequently, 12 modules were determined using the selected soft-thresholding power (Fig. S1e–f). Module–trait relationships were subsequently examined by defining the expression levels of OR13A1, OR11H7, and OR2B6, as phenotypic traits (Fig. 2i). As a result, notable associations were observed between individual gene expression and specific module: OR2B6 correlated with MEred (r = 0.32) and MEpink (r = 0.29); OR13A1 with MEpink (r = 0.41) and MEpurple (r = 0.39); and OR11H7 with MEblue (r = 0.58) and MEred (r = 0.48).

To further elucidate the biological functions of the three OR genes included in the prognostic model, the genes belonging to the modules most strongly correlated with each OR gene and their associated pathways were analyzed. The MEpink module, which was correlated with both OR2B6 and OR13A1, was significantly enriched in cancer-associated pathways such as cell cycle and proliferation (Fig. S1e). Another OR13A1-associated module, MEpurple (160 genes), was primarily linked to regulatory B cell (B reg) functions, including IL-10 production, B cell receptor (BCR) signaling, and immunoglobulin production (Fig. 2j) [40, 41]. Meanwhile, two distinct modules were identified for OR11H7: MEblue (1,405 genes) and MEred (594 genes). The MEblue module was enriched for cilium-based motility, a process increasingly recognized for its role in cancer cell migration (Fig. S1f) [42]. In contrast, no significantly enriched pathways were detected for the MEred module. Collectively, these findings suggest that among the OR genes included in the prognostic model, OR13A1 is primarily associated with B reg-related pathways; meanwhile, OR11H7 and OR2B6 are predominantly linked to cancer cell-intrinsic pathways. These associations may underlie the adverse prognostic impact observed in ccRCC patients.

Classification of tumor samples with OR gene expression based on consensus clustering

While the preceding analysis focused on identifying significant OR genes affecting patient prognosis and constructing a predictive model, the subsequent investigation aimed to explore the characteristics of clusters derived from consensus clustering of tumor samples based on the expression patterns of 125 tumor-upregulated OR genes [43]. Additionally, the relationships among OR genes exhibiting cluster-specific expression patterns were examined.

Based on consensus clustering, five clusters were determined as the optimal solution (Fig. 3a). The resulting clusters, designated as Cluster 1–5, comprised 304, 107, 36, 73, and 19 tumor samples, respectively. Subsequently, an xCell analysis was performed to assess the cell-type and immune/stromal scores across the identified clusters (Fig. 3b) [44]. These results revealed that Cluster 1 exhibited a high stromal score and low immune score, whereas Clusters 3 and 5 displayed the opposite pattern, characterized by low stromal scores and high immune scores. Clusters 2 and 4 showed intermediate levels of both stromal and immune scores (Fig. 3c). Consistent with xCell results, analysis with EPIC further revealed that Cluster 3 and 5 exhibited higher cell fraction, including CD8⁺ T cells and macrophages, thereby supporting the classification of these cluster as immune-active. In contrast, the estimated fractions of CD4⁺ T cells did not differ significantly among Clusters 1 through 5 (Fig. S2a) [45].

These findings suggest that clustering based on OR gene expression effectively stratified samples according to distinct immune and stromal microenvironmental profiles.

Based on these findings, cluster-specific pathways and OR genes were identified. To further delineate the molecular distinctions among transcriptionally defined subpopulations, Clusters 3 and 5 were initially compared. Cluster 3 was enriched for pathways related to multi-cancer invasiveness and cell adhesion (Fig. S2b). In contrast, Cluster 5 showed enrichment for gene sets associated with the T cell receptor (TCR) complex, ribosomal activity, and translation (Fig. S2c). Notably, Cluster 3 also demonstrated relatively higher enrichment for TCR complex-related pathways compared to the other four clusters, despite a primary association with invasive signatures (Fig. S2d). Despite these differences, both clusters exhibited similarly high immune scores and low stromal scores, with minimal divergence across most immune-related pathways, except for TCR. Given the comparable immunological characteristics and the relatively small number of samples in Clusters 3 and 5, these clusters were merged and analyzed as a single group for downstream comparisons.

The 30 OR genes, including the tumor-upregulated OR51E1 and OR51E2, exhibited relatively high expression in Cluster 1 (Fig. S2e). Consistent with the previously observed elevated stromal scores in this cluster, the enrichment analysis revealed significant activation of pathways related to collagen organization and extracellular matrix (ECM) remodeling. Conversely, immune-associated pathways showed markedly negative enrichment scores (Fig. 3f). Clusters 3/5 were characterized by increased expression of six OR genes (OR52N4, OR2L13, OR52K3P, OR1H1P, OR2I1P, and OR13A1) (Fig. S2e). These clusters were previously defined by low stromal and high immune scores. In accordance with these features, the GSEA revealed a strong upregulation of a broad spectrum of immune functions (Fig. 3f).

Three OR genes (OR2I1P, OR9A4, and OR7A8P) were found to be overexpressed in Cluster 4 (Fig. S2e). Consistent with the intermediate levels of both stromal and immune scores for this cluster, this group of genes displayed enhanced activity in pathways related to the response to the ECM, as well as immune responses involving T/NK cells, interferon signaling, PD-1 signaling, major histocompatibility complex (MHC), and antigen presentation. However, unlike Clusters 3 and 5, Cluster 4 showed the downregulation of immune signaling associated with B cells, immunoglobulins, interleukins (particularly IL-10), complement activation, phagocytosis, Fc receptors, scavenger receptors, and complement receptors (Fig. 3f). Notably, differentially expressed OR genes were not detected in Cluster 2.

Compared with Cluster 4, B-reg- and macrophage-related pathways were specifically enriched in Clusters 3/5 [46–48]. Given the well-documented involvement of B regs and tumor-associated macrophages (TAMs) in mediating immunosuppression and contributing to poor cancer outcomes, the prognostic relevance of the identified clusters was subsequently investigated [49, 50]. Although differences in OS across all five clusters were not statistically significant, stratification into grouped categories—Cluster 3/5 versus all others or Cluster 3/5 versus Cluster 2/4 and Cluster 1—revealed that patients in Clusters 3/5 exhibited significantly worse survival outcomes (Fig. 3e). These findings suggest that the enrichment of B reg- and TAM-associated pathways in Clusters 3 and 5 may contribute to the adverse prognosis observed in these patients.

Hence, to determine whether the OR genes upregulated in Clusters 3 and 5 were functionally associated with immunosuppression, an immune-regulatory score was computed: (B reg score × M2 macrophage score / M1 macrophage score) × 100. The OR gene signature derived from Clusters 3/5 presented strong positive correlations with the immune-regulatory score, TAM, and B reg scores, respectively, suggesting a potential link between OR expression and immunosuppressive dynamics within the TME (Fig. 3g). Notably, this observation is consistent with earlier findings, which showed that OR13A1 was correlated with B reg activity (Fig. 2i). These results suggest that ORs enriched in Clusters 3 and 5 are closely linked to immunoregulatory cell types.

Collectively, these findings highlight that OR gene expression can serve as a robust basis for both TME-based classification and prognostic prediction in ccRCC patients.

Overall expression profiles of OR genes at a single-cell level

To identify cell-type-specific expression patterns of OR genes at the single-cell level, scRNA-seq was performed on tumor specimens from 13 patients with ccRCC. After rigorous quality control and filtering, 55,709 cells were retained for downstream analysis. These cells were clustered and visualized, leading to the identification and annotation of 16 distinct cell types within the TME (Fig. 4a). A total of 40 OR genes were detected among the tumor-derived cells. Notably, several OR genes with a relatively high expression frequency were predominantly observed in cancer cells, including OR2T2, OR2G6, OR2T10, OR2T11, and OR2A1-AS1, which belong to the OR2 gene family (Figs. 4b and S3a–d). In addition, OR51E1 and OR51E2 were found to have specific expression patterns in mesenchymal cell populations (Figs. 4b and S3e). OR genes, such as OR2B11, OR5H14, OR4D9, OR4D1, OR7C1, and OR2A25, were also detected in innate immune cells, including dendritic cells (DCs), monocytes, and T cells, suggesting a broader distribution of OR gene activity across multiple cell types within the TME (Figs. 4b and S3f–k).

Fig. 4 — Overall expression pattern of ORs at the single-cell level. a UMAP plot visualizing single-cell transcriptomic profiles of tumor tissues from 13 patients, integrating data from three independent GEO datasets. A total of 16 distinct cell types were annotated based on transcriptional signatures. b Dot plot of the expression pattern of OR genes across annotated cell types detected within the tumor samples. Dot size represents the proportion of cells expressing each gene, and color intensity reflects the average expression level of each gene. c UMAP plot of the single-cell transcriptomic landscapes of normal kidney tissues from 10 individuals, based on the integration of two independent GEO datasets. A total of 20 cell clusters were defined according to the transcriptional profiles. d Bar plot of the number of expressed OR genes between tumor and normal kidney single-cell datasets. A total of 13 OR gene families were detected, and the number of expressed genes in each family was quantified and visualized. e UMAP plot of the single-cell ATAC-seq data from 19 patients, with 25 cell types annotated using chromatin accessibility signatures. f Heatmap of the chromatin accessibility of OR genes with high scATAC-seq peak signals, stratified by cell type. Only OR genes with prominent accessibility peaks were included, and accessibility was compared across annotated cell populations

Given that cancer cells often arise from normal kidney epithelial cells, such as those in the proximal renal tubules, the expression patterns of OR genes in normal kidney tissues were examined to explore potential lineage relationships. As a result, OR2T10 expression was confined to proximal renal tubular cells in normal samples (Figs. 4c and S4a). Consistently, a strong positive correlation was observed between OR2T10 expression and the proximal tubule cell gene expression signature at the bulk transcriptomic level (Fig. S4b). These findings suggest that OR2T10 may represent a lineage-associated OR gene whose expression is maintained during the transition from normal to malignant epithelial states.

Similarly, OR2B11 and OR5H14, which were detected in macrophage and monocyte clusters in normal samples, were also consistently expressed in the corresponding monocyte and macrophage populations in the tumor samples (Fig. S4c–d). In addition, OR2A25, OR4D1, and OR4D9 were commonly expressed in T cells across both normal and tumor tissues, indicating a subset of OR genes for which the cell-type-specific expression is conserved regardless of disease state (Fig. S4e–g). In contrast, OR51E1 and OR51E2, which were prominently expressed in the mesenchymal cells in the TME, were not detected in any corresponding mesenchymal-related clusters of normal kidney tissues. This suggests that these ORs may represent tumor-associated genes, reflecting the cellular and metabolic changes that occur during malignant transformation or stromal remodeling within the tumor milieu.

When the repertoire of OR genes detected in the tumor and normal samples was compared, a notable expansion of OR gene expression was observed in the TME. In particular, more OR2 genes were detected in the tumor samples than in the normal samples. Similarly, OR genes from the OR3, OR51, and OR52 families were also more frequently observed in tumor cells (Fig. 4d). The elevated detection of OR51 and OR52 family genes at the single-cell level is consistent with the earlier observation (Fig. 1b), where these gene families exhibited higher family-related scores in tumors than in normal tissues. Overall, the number of detected OR genes was substantially higher in tumor tissues than in normal counterparts, suggesting that malignant transformation may be accompanied by broader transcriptional activation of OR gene families, potentially reflecting changes in cellular identity or microenvironmental cues.

To further validate the cell-type-specific expression of OR genes at the chromatin level, scATAC-seq was performed on tumor tissues from a distinct cohort of 19 ccRCC patients. After conducting the quality control, a total of 58,624 cells were visualized (Fig. 4e). The chromatin accessibility peaks associated with OR genes were investigated, and OR51E1 and OR51E2 were found to exhibit distinct accessibility peaks in the pericyte populations, which belong to mesenchymal stromal cells (MSCs) (Fig. 4f). In addition, OR genes, such as OR2T10, showed enriched chromatin accessibility in cancer cells, consistent with the associated expression profiles observed in the scRNA-seq data (Fig. 4f). These findings offer complementary epigenomic insights into the cell-type-specific patterns of OR genes, which are consistent with the observed transcriptional signatures in specific tumor-associated cellular components.

OR51E1 is specifically expressed in pericytes and is associated with angiogenesis

Among the OR genes consistently detected at both single-cell transcriptomic and chromatin accessibility levels, OR51E1 was selected as an initial focus to explore the potential contribution of this gene to tumor progression. Notably, OR51E1 expression was highly enriched in MSCs from tumor tissues and was nearly absent in other cell types within the TME (Fig. 5a). To further resolve the cellular specificity of OR51E1, sub-clustering was performed in MSCs, which demonstrated that OR51E1 expression was prominently localized to the pericyte population (Figs. 5b–c and S5a–b). Consistent with these findings, scATAC-seq analysis revealed a strong, cell-type-specific chromatin accessibility peak at the OR51E1 locus in pericytes, supporting active transcriptional regulation (Fig. 5d). Together, these results reveal that OR51E1 is selectively expressed in pericytes, suggesting a pericyte-specific expression profile that may be relevant to renal tumor biology.

Fig. 5 — *OR51E1* expression is restricted to pericytes and functionally linked to angiogenesis. a Feature plot of the cell type-specific expression of *OR51E1* in MSCs. b Sub-clustering of MSCs to annotate ten discrete cell populations based on transcriptional profiles. c Dot plot of the expression of canonical marker genes used to annotate MSC sub-clusters, including pericytes, fibroblasts, and SMCs. d Coverage plot of the chromatin accessibility peaks at the *OR51E1* locus across different cell types. Gene annotation and genomic coordinates on chromosome 11 are shown below. e Correlation analysis between *OR51E1* expression and pathway scores derived from 21,899 gene sets in MSigDB collections (C2, C5, and Hallmark). Among the top 50 pathways ranked by coefficient, pathways related to angiogenesis are highlighted. The p-values were adjusted using the Bonferroni correction method. f Heatmap of the angiogenesis-related pathway scores, previously identified at the bulk level, compared between *OR51E1*-positive and -negative pericytes at the single-cell level. Pathway activity was quantified using both the ssGSEA and AUCell methods. g Violin plots of the differential expression of pro-angiogenic growth factors between *OR51E1*-positive and negative pericytes at the single-cell level. Adjusted p-values were assessed using Student’s t-test. h TF footprints of *EBF2*, *NFAT5* and *RFX5*, with consistently elevated activity in *OR51E1*-positive pericytes across both ATAC-seq and scRNA-seq levels, compared to *OR51E1*-negative counterparts. These factors are among the top 10 enriched in *OR51E1*-active pericytes with minimal background noise. The Tn5 insertion bias tracks are included below

To investigate the potential biological functions of OR51E1 in pericytes within the TME, the association with 21,889 pathways was assessed in KIRC tumor samples. A correlation analysis between OR51E1 expression and pathway activity scores revealed a strong and significant positive correlation with pathways related to angiogenesis (Fig. 5e). Consistent with these observations, analyses at the single-cell level revealed that OR51E1 expression in pericytes was strongly associated with angiogenesis-related programs (Fig. 5f). Furthermore, DEG analyses at the single-cell transcriptomic and chromatin accessibility levels identified a set of concordantly upregulated genes in OR51E1 + pericytes. GSEA based on these shared DEGs consistently revealed significant enrichment of angiogenesis-related pathways (Fig. S5c).

To confirm functional receptor activity, we evaluated canonical GPCR downstream signatures. OR51E1 + pericytes showed significantly higher scores for MAPK and cAMP-mediated signaling compared to negative counterparts, supporting the biological plausibility of active OR51E1 signaling in pro-angiogenic regulation (Fig. S5e).

Therefore, to explore how OR51E1 is linked to angiogenesis, we aimed to investigate the expression of key pro-angiogenic factors, including ANGPT2, VEGFA, PDGFA, PDGFB, NGF, and PGF, at the single-cell level (Fig. 5g) [51–53]. While VEGFA expression did not differ substantially between OR51E1-positive and -negative pericytes, the remaining pro-angiogenic genes were consistently upregulated in OR51E1-expressing pericytes, suggesting that potential paracrine or autocrine mechanisms contribute to angiogenic signaling.

Next, we examined the differences in TF activity between OR51E1-positive and -negative pericytes, using both scRNA-seq and scATAC-seq data. Transcriptional activity was inferred at the single-cell transcriptomic level, while the accessibility of the TF binding motif was calculated using chromatin accessibility data. By intersecting the top 10 TF motifs with increased accessibility in OR51E1-active pericytes and TFs with elevated activity scores in the same population at the transcriptomic level, a shared set of transcriptional regulators was identified: RFX5, EBF2, and NFAT5 (Figs. 5h and S5d). Among these, NFAT5 has previously been implicated in promoting angiogenesis, supporting the potential contribution to the pro-angiogenic signaling observed in the OR51E1-expressing pericytes, whereas the role of RFX5 and EBF2 to tumor-associated angiogenesis remains unestablished [54, 55].

OR2 family genes have cancer cell-specific expression and are associated with invasion and metastasis

Following the analysis of OR51E1, the OR genes belonging to the OR2 gene family, which showed markedly higher detection in tumors compared to normal tissues, were examined to explore potential roles in tumor progression (Fig. 4d).

A chromatin accessibility analysis revealed increased peaks for several OR2 family members, particularly OR2T10, OR2T11, OR2G6, and OR2T2, within cancer cells (Fig. 4f). Interestingly, OR2T10 was robustly expressed in ccRCC cells and showed baseline expression in proximal tubule cells from normal kidney samples (Figs. 6a and S4a–b). In contrast, adjacent OR2 genes, such as OR2G6 and OR2T11, despite the genomic proximity of these genes to OR2T10, were undetectable in normal samples, but displayed selective expression in tumor-derived cancer cells from tumor tissues (Fig. S3a–c).

Fig. 6 — OR2 family genes exhibit cancer cell-specific expression and are implicated in the development of malignant phenotypes. a Feature plot of the cell type-specific expression of *OR51E1* in cancer cells. b Sub-clustering of cancer cell-1 clusters, resulting in the annotation of four discrete cell populations based on transcriptional profiles. c Coverage plot of the chromatin accessibility peaks at the *OR2T10* locus across different cell types. Gene annotation and genomic coordinates on chromosome 1 are shown below. d Hi-C contact heatmap (top) generated from 786-O ccRCC cells of the genomic region spanning Chr1: 247.0–249.0 Mb. The lower panels show predicted TADs, inferred using insulation score profiling. e Magnified view of the TAD regions encompassing the OR2 gene loci. f Enrichment map of the GSEA results of commonly upregulated genes in *OR2T10*-positive cancer cells, based on both scRNA-seq and scATAC-seq analyses. g A box plot of the ssGSEA-derived scores for cancer cell malignancy-associated pathways across cancer cell-1 sub-clusters. Scores were compared among *OR2T10*-positive ccRCC-1, *OR2T10*-negative ccRCC-1, ccRCC-2, 3, 4 subgroups within cancer cell-1 cluster. Statistical significance was assessed using Student’s t-test. h Violin plot of the transcriptional activity of *TFAM*, inferred at the single-cell level, across ccRCC-1 sub-clusters. i TF footprint of *TFAM*, the most enriched TF in *OR2T10*-positive cells at both the ATAC-seq and scRNA-seq levels. The Tn5 insertion bias tracks are located below

Consistent with the transcriptomic findings, open chromatin peaks were detected at the loci of all four OR2 genes, specifically in cancer cells (Fig. S6a–c). Among these genes, OR2T10 exhibited the strongest chromatin accessibility signal, supporting the prominent activation of this gene in the malignant population (Fig. 6b). While these OR2 genes have yet to be functionally characterized in cancer-related processes in prior literature, the selective expression and chromatin accessibility of these genes in cancer cells suggest potential tumor-specific regulatory mechanisms.

Given the high diversity of cancer cells shaped by their spatial context within the tumor and interactions with surrounding components of the TME, cancer cells were sub-clustered to identify the transcriptional subtypes that preferentially express these OR genes (Fig. 6c) [56]. This analysis identified four distinct cancer cell subtypes (ccRCC-1 to ccRCC-4); the OR2 family genes were specifically enriched in the ccRCC-1 subtype (Fig. S6e). Analysis of the DE between ccRCC-1 and other subtypes, followed by GO analysis, revealed that ccRCC-1 showed enhanced responses to chemical stimuli and enrichment of terms related to the negative regulation of programmed cell death and cell population proliferation, indicating functional divergence from other tumor cell clusters (Fig. S6d).

Thus, to explore the regulatory basis underlying the coordinated expression of genomically adjacent OR2 genes within the same cancer cell cluster, we hypothesized that the co-expression of these genes might be facilitated by shared chromatin topology, such as inclusion within a common topologically associated domain (TAD). Since genes within the same TAD are more likely to be co-regulated due to their spatial proximity and shared regulatory environment, we reasoned that the observed co-expression of OR2T10, OR2T11, OR2T2, and OR2G6 in the ccRCC-1 subtype may reflect such high-order chromatin architecture [57].

Accordingly, to investigate whether these OR2 genes are encompassed within a common TAD that may underlie their transcriptional coordination, we analyzed Hi-C data from the 786-O cell line to delineate TAD structures on chromosome 1. While all four OR2 genes were co-expressed within the same cancer sub-cluster, these genes were found to reside in separate TADs, specifically, OR2T2 and OR2G6 (transcribed in the 5’ to 3’ direction) formed one domain, while OR2T10 and OR2T11 (transcribed in the opposite direction) occupied another (Figs. 6d–e and S6f). To elucidate the chromatin context underlying the coordinated expression of the OR2 family genes, the compartment organization of these genes on chromosome 1 was examined. Chromosomal compartments, broadly categorized into transcriptionally active (A) and inactive (B) regions, provide an additional layer of spatial genome organization that influences gene regulation. As a result, all four OR2 genes, OR2T10, OR2T11, OR2T2, and OR2G6, were localized within the compartment A region, which is characteristically associated with open chromatin, high gene density, and an active transcriptional state (Fig. S6g) [58]. This finding suggests a potential link between the 3D genomic environment and the coordinated expression of these OR2 genes; specifically, their elevated expression in ccRCC cells appears to be associated with not only their TAD-level proximity but also their localization within a transcriptionally active compartment. Collectively, these results highlight a potential association between the OR2 family and ccRCC progression, which appears to be supported by a favorable three-dimensional (3D) genome architecture and an active regulatory environment.

Therefore, to delineate the potential involvement of the OR2 gene family—particularly OR2T10—in cancer progression, genes commonly upregulated in two distinct comparisons were identified: OR2T10-positive versus OR2T10-negative cancer cells at the single-cell (1) transcriptomic and (2) ATAC-levels. GSEA of the intersected upregulated genes revealed significant enrichment of pathways implicated in cancer progression, including drug resistance, EMT, and metastasis (Fig. 6f). Moreover, cancer progression-associated pathway scores were significantly higher in OR2T10-expressing cells than in the associated OR2T10-negative counterparts (Fig. S6g). Importantly, the increased scores for cAMP and MAPK signaling pathways in OR2T10 + cells further support the biological plausibility of active OR signaling in this malignancy-associated population (Fig. S6h).

Building on these findings, the potential contribution of OR2T10 and other OR2 family members to ccRCC progression were further investigated. Accordingly, WGCNA was performed using transcriptomic profiles of cancer cells (Fig. S6i). Consequently, co-expression modules indicated that genes co-clustered with OR2T10 and OR2G6 were predominantly associated with oxidative phosphorylation (OXPHOS) (Fig. S6j–k). Although the role of OXPHOS may vary depending on cancer type and TME context, this process has been implicated in supporting the energetic demands of invasive and metastatic cancer cells [59, 60]. These findings suggest a potential link between OR2 family genes and cancer progression, providing further evidence that these genes may contribute to the metabolic reprogramming associated with tumor invasiveness.

To characterize the regulatory architecture underlying OR2T10 activation, a comparative analysis of TF activity was conducted, using an approach identical to that for OR51E1. Through this integrative analysis, TFAM emerged as the most differentially active TF in OR2T10 + cells, consistent with its established roles in promoting metabolic reprogramming and metastasis in cancers [61, 62]. Hence, to evaluate whether TFAM binding is linked to OR2T10 expression, a targeted footprint analysis was conducted using the sequence of the mitochondrial LSP (light strand promoter)—previously characterized as a TFAM binding site [63]. Consistent with elevated TFAM activity, OR2T10 + cancer cells exhibited higher TFAM binding signals at the LSP motif compared to the contrasting cells, suggesting that TFAM activity is closely aligned with the distinct metabolic and transcriptional landscape of OR2T10 + cells.

Collectively, these findings suggest that OR2T10, along with the associated genomically adjacent OR2 family members, resides in transcriptionally active chromatin compartments and neighboring TADs, supporting the coordinated regulation of these members in ccRCC cells. The convergence of elevated malignancy-related pathway activity, enrichment in oxidative phosphorylation-associated gene networks, and increased TFAM binding and activity in OR2T10 + cells provides biological implications for the potential contribution of OR2T10 to ccRCC progression.

Characterization of cell-to-cell communication networks associated with OR gene expression

Given the tumor-associated expression patterns of OR51E1 and OR2T10, we investigated the potential associations between these genes and the intercellular communication networks within the TME. First, pericytes were stratified by OR51E1 expression, as previously defined, to investigate the dynamics of intercellular interactions. Given the angiogenic features associated with OR51E1, subsequent analyses focused on their roles as angiogenic signal-sending cells (senders) within the TME. As a result, OR51E1 + pericytes exhibited the highest number of predicted interactions with endothelial cells (ECs), followed by fibroblasts, and cancer cells (Fig. 7a–b). Interestingly, although OR51E1-negative pericytes were more abundant, OR51E1-positive pericytes still demonstrated a greater number of EC-directed interactions as senders, suggesting that OR51E1 expression is characterized by the pro-angiogenic signaling capacity and linked to tumor-associated vascular remodeling.

Fig. 7 — Characterization of cell-to-cell communication networks associated with OR gene expression. a Heatmap of the differential interaction counts in cell–cell communication networks, stratified by *OR51E1* expression status in pericytes. b Bar plot of the number of inferred cell–cell interactions mediated by *OR51E1*-positive and -negative pericytes acting as senders within the communication network. c Dot plot of the top 10 predicted ligand–receptor interactions (excluding ECM–receptor interactions) between *OR51E1*-positive pericytes (as sender) and ECs, ranked by communication probability. d Heatmap of the centrality scores of *OR51E1*-positive and -negative pericytes in ANGPT (left) and VEGF (right) intercellular communication networks, highlighting their roles as senders, receivers, mediators, and influencers. e Dot plot of the ligand–receptor interactions with the ECM–receptor interactions removed, uniquely detected in *OR2T10*-positive cells acting as the sender, but not in *OR2T10*-negative cells. f Dot plot of the ligand–receptor interactions (omitting the ECM–receptor interactions) exclusively detected in *OR2T10*-positive cells serving as the receiver, but not in *OR2T10*-negative cells. g Inferred outgoing communication patterns of mesenchymal stromal cells (pericytes, fibroblasts, and vSMCs), ECs, and ccRCC-1 cells, illustrating the alignment of inferred signaling programs, cell groups, and relevant signaling pathways

To clarify these findings, we next focused on interactions in which ECs were designated as receivers. Excluding ECM-receptor interactions, the top 10 ligand–receptor pairs with the highest communication probabilities specific to OR51E1 + pericytes were prioritized (Fig. 7c). This analysis revealed a higher likelihood of signaling pathways known to promote angiogenesis—including PGF and ANGPT signaling—than those originating from negative counterparts. In addition, OR51E1 + pericytes consistently exhibited greater importance as senders within angiogenic pathways, including ANGPT, VEGF, and NOTCH signaling axes, compared to their contrasting cells (Figs. 7d and S7a) [64]. These findings collectively highlight the potential involvement of OR51E1 in pericyte-mediated pro-angiogenic communication within the TME.

Following stratification of cancer cells by OR2T10 expression, we examined whether OR2T10 + cells displayed distinct intercellular communication patterns within the TME. Given the association of OR2T10 with malignant features, subsequent analyses focused on evaluating roles of OR2T10 + cells as both senders and receivers of tumor-promoting signals. These analyses focused on OR2T10-expressing cells, which were predominantly found in the ccRCC-1 subtype. Accordingly, stratification based on OR2T10 expression was applied only within the ccRCC-1 sub-cluster from cancer cell-1, while the remaining cancer cell subtypes (ccRCC-2, 3, and 4) were analyzed without further division.

Among the OR2T10 + cells acting as senders, the most frequent interactions were observed with ECs, as well as with OR2T10-positive and -negative cells. In contrast, OR2T10-negative cells predominantly interacted with ECs, OR2T10 + cells, and DCs when serving as senders. As receivers, OR2T10-positive and -negative cells showed similar interaction profiles, which represent enriched communication with other cancer cells (ccRCC-1 and ccRCC-2) and monocytes (Fig. S7b–c).

To elucidate the biological implication of OR2T10 in cancer cells, the distinct intercellular interactions were investigated. When serving as receivers, OR2T10 + cells exhibited an elevated responsiveness to ligands involved in oncogenic signaling cascades, including epidermal growth factor (EGF), macrophage migration inhibition factor (MIF), and midkine (MDK) (Fig. 7e) [65–67]. In addition, as senders, these cells prominently engaged in signaling pathways, such as fibroblast growth factor (FGF), hepatocyte growth factor (HGF), and EGF; all of which are closely associated with tumor progression, proliferation, and invasiveness (Fig. 7f) [65, 68, 69]. In line with the previous analyses, OR2T10 + cells consistently exhibited greater centrality as key signal senders in the HGF and FGF pathways, and as principal signal receivers in the EGF, FGF, and MDK pathways, compared to their negative counterparts (Fig. S7d–e). These findings suggest that OR2T10 expression is associated with an elevated capacity for malignant signaling integration and dissemination, potentially representing the aggressive tumorigenic properties of cancer cells in a subtype-specific manner.

Subsequently, the analysis was further expanded to investigate whether the expression of OR51E1 and OR2T10 could serve as markers for distinct cellular subpopulations based on their intercellular communication dynamics. Within this framework, stratification of pericytes by OR51E1 expression revealed clearly separable communication patterns, particularly in their roles as signal senders (Figs. 7g and S7f). Among the outgoing signaling programs enriched in OR51E1 + pericytes, three distinct patterns—including dehydroepiandrosterone sulfate (DHEAS) and WNT signaling—were identified, both of which are known to play pro-angiogenic roles within the TME [70, 71]. These findings are consistent with earlier observations indicating the heightened angiogenic potential of OR51E1 + pericytes.

Stratification of cancer cells by OR2T10 expression revealed distinct outgoing and incoming signaling profiles within TME. In the outgoing communication landscape, OR2T10 + cells contributed more prominently to signaling pattern characterized by enhanced HGF and FGF pathways. Similarly, incoming signaling profiles revealed OR2T10 + cells were preferentially targeted by a malignancy-associated communication pattern marked by FGF, HGF, and oncostatin M (OSM) signaling [72]. These findings suggest that OR2T10 expression delineates cancer cell subpopulations engaged in pro-tumorigenic communication circuits within the TME.

Collectively, these results highlights that OR51E1 and OR2T10 delineate distinct signaling landscapes that align with the respective roles of the genes in angiogenesis and malignancy, respectively.

Drug response profiles associated with OR gene expression

Building upon these observations, the potential associations between OR51E1 and OR2T10 and therapeutic responsiveness in ccRCC were further explored to evaluate the clinical relevance of these putative drivers in tumor progression.

Although OR51E1 expression was primarily localized to pericytes, a minor population of cancer cells also exhibited detectable expression. The observed association between elevated OR51E1 levels and increased resistance to bortezomib and bryostatin-1 suggests that OR51E1 expression in malignant cells may contribute to drug tolerance (Fig. 8a, c). Moreover, reduced sensitivity to compounds targeting genome integrity, chromatin modification, protein stability, and degradation further supports a role for OR51E1 in promoting therapeutic resistance through multiple cellular mechanisms (Fig. 8d). Although the functional characterization of OR51E1 in this study primarily focused on the pro-angiogenic role of OR51E1 in pericytes, these findings collectively indicate that ectopic expression of OR51E1 in cancer cells may additionally facilitate treatment resistance, thereby reinforcing the contribution to ccRCC progression.

Likewise, the cancer cell-specific expression of OR2T10 was associated with broad resistance to numerous anticancer agents (Fig. 8b). Such resistance was particularly evident for drugs targeting pathways fundamental to tumor proliferation and survival, including proliferation, DNA replication, apoptosis regulation, genome integrity, and angiogenesis (Fig.8e–f). This observation is consistent with the previous result, which linked OR2T10 to enhanced resistance of multiple drugs (Fig. 6g). These findings suggest that OR2T10 expression may confer both metabolic adaptability and therapeutic resistance, underscoring its potential as a novel molecular target in ccRCC.

Discussion

This study comprehensively investigated the clinical and biological relevance of ectopically expressed OR genes to tumor progression in ccRCC by integrating multi-omics datasets, including bulk RNA-seq, scRNA-seq, scATAC-seq, and Hi-C data.

Across multiple analyses at bulk transcriptomic level, a subset of OR genes consistently showed tumor-associated upregulation and was linked to aggressive tumor phenotypes, including invasion, metastasis, and immune evasion (Fig. 9). Among the identified upregulated genes found in tumors, several have been previously reported to exert non-olfactory biological functions. For instance, OR2A4 is expressed in suprabasal keratinocytes and basal melanocytes of the epidermis, where this protein modulates key cellular processes including cytokinesis and proliferation [73]. Likewise, OR7E47P is widely expressed in lung tissue and functions as a positive regulator in the TME of lung adenocarcinoma (LUAD) [74]. Conversely, four OR genes (OR7E102P, OR7E14P, OR7E91P, and OR10AB1P) were identified as significantly downregulated in tumor samples. Notably, OR7E14P has been described as a favorable prognostic marker in polycystic ovary syndrome and ovarian cancer [75]; meanwhile, OR10AB1P was found to be downregulated in neuroblastoma patients with poor survival outcomes compared to long-term survivors [76]. These observations support the notion that certain OR genes, despite their olfactory origin, may play functionally significant roles across and may represent prognostic biomarkers and druggable therapeutic targets across multiple cancer types.

Fig. 9 — Summary of the context- oncogenic roles of ectopically expressed ORs in ccRCC. This schematic summarizes the functional heterogeneity of ectopically expressed ORs, characterized by integrated multi-omics analyses. Prognostic ORs (top left): A predictive model identified *OR2B6*, *OR11H7*, and *OR13A1* as risk-associated genes, where *OR2B6* and *OR11H7* correlate with malignant gene signatures, while *OR13A1* is linked to the B reg signature. Immunoregulatory ORs (top right): Consensus clustering and GSEA of immune-related features identified a subset of ORs enriched in regulatory immune signatures. Pro-angiogenic OR (bottom left): *OR51E1* was specifically expressed in pericytes and associated with the upregulation of pro-angiogenic growth factors, suggesting a role in vascular remodeling within the TME. Malignancy-enhancing OR (bottom right): *OR2T10* was predominantly expressed in a highly malignant cancer cell subtype, promoting mitochondrial *TFAM* activity and oxidative phosphorylation. *OR2T10* + cells exhibited cell–cell communication signatures linked to increased tumor aggressiveness. Created with BioRender.com.

In particular, several of these ORs were further associated with unfavorable clinical outcomes, indicating their potential relevance to prognosis of ccRCC. Among these, OR2B6, OR11H7, and OR13A1 emerged as independent prognostic markers, each correlated with poor survival in ccRCC patients (Fig. 9). Notably, OR2B6 has been previously studied in breast cancer, where its expression was linked to tumor progression [16, 77]. The consistency of these observations across cancer types supports the role of OR2B6 as a potential oncogenic factor in ccRCC. Similarly, the expression of OR11H7 was reported in human proximal tubule cell lines and activated by short-chain fatty acids (SCFAs), suggesting that OR11H7 upregulation during malignant transformation may be associated with tumor aggressiveness through metabolic reprogramming [78]. In contrast, OR13A1 has yet to be characterized in the context of cancer. Hence, the identification of OR13A1 in this study, to our knowledge, highlights a novel candidate OR gene with potential oncogenic relevance in ccRCC, warranting further experimental and clinical cohort validation to elucidate the function and regulatory mechanisms of this gene. Importantly, the association between OR13A1 and B regs aligns with observations from consensus clustering, which classified OR13A1 within an immunoregulatory OR gene subset, further supporting the potential role of this gene in modulating the tumor immune microenvironment.

We further clarified cell-type-specific functions of ectopically expressed ORs in ccRCC TME. In this context, OR51E1 emerged as a stromal cell–restricted OR, with expression predominantly confined to pericytes rather than malignant cells (Fig. 9). This finding highlights a previously underappreciated role of ORs in non-cancerous components of the tumor niche. Rather than directly influencing cancer cell–intrinsic programs, OR51E1 expression was associated with transcriptional and regulatory features linked to angiogenesis, suggesting that this receptor may contribute to tumor progression indirectly by modulating the vascular microenvironment.

Consistent with this interpretation, NFAT5 was identified as a transcriptional regulator exhibiting consistently elevated activity in OR51E1 + pericytes at both single-cell transcriptomic and ATAC-level. NFAT5 has been previously implicated in the promotion of angiogenesis, and notably, subsequent intercellular communication analyses further revealed that WNT signaling was selectively enriched in OR51E1 + pericytes compared to their negative counterparts. Given that NFAT transcription factors have been shown to activate WNT signaling-related genes, these results collectively support the potential involvement of OR51E1 + pericytes in pro-angiogenic processes [70].

In addition, RFX5 and EBF2 also showed increased activity in OR51E1 + pericytes. Although their direct involvement in tumor-associated angiogenesis remains unestablished, previous studies have demonstrated that RFX5 can upregulate JAG1 expression, a Notch ligand known to promote angiogenesis and tumor vascular development, thereby suggesting a potential indirect contribution of RFX5 to pro-angiogenic programs through the JAG1–Notch axis [64, 79]. Consistent with this notion, JAG1 expression was higher in OR51E1 + pericytes than in their negative counterparts, further supporting the involvement of this axis in the angiogenic programs associated with OR51E1 expression (Fig. S5f). Likewise, EBF2 has yet to be directly linked to the regulation of angiogenesis. Nonetheless, prior research has implicated EBF2 in metastatic behavior and resistance to apoptosis in osteosarcoma, suggesting a broader role in tumor biology [80].

Although OR51E1 expression was largely restricted to pericytes, low-level expression was also detectable in a subset of ccRCC cells. In this context, drug response analyses in ccRCC cell lines revealed that OR51E1 expression is associated with increased resistance to bortezomib and byrostatin-1 significantly. Given that bortezomib induces apoptosis by inhibiting proteasomal degradation of key regulatory proteins and bryostatin-1 modulates PKC-mediated signaling pathways involved in cell growth, differentiation, and chemosensitization—enhancing the cytotoxic efficacy of conventional chemotherapeutic agents such as cisplatin—resistance to these agents suggests that OR51E1 expression may contribute to ccRCC progression [81].

OR51E1 is one of the few ORs that has been studied previously in various cancer types. Indeed, the expression of OR51E1 has been reported in small-intestine neuroendocrine carcinomas, where OR51E1 was proposed as a potential tissue biomarker [82]. In addition, OR51E1 overexpression has also been observed in colorectal cancer and early-stage gastric cancer [83]. In prostate cancer, both OR51E1 and its close homolog OR51E2 have been shown to suppress tumor cell proliferation and promote cell death, underscoring the context-dependent functionality of these receptors [84]. In contrast, our findings suggest that in ccRCC, OR51E1 may contribute to tumor progression through its pro-angiogenic activities, warranting further investigation into the functional role of OR51E1 in the renal tumor vasculature.

Similarly, OR2T10 was identified as a cancer cell–specific OR associated with an aggressive malignant phenotype in ccRCC (Fig. 9). Notably, during malignant transformation, epigenetic reorganization of chromatin architecture may permit OR2T10 and adjacent OR2 family genes to become repositioned into transcriptionally active A compartments, which is associated with their coordinated upregulation.

OR2T10 expression was linked to metabolic programs that support tumor progression, particularly those related to mitochondrial energy production via OXPHOS. This association is particularly noteworthy given that renal cell carcinomas frequently exhibit metabolic heterogeneity, often segregating based on mitochondrial DNA content and mTOR-dependent metabolic rewiring [85]. In this context, the correlation between OR2T10 expression and OXPHOS programs may indicate a specific metabolic state rather than a direct signaling function alone. Interpreting these receptors as markers of mitochondrial metabolic adaptation across renal cancer subtypes allows for the integration of our observations into a broader biological framework of ccRCC metabolism.

TFAM emerged as a key transcriptional regulator with elevated activity in OR2T10 + cells, consistent across both at the single-cell transcriptomic and ATAC-level. Given that TFAM has been implicated in pro-tumorigenic processes such as metabolic reprogramming and metastasis, its heightened activity alongside the enrichment of OXPHOS signatures in OR2T10 + cells suggests that the expression of OR2T10 may be closely linked to aggressive tumor behaviors and reflects a specific state of metabolic adaptation. [59–62]. Together, these findings suggest that ectopically expressed ORs may contribute to tumor progression not only through the modulation of stromal components but also by being closely associated with the reorganization of malignant cell–intrinsic programs.

Consistent with the malignant features associated with OR2T10, elevated OR2T10 expression was also associated with increased resistance to multiple therapeutic agents significantly in ccRCC cell lines. Notably, these drug target core cancer hallmarks, including proliferation, apoptosis regulation, and DNA regulation. The observed multi-drug resistance pattern suggests that OR2T10 may contribute not only metastatic potential but also to multiple facets of tumor aggressiveness, further underscoring its role as a key feature associated with malignant cell–intrinsic programs in ccRCC.

While the involvement of OR2T2, OR2T11, and OR2G6 in cancer biology has not been described, prior studies have implicated OR2T10 in non-small cell lung cancer (NSCLC), where genetic alterations in OR2T10 locus were associated with metastatic potential [86]. The concordance between these reports and the present findings supports a broader role for OR2T10 as a pro-tumorigenic OR across cancer types. Collectively, these observations position OR2T10, and potentially adjacent homologous OR2 family members, as key molecular features linked to cancer cell metabolism and malignancy in ccRCC, underscoring the need for further mechanistic and experimental investigation.

Intercellular communication analyses further validated the tumor-associated functions of OR51E1 and OR2T10, revealing their involvement in distinct pro-tumorigenic signaling networks within the TME. Together, these findings suggest that ectopically expressed ORs are embedded within cell type–specific signaling networks associated with tumor progression. In particular, OR51E1 expression in pericytes was linked to pro-angiogenic communication toward ECs, lending further support to its potential involvement in an angiogenic signaling milieu of the TME. Likewise, OR2T10 expression delineated a subset of cancer cells preferentially engaged in tumor promoting signaling interactions, consistent with its association with aggressive malignant states. Collectively, these findings suggest that OR-associated signaling extends beyond gene expression changes and is integrally linked to pro-tumorigenic intercellular communication within ccRCC.

A major limitation of the present study is the lack of experimental validation using in vitro or in vivo biological assays; consequently, the specific ligands and downstream signaling pathways of ectopically expressed ORs remain largely uncharacterized. Furthermore, while our multi-omics integration provides a strong foundation, the inferred regulatory programs and intercellular communications rely on in silico predictions, which may not fully capture the complexity of actual signaling dynamics. Specifically, whether these ORs function as putative drivers or serve as high-fidelity metabolic markers reflecting specific mitochondrial states, as suggested in recent studies, requires definitive functional characterization [85]. Nevertheless, to mitigate this limitation and enhance the robustness of our findings, we employed a comprehensive multi-omics approach by integrating analyses across multiple layers of data, including bulk and scRNA-seq, scATAC-seq, and Hi-C datasets. This integrative strategy enabled the systematic evaluation of OR-associated regulatory programs in ccRCC across transcriptional, epigenetic, and chromatin structural dimensions, providing a strong foundation for future functional and mechanistic investigations. In addition, our findings highlight ectopically expressed OR genes as potential biomarkers and therapeutic targets in ccRCC, with clear clinical relevance in ccRCC and potential pan-cancer utility, warranting prospective validation.

Conclusions

In summary, this study highlights the potential involvement of ectopically expressed ORs in ccRCC progression through distinct yet complementary associations, encompassing stromal-mediated angiogenic remodeling and cancer cell–intrinsic metabolic adaptation. Although the present work focuses on ccRCC, the recurrent identification of ORs across diverse cancer types suggests that these signaling signatures may reflect broader regulatory patterns in other malignancies. These findings position ectopically expressed ORs as promising candidates for biomarker development and therapeutic targeting—especially in tumors characterized by aberrant tumor–stroma interactions or specific metabolic dependencies—warranting validation in prospective clinical cohorts and functional investigation of OR-directed therapeutic strategies.

Supplementary Information

Below is the link to the electronic supplementary material.

10238_2026_2118_MOESM1_ESM.pdf^{(1MB, pdf)}

Supplementary file1 Fig. S1 Construction of a prognostic model using the OR gene expression. a Bar plot of LASSO selection frequency of candidate prognostic OR genes identified through 1,000 bootstrap iterations. b Forest plot of multivariate Cox regression including three OR genes filtered by LASSO regression, alongside age, AJCC pathological stage, gender, and tumor purity as covariates. c Hierarchical clustering dendrogram of genes from KIRC tumor samples based on pairwise topological overlap, constructed using a WGCNA. d Sample clustering dendrogram used for outlier detection prior to WGCNA. Based on the clustering dendrogram, one tumor sample with distinct expression patterns were classified as outliers (highlighted in red) and excluded from downstream network construction. e–f The scale-free topology fit index and mean connectivity across soft-thresholding powers (β). The optimal β was selected based on the criterion of achieving a scale-free R² of 0.85. The left panel shows the relationship between β and the fit index, and the right panel displays β versus mean connectivity. g–h Dot plot of the top 30 enriched pathways identified by GSEA on genes from the OR-correlated co-expression modules. g OR2B6 and OR13A1-correlated modules (MEpink); h OR11H7-correlated modules (MEblue)

10238_2026_2118_MOESM2_ESM.pdf^{(3MB, pdf)}

Supplementary file2 Fig. S2 Classification of tumor samples with OR gene expression by consensus clustering. a Box plot of the estimated fractions of CD8⁺ T cells (left) and CD4⁺ T cells (right) across the five consensus clusters assessed by EPIC. Statistical significance between clusters was calculated using Student’s t-test. b–d GSEA to compare Cluster 3 with other clusters. b Top three pathways with the highest positive NES scores in Cluster 3 versus Cluster 5. c Bottom three pathways with the most negative NES scores in Cluster 3 versus Cluster 5. d The top three upregulated pathways in Cluster 3 compared to Clusters 1,2,3, and 5. e Heatmap of the expression patterns of the upregulated ORs in each consensus cluster. Genes are grouped by cluster-specific upregulation from the DEG analysis, and samples are ordered by assigned cluster identity. Upregulated OR genes were defined based on a logFC > 0.3 and an adjusted p-value < 0.05

10238_2026_2118_MOESM3_ESM.pdf^{(174.1KB, pdf)}

Supplementary file3 Fig. S3 Expression of OR genes at the single-cell level from tumor samples. a–k Feature plot of the transcriptomic data for the expression of selected OR genes across the single-cell tumor samples. Each panel represents the expression of an individual gene: a OR2T11, b OR2T2, c OR2G6, d OR2A1-AS1, e OR51E2, f OR7C1, g OR2B11, h OR4D9, i OR5H14, j OR2A25, k OR4D1

10238_2026_2118_MOESM4_ESM.pdf^{(195.6KB, pdf)}

Supplementary file4 Fig. S4 Expression of the OR genes from normal kidney samples at the single-cell level. a Feature plot of OR2T10 expression in normal samples. b Correlation plot of the relationship between the score of proximal tubule cells and OR2T10 expression (log2 FPKM) based on the Pearson correlation analysis. The p-value was derived from Student’s t-test. c–g Feature plot of the expression of selected OR genes across single-cell normal kidney samples transcriptomic data. Each panel represents the expression of an individual gene: c OR2B11, d OR5H14, e OR4D1, f OR2A25, g OR4D9

10238_2026_2118_MOESM5_ESM.pdf^{(415.2KB, pdf)}

Supplementary file5 Fig. S5 OR51E1 expression is restricted to pericytes and functionally linked to angiogenesis. a Violin plot and b feature plot of the pericyte-specific expression of OR51E1 in tumor samples transcriptomic data. c Enrichment map of the GSEA results of commonly upregulated genes in OR51E1+ pericytes, based on both scRNA-seq and scATAC-seq levels. d Violin plots of the inferred activity scores of EBF2, NFAT5, and RFX5 in OR51E1-positive versus OR51E1-negative pericytes at the single-cell transcriptomic level. e Box plots of the scores of GPCR downstream signaling pathways (MAPK and cAMP) in pericytes according to OR51E1 expression levels. f Violin plot of the expression of JAG1 in OR51E1-positive and OR51E1-negative pericytes at the single-cell transcriptomic level

10238_2026_2118_MOESM6_ESM.pdf^{(4.4MB, pdf)}

Supplementary file6 Fig. S6 OR2 family genes exhibit cancer cell-specific expression and are implicated in the development of malignant phenotypes. a–c Coverage plot of selected OR2 family genes across scATAC-seq data. Each panel represents the signals of open chromatin accessibility for individual genes: a) OR2T2, b) OR2G6, and c) OR2T11. d Bar plot of the results for Gene Ontology enrichment using genes upregulated in the ccRCC-1 subtype compared to ccRCC-2, 3, 4. Displayed pathways were filtered for significance with an adjusted p-value < 0.05 and –log10 (adjusted p-values) ≥ 5. e Feature plots of the expression of OR2 family genes (OR2T10, OR2T11, OR2T2, and OR2G6) in sub-clustered cancer cell-1. f Hi-C contact heatmap generated from 786-O ccRCC cells, visualizing the genomic region spanning Chr1: 247.0–249.0 Mb. Blue dots on the heatmap represent TADs identified using the TopDom algorithm. g Hi-C correlation matrix of a 100 Mb region on chromosome in 786-O ccRCC cells. The matrix represents the pairwise Pearson correlations between interaction profiles of genomic bins, each 500 kb in size. Color intensity reflects the degree of similarity between interaction profiles of genomic bins. This correlation matrix was used to infer chromatin compartments, and the highlighted region indicates that the genomic loci harboring OR2 family genes fall within compartment A. h Box plots of the scores of GPCR downstream signaling pathways (MAPK and cAMP) in cancer cells according to OR2T10 expression levels. i Hierarchical clustering dendrogram of genes from ccRCC cancer cells at the single-cell transcriptomic level, constructed using hdWGCNA. j–k Enrichment graph and table with a summary of the fgsea pathway enrichment results based on gene ranking by eigengene-based connectivity (kME) within modules containing j OR2T10 and k OR2G6

10238_2026_2118_MOESM7_ESM.pdf^{(412.8KB, pdf)}

Supplementary file7 Fig. S7 Characterization of cell-to-cell communication networks associated with OR gene expression. a A heatmap of the centrality scores of OR51E1-positive and -negative pericytes in NOTCH intercellular communication networks, highlighting the roles of these pericytes as senders, receivers, and influencers. b Bar plot of the number of inferred cell–cell interactions initiated by OR2T10-positive and -negative cancer cell-1 subpopulations acting as the sender. c Bar plot of the total number of inferred incoming signals targeting OR2T10-positive versus -negative cancer cell-1 subtypes. d–e Heatmap of the centrality scores of intercellular communication networks specific to OR2T10-positive and -negative cells. d Signaling patterns specifically detected in OR2T10-positive cells as senders. e Pathways uniquely enriched in OR2T10-positive cells as receivers. f A heatmap of the inferred outgoing communication patterns, with the pattern usage across cell types, including pericytes stratified by OR51E1 expression, fibroblasts, vSMCs, ECs, and cancer cells (left) and associated signaling pathways for each pattern (right). g–h Heatmap of the outgoing and incoming communication patterns. g Outgoing signaling patterns across diverse cell types, including ccRCC cancer cell subtypes stratified by OR2T10 expression, monocytes, ECs, pericytes, and γδ T cells (left), with corresponding pathways for each pattern (right). h Incoming communication patterns mapped across ccRCC subsets, including ccRCC cell subsets stratified by OR2T10 expression, DCs, pericytes, monocytes, and fibroblast (left), alongside linked signaling pathways (right)

Supplementary file8^{(11.5KB, xlsx)}

Supplementary file9^{(9.3KB, xlsx)}

Acknowledgements

Not applicable.

Abbreviations

AJCC: American joint committee on cancer
ATAC-seq: Assay for transposase-accessible chromatin sequencing
AUC: Area under the curve
BCR: B cell receptor
B reg: Regulatory B cell
ccRCC: Clear cell renal cell carcinoma
DC: Dendritic cell
DE: Differential expression
DEG: Differentially expressed gene
DHEAS: Dehydroepiandrosterone sulfate
EC: Endothelial cell
ECM: Extracellular matrix
EGF: Epidermal growth factor
EMT: Epithelial-to-mesenchymal transition
FGF: Fibroblast growth factor
FPKM: Fragments per kilobase of transcript per million mapped reads
GDSC: Genomics of drug sensitivity in cancer
GEO: Gene expression omnibus
GSEA: Gene set enrichment analysis
GPCR: G-protein coupled receptor
GO: Gene ontology
HGF: Hepatocyte growth factor
Hi-C: High-throughput chromosome conformation capture
KIRC: Kidney renal clear cell carcinoma
LASSO: Least absolute shrinkage and selection operator
LPA: Lysophosphatidic acid
LSP: Light strand promoter
LUAD: Lung adenocarcinoma
MAD: Median absolute deviation
MDK: Midkine
MHC: Major histocompatibility complex
MIF: Macrophage migration inhibitory factor
MSC: Mesenchymal stromal cell
NES: Normalized enrichment score
NSCLC: Non-small cell lung cancer
OR: Olfactory receptor
OS: Overall survival
OSM: Oncostatin M
OSN: Olfactory sensory neuron
OXPHOS: Oxidative phosphorylation
QC: Quality control
ROC: Receiver operating characteristic
SCFA: Short-chain fatty acid
scATAC-seq: Single-cell assay for transposase-accessible chromatin sequencing
scRNA-seq: Single-cell RNA sequencing
ssGSEA: Single-sample gene set enrichment analysis
SVD: Singular value decomposition
TAD: Topologically associating domain
TAM: Tumor-associated macrophage
TCGA: The cancer genome atlas
TCR: T cell receptor
TF: Transcription factor
TF-IDF: Term frequency-inverse document frequency
TME: Tumor microenvironment
TNBC: Triple-negative breast cancer
TSS: Transcription start site
UMI: Unique molecular identifier
VST: Variance stabilizing transformation
WGCNA: Weighted gene co-expression network analysis

Authors contributions

D.J.Y. conceptualized the study, performed bioinformatics analyses, interpreted results, and wrote the first draft. H.J.C. supervised the project, acquired funding, validated analyses, and revised the manuscript critically for important intellectual content. Both authors had full access to all data, approved the final version, and agree to be accountable for all aspects of the work. D.J.Y. verified the underlying data.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS-2023–00209741) and a Korean Fund for Regenerative Medicine (KFRM) grant funded by the Korean government (the Ministry of Science and ICT, the Ministry of Health & Welfare) (grant number: 23A0105L1).

Data availability

This study used only publicly available datasets and did not generate any new individual participant data. Bulk RNA-seq data were obtained from The Cancer Genome Atlas (TCGA), and single-cell RNA-seq and ATAC-seq datasets were retrieved from publicly accessible repositories including GEO and ArrayExpress, as specified in the Methods section. These datasets represent the minimal data required to interpret, reproduce, and build upon the findings of this study. Custom R code for OR gene filtering and consensus clustering of bulk RNA-seq data (ConsensusClusterPlus) is publicly available via Zenodo (https://doi.org/10.5281/zenodo.18296557). All remaining analyses were performed using standard pipeline described in Methods.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

Not applicable. This study is a bioinformatics re-analysis of de-identified, publicly available gene expression data from GEO datasets. No new human or animal subjects were involved, and all original studies obtained appropriate institutional review board approvals.

Consent to participate

Not applicable. This bioinformatics study utilized de-identified public datasets from GEO database.

Consent to publish

Not applicable. No individual patient data or identifiable information was included.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1.Clear cell renal cell carcinoma - National Cancer Institute. 2025 [cited 2025 June 20]; Available from: https://www.cancer.gov/pediatric-adult-rare-tumor/rare-tumors/rare-kidney-tumors/clear-cell-renal-cell-carcinoma.
2.Manley BJ, et al. Integration of recurrent somatic mutations with clinical outcomes: a pooled analysis of 1049 patients with clear cell renal cell carcinoma. Eur Urol Focus. 2017;3(4–5):421–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Drabkin HA, Gemmill RM. Cholesterol and the development of clear-cell renal carcinoma. Curr Opin Pharmacol. 2012;12(6):742–50. [DOI] [PubMed] [Google Scholar]
4.Kidney cancer statistics. [cited 2025 June 20]; Available from: https://www.wcrf.org/preventing-cancer/cancer-statistics/kidney-cancer-statistics/.
5.Siegel RL, et al. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17–48. [DOI] [PubMed] [Google Scholar]
6.Flegel C, et al. Expression profile of ectopic olfactory receptors determined by deep sequencing. PLoS ONE. 2013;8(2):e55368. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Barnes IHA, et al. Expert curation of the human and mouse olfactory receptor gene repertoires identifies conserved coding regions split across two exons. BMC Genomics. 2020;21(1):196. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Armelin-Correa LM, et al. Nuclear compartmentalization of odorant receptor genes. Proc Natl Acad Sci U S A. 2014;111(7):2782–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Maßberg D, Hatt H. Human olfactory receptors: Novel cellular functions outside of the nose. Physiol Rev. 2018;98(3):1739–63. [DOI] [PubMed] [Google Scholar]
10.Ansoleaga B, et al. Dysregulation of brain olfactory and taste receptors in AD, PSP and CJD, and AD-related model. Neuroscience. 2013;248:369–82. [DOI] [PubMed] [Google Scholar]
11.Garcia-Esparcia P, et al. Functional genomics reveals dysregulation of cortical olfactory receptors in Parkinson disease: Novel putative chemoreceptors in the human brain. J Neuropathol Exp Neurol. 2013;72(6):524–39. [DOI] [PubMed] [Google Scholar]
12.Chaudhary PK, Kim S. An insight into GPCR and G-Proteins as cancer drivers. Cells. 2021. 10.3390/cells10123288. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Balijepalli P, Sitton CC, Meier KE. Lysophosphatidic acid signaling in cancer cells: What makes LPA so special? Cells. 2021. 10.3390/cells10082059. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Chung C, et al. Odorant receptors in cancer. BMB Rep. 2022;55(2):72–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Morita R, et al. Olfactory Receptor Family 7 Subfamily C Member 1 is a novel marker of Colon cancer-initiating cells and is a potent target of immunotherapy. Clin Cancer Res. 2016;22(13):3298–309. [DOI] [PubMed] [Google Scholar]
16.Masjedi S, Zwiebel LJ, Giorgio TD. Olfactory receptor gene abundance in invasive breast carcinoma. Sci Rep. 2019;9(1):13736. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Massenet-Regad L, et al. Large-scale analysis of cell-cell communication reveals angiogenin-dependent tumor progression in clear cell renal cell carcinoma. iScience. 2023;26(12):108367. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Su C, et al. Single-cell RNA sequencing in multiple pathologic types of renal cell carcinoma revealed novel potential tumor-specific markers. Front Oncol. 2021;11:719564. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Zvirblyte J, et al. Single-cell transcriptional profiling of clear cell renal cell carcinoma reveals a tumor-associated endothelial tip cell phenotype. Commun Biol. 2024;7(1):780. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Yu Z, et al. Integrative single-cell analysis reveals transcriptional and epigenetic regulatory features of clear cell renal cell carcinoma. Cancer Res. 2023;83(5):700–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Rodrigues P, et al. NF-κB-dependent lymphoid enhancer co-option promotes renal carcinoma metastasis. Cancer Discov. 2018;8(7):850–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Lake, B.B., et al. An atlas of healthy and injured cell states and niches in the human kidney. bioRxiv, 2021: p. 2021.07.28.454201. [DOI] [PMC free article] [PubMed]
23.Hao Y, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573-3587.e29. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Aran D, et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat Immunol. 2019;20(2):163–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Fornes O, et al. JASPAR 2020: update of the open-access database of transcription factor binding profiles. Nucleic Acids Res. 2020;48(D1):D87–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Akdemir KC, Chin L. HiCPlotter integrates genomic data with interaction matrices. Genome Biol. 2015;16(1):198. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Kruse K, Hug CB, Vaquerizas JM. FAN-C: a feature-rich framework for the analysis and visualisation of chromosome conformation capture data. Genome Biol. 2020;21(1):303. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Șenbabaoğlu Y, Michailidis G, Li JZ. Critical limitations of consensus clustering in class discovery. Sci Rep. 2014;4:6207. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Chen Y, et al. edgeR v4: powerful differential analysis of sequencing data with expanded functionality and improved support for small counts and larger datasets. Nucleic Acids Res. 2025. 10.1093/nar/gkaf018. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Ritchie ME, et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Liberzon A, et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011;27(12):1739–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Kolberg L, et al. g:Profiler-interoperable web service for functional enrichment analysis and gene identifier mapping (2023 update). Nucleic Acids Res. 2023;51(W1):W207–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Yang, W., et al., Genomics of Drug Sensitivity in Cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. Nucleic Acids Res, 2013. 41(Database issue): p. D955–61. [DOI] [PMC free article] [PubMed]
35.Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature, 2013. 499(7456): p. 43–9. [DOI] [PMC free article] [PubMed]
36.Menyhart O, Kothalawala WJ, Győrffy B. A gene set enrichment analysis for cancer hallmarks. J Pharm Anal. 2025;15(5):101065. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Guardia GDA, et al. Acute myeloid leukemia expresses a specific group of olfactory receptors. Cancers (Basel). 2023. 10.3390/cancers15123073. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Sato Y, et al. Integrated molecular analysis of clear-cell renal cell carcinoma. Nat Genet. 2013;45(8):860–7. [DOI] [PubMed] [Google Scholar]
39.Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Rosser EC, Mauri C. The emerging field of regulatory B cell immunometabolism. Cell Metab. 2021;33(6):1088–97. [DOI] [PubMed] [Google Scholar]
41.Hu HT, et al. Characterization of intratumoral and circulating IL-10-producing B cells in gastric cancer. Exp Cell Res. 2019;384(2):111652. [DOI] [PubMed] [Google Scholar]
42.Collinson R, Tanos B. Primary cilia and cancer: a tale of many faces. Oncogene. 2025;44(21):1551–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26(12):1572–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Aran D. Cell-type enrichment analysis of bulk transcriptomes using xCell. Methods Mol Biol. 2020;2120:263–76. [DOI] [PubMed] [Google Scholar]
45.Racle J, et al. Simultaneous enumeration of cancer and immune cell types from bulk tumor gene expression data. Elife. 2017. 10.7554/eLife.26476. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Lecoultre M, Dutoit V, Walker PR. Phagocytic function of tumor-associated macrophages as a key determinant of tumor progression control: a review. J Immunother Cancer. 2020. 10.1136/jitc-2020-001408. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Luan X, et al. Blockade of C5a receptor unleashes tumor-associated macrophage antitumor response and enhances CXCL9-dependent CD8(+) T cell activity. Mol Ther. 2024;32(2):469–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Xu Z, et al. Scavenger receptor CD36 in tumor-associated macrophages promotes cancer progression by dampening type-I IFN signaling. Cancer Res. 2025;85(3):462–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Zhang WJ, et al. Tumor-associated macrophages correlate with phenomenon of epithelial-mesenchymal transition and contribute to poor prognosis in triple-negative breast cancer patients. J Surg Res. 2018;222:93–101. [DOI] [PubMed] [Google Scholar]
50.Lv Y, Wang H, Liu Z. The role of regulatory B cells in patients with acute myeloid leukemia. Med Sci Monit. 2019;25:3026–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Jiang X, et al. The role of microenvironment in tumor angiogenesis. J Exp Clin Cancer Res. 2020;39(1):204. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Vera C, et al. Role of nerve growth factor and its TRKA receptor in normal ovarian and epithelial ovarian cancer angiogenesis. J Ovarian Res. 2014;7:82. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Aoki S, et al. Placental growth factor promotes tumour desmoplasia and treatment resistance in intrahepatic cholangiocarcinoma. Gut. 2022;71(1):185–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Lin XC, et al. NFAT5 promotes arteriogenesis via MCP-1-dependent monocyte recruitment. J Cell Mol Med. 2020;24(2):2052–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Yu H, et al. Transcription factor NFAT5 promotes glioblastoma cell-driven angiogenesis via SBF2-AS1/miR-338-3p-mediated EGFL7 expression change. Front Mol Neurosci. 2017;10:301. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Mo CK, et al. Tumour evolution and microenvironment interactions in 2D and 3D space. Nature. 2024;634(8036):1178–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Long HS, et al. Making sense of the linear genome, gene function and TADs. Epigenetics Chromatin. 2022;15(1):4. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Lieberman-Aiden E, et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science. 2009;326(5950):289–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Uslu C, Kapan E, Lyakhovich A. Cancer resistance and metastasis are maintained through oxidative phosphorylation. Cancer Lett. 2024;587:216705. [DOI] [PubMed] [Google Scholar]
60.LeBleu VS, et al. PGC-1α mediates mitochondrial biogenesis and oxidative phosphorylation in cancer cells to promote metastasis. Nat Cell Biol. 2014;16(10):992–1003 (1-15). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Araujo LF, et al. Mitochondrial transcription factor A (TFAM) shapes metabolic and invasion gene signatures in melanoma. Sci Rep. 2018;8(1):14190. [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Zuo Y, et al. The HIF-1/SNHG1/miR-199a-3p/TFAM axis explains tumor angiogenesis and metastasis under hypoxic conditions in breast cancer. BioFactors. 2021;47(3):444–60. [DOI] [PubMed] [Google Scholar]
63.Cuppari A, et al. DNA specificities modulate the binding of human transcription factor A to mitochondrial DNA control region. Nucleic Acids Res. 2019;47(12):6519–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Li D, et al. The notch ligand JAGGED1 as a target for anti-tumor therapy. Front Oncol. 2014;4:254. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Normanno N, et al. Epidermal growth factor receptor (EGFR) signaling in cancer. Gene. 2006;366(1):2–16. [DOI] [PubMed] [Google Scholar]
66.Jäger B, et al. CXCR4/MIF axis amplifies tumor growth and epithelial-mesenchymal interaction in non-small cell lung cancer. Cell Signal. 2020;73:109672. [DOI] [PubMed] [Google Scholar]
67.Kadomatsu K, Kishida S, Tsubota S. The heparin-binding growth factor midkine: the biological activities and candidate receptors. J Biochem. 2013;153(6):511–21. [DOI] [PubMed] [Google Scholar]
68.Bozkaya G, et al. Cooperative interaction of MUC1 with the HGF/c-Met pathway during hepatocarcinogenesis. Mol Cancer. 2012;11:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Babina IS, Turner NC. Advances and challenges in targeting FGFR signalling in cancer. Nat Rev Cancer. 2017;17(5):318–32. [DOI] [PubMed] [Google Scholar]
70.Mankuzhy P, et al. The role of Wnt signaling in mesenchymal stromal cell-driven angiogenesis. Tissue Cell. 2023;85:102240. [DOI] [PubMed] [Google Scholar]
71.Liu D, et al. Dehydroepiandrosterone stimulates endothelial proliferation and angiogenesis through extracellular signal-regulated kinase 1/2-mediated mechanisms. Endocrinology. 2008;149(3):889–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Araujo AM, et al. Stromal oncostatin m cytokine promotes breast cancer progression by reprogramming the tumor microenvironment. J Clin Invest. 2022. 10.1172/JCI165107. [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Tsai T, et al. Two olfactory receptors-OR2A4/7 and OR51B5-differentially affect epidermal proliferation and differentiation. Exp Dermatol. 2017;26(1):58–65. [DOI] [PubMed] [Google Scholar]
74.Zhao YQ, et al. Prediction of tumor microenvironment characteristics and treatment response in lung squamous cell carcinoma by pseudogene OR7E47P-related immune genes. Curr Med Sci. 2023;43(6):1133–50. [DOI] [PubMed] [Google Scholar]
75.Zou J, et al. Identification of key genes associated with polycystic ovary syndrome (PCOS) and ovarian cancer using an integrated bioinformatics analysis. J Ovarian Res. 2022;15(1):30. [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Giwa A, et al. Identification of novel prognostic markers of survival time in high-risk neuroblastoma using gene expression profiles. Oncotarget. 2020;11(46):4293–305. [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Weber L, et al. Olfactory receptors as biomarkers in human breast carcinoma tissues. Front Oncol. 2018;8:33. [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Kalbe B, et al. Olfactory signaling components and olfactory receptors are expressed in tubule cells of the human kidney. Arch Biochem Biophys. 2016;610:8–15. [DOI] [PubMed] [Google Scholar]
79.Qiao X, et al. JAG1 is associated with the prognosis and metastasis in breast cancer. Sci Rep. 2022;12(1):21986. [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Patiño-García A, et al. Profiling of chemonaive osteosarcoma and paired-normal cells identifies EBF2 as a mediator of osteoprotegerin inhibition to tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis. Clin Cancer Res. 2009;15(16):5082–91. [DOI] [PubMed] [Google Scholar]
81.Mohanty S, Huang J, Basu A. Enhancement of cisplatin sensitivity of cisplatin-resistant human cervical carcinoma cells by bryostatin 1. Clin Cancer Res. 2005;11(18):6730–7. [DOI] [PubMed] [Google Scholar]
82.Cui T, et al. Olfactory receptor 51E1 protein as a potential novel tissue biomarker for small intestine neuroendocrine carcinomas. Eur J Endocrinol. 2013;168(2):253–61. [DOI] [PubMed] [Google Scholar]
83.Zhang P, et al. Dissecting the single-cell transcriptome network underlying gastric premalignant lesions and early gastric cancer. Cell Rep. 2019;27(6):1934-1947.e5. [DOI] [PubMed] [Google Scholar]
84.Pronin A, Slepak V. Ectopically expressed olfactory receptors OR51E1 and OR51E2 suppress proliferation and promote cell death in a prostate cancer cell line. J Biol Chem. 2021;296:100475. [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Marquardt A, et al. Subgroup-independent mapping of renal cell carcinoma-machine learning reveals prognostic mitochondrial gene signature beyond histopathologic boundaries. Front Oncol. 2021;11:621278. [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Wu Y, et al. Driver and novel genes correlated with metastasis of non-small cell lung cancer: a comprehensive analysis. Pathol Res Pract. 2021;224:153551. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

10238_2026_2118_MOESM1_ESM.pdf^{(1MB, pdf)}

10238_2026_2118_MOESM2_ESM.pdf^{(3MB, pdf)}

10238_2026_2118_MOESM3_ESM.pdf^{(174.1KB, pdf)}

10238_2026_2118_MOESM4_ESM.pdf^{(195.6KB, pdf)}

10238_2026_2118_MOESM5_ESM.pdf^{(415.2KB, pdf)}

10238_2026_2118_MOESM6_ESM.pdf^{(4.4MB, pdf)}

10238_2026_2118_MOESM7_ESM.pdf^{(412.8KB, pdf)}

Supplementary file8^{(11.5KB, xlsx)}

Supplementary file9^{(9.3KB, xlsx)}

Data Availability Statement

[CR1] 1.Clear cell renal cell carcinoma - National Cancer Institute. 2025 [cited 2025 June 20]; Available from: https://www.cancer.gov/pediatric-adult-rare-tumor/rare-tumors/rare-kidney-tumors/clear-cell-renal-cell-carcinoma.

[CR2] 2.Manley BJ, et al. Integration of recurrent somatic mutations with clinical outcomes: a pooled analysis of 1049 patients with clear cell renal cell carcinoma. Eur Urol Focus. 2017;3(4–5):421–7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Drabkin HA, Gemmill RM. Cholesterol and the development of clear-cell renal carcinoma. Curr Opin Pharmacol. 2012;12(6):742–50. [DOI] [PubMed] [Google Scholar]

[CR4] 4.Kidney cancer statistics. [cited 2025 June 20]; Available from: https://www.wcrf.org/preventing-cancer/cancer-statistics/kidney-cancer-statistics/.

[CR5] 5.Siegel RL, et al. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17–48. [DOI] [PubMed] [Google Scholar]

[CR6] 6.Flegel C, et al. Expression profile of ectopic olfactory receptors determined by deep sequencing. PLoS ONE. 2013;8(2):e55368. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Barnes IHA, et al. Expert curation of the human and mouse olfactory receptor gene repertoires identifies conserved coding regions split across two exons. BMC Genomics. 2020;21(1):196. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Armelin-Correa LM, et al. Nuclear compartmentalization of odorant receptor genes. Proc Natl Acad Sci U S A. 2014;111(7):2782–7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Maßberg D, Hatt H. Human olfactory receptors: Novel cellular functions outside of the nose. Physiol Rev. 2018;98(3):1739–63. [DOI] [PubMed] [Google Scholar]

[CR10] 10.Ansoleaga B, et al. Dysregulation of brain olfactory and taste receptors in AD, PSP and CJD, and AD-related model. Neuroscience. 2013;248:369–82. [DOI] [PubMed] [Google Scholar]

[CR11] 11.Garcia-Esparcia P, et al. Functional genomics reveals dysregulation of cortical olfactory receptors in Parkinson disease: Novel putative chemoreceptors in the human brain. J Neuropathol Exp Neurol. 2013;72(6):524–39. [DOI] [PubMed] [Google Scholar]

[CR12] 12.Chaudhary PK, Kim S. An insight into GPCR and G-Proteins as cancer drivers. Cells. 2021. 10.3390/cells10123288. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Balijepalli P, Sitton CC, Meier KE. Lysophosphatidic acid signaling in cancer cells: What makes LPA so special? Cells. 2021. 10.3390/cells10082059. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Chung C, et al. Odorant receptors in cancer. BMB Rep. 2022;55(2):72–80. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Morita R, et al. Olfactory Receptor Family 7 Subfamily C Member 1 is a novel marker of Colon cancer-initiating cells and is a potent target of immunotherapy. Clin Cancer Res. 2016;22(13):3298–309. [DOI] [PubMed] [Google Scholar]

[CR16] 16.Masjedi S, Zwiebel LJ, Giorgio TD. Olfactory receptor gene abundance in invasive breast carcinoma. Sci Rep. 2019;9(1):13736. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Massenet-Regad L, et al. Large-scale analysis of cell-cell communication reveals angiogenin-dependent tumor progression in clear cell renal cell carcinoma. iScience. 2023;26(12):108367. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Su C, et al. Single-cell RNA sequencing in multiple pathologic types of renal cell carcinoma revealed novel potential tumor-specific markers. Front Oncol. 2021;11:719564. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Zvirblyte J, et al. Single-cell transcriptional profiling of clear cell renal cell carcinoma reveals a tumor-associated endothelial tip cell phenotype. Commun Biol. 2024;7(1):780. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Yu Z, et al. Integrative single-cell analysis reveals transcriptional and epigenetic regulatory features of clear cell renal cell carcinoma. Cancer Res. 2023;83(5):700–19. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Rodrigues P, et al. NF-κB-dependent lymphoid enhancer co-option promotes renal carcinoma metastasis. Cancer Discov. 2018;8(7):850–65. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Lake, B.B., et al. An atlas of healthy and injured cell states and niches in the human kidney. bioRxiv, 2021: p. 2021.07.28.454201. [DOI] [PMC free article] [PubMed]

[CR23] 23.Hao Y, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573-3587.e29. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Aran D, et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat Immunol. 2019;20(2):163–72. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Fornes O, et al. JASPAR 2020: update of the open-access database of transcription factor binding profiles. Nucleic Acids Res. 2020;48(D1):D87–92. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Akdemir KC, Chin L. HiCPlotter integrates genomic data with interaction matrices. Genome Biol. 2015;16(1):198. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Kruse K, Hug CB, Vaquerizas JM. FAN-C: a feature-rich framework for the analysis and visualisation of chromosome conformation capture data. Genome Biol. 2020;21(1):303. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Șenbabaoğlu Y, Michailidis G, Li JZ. Critical limitations of consensus clustering in class discovery. Sci Rep. 2014;4:6207. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Chen Y, et al. edgeR v4: powerful differential analysis of sequencing data with expanded functionality and improved support for small counts and larger datasets. Nucleic Acids Res. 2025. 10.1093/nar/gkaf018. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Ritchie ME, et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Liberzon A, et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics. 2011;27(12):1739–40. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Kolberg L, et al. g:Profiler-interoperable web service for functional enrichment analysis and gene identifier mapping (2023 update). Nucleic Acids Res. 2023;51(W1):W207–12. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Yang, W., et al., Genomics of Drug Sensitivity in Cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. Nucleic Acids Res, 2013. 41(Database issue): p. D955–61. [DOI] [PMC free article] [PubMed]

[CR35] 35.Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature, 2013. 499(7456): p. 43–9. [DOI] [PMC free article] [PubMed]

[CR36] 36.Menyhart O, Kothalawala WJ, Győrffy B. A gene set enrichment analysis for cancer hallmarks. J Pharm Anal. 2025;15(5):101065. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Guardia GDA, et al. Acute myeloid leukemia expresses a specific group of olfactory receptors. Cancers (Basel). 2023. 10.3390/cancers15123073. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Sato Y, et al. Integrated molecular analysis of clear-cell renal cell carcinoma. Nat Genet. 2013;45(8):860–7. [DOI] [PubMed] [Google Scholar]

[CR39] 39.Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9:559. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Rosser EC, Mauri C. The emerging field of regulatory B cell immunometabolism. Cell Metab. 2021;33(6):1088–97. [DOI] [PubMed] [Google Scholar]

[CR41] 41.Hu HT, et al. Characterization of intratumoral and circulating IL-10-producing B cells in gastric cancer. Exp Cell Res. 2019;384(2):111652. [DOI] [PubMed] [Google Scholar]

[CR42] 42.Collinson R, Tanos B. Primary cilia and cancer: a tale of many faces. Oncogene. 2025;44(21):1551–66. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26(12):1572–3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Aran D. Cell-type enrichment analysis of bulk transcriptomes using xCell. Methods Mol Biol. 2020;2120:263–76. [DOI] [PubMed] [Google Scholar]

[CR45] 45.Racle J, et al. Simultaneous enumeration of cancer and immune cell types from bulk tumor gene expression data. Elife. 2017. 10.7554/eLife.26476. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Lecoultre M, Dutoit V, Walker PR. Phagocytic function of tumor-associated macrophages as a key determinant of tumor progression control: a review. J Immunother Cancer. 2020. 10.1136/jitc-2020-001408. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Luan X, et al. Blockade of C5a receptor unleashes tumor-associated macrophage antitumor response and enhances CXCL9-dependent CD8(+) T cell activity. Mol Ther. 2024;32(2):469–89. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Xu Z, et al. Scavenger receptor CD36 in tumor-associated macrophages promotes cancer progression by dampening type-I IFN signaling. Cancer Res. 2025;85(3):462–76. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Zhang WJ, et al. Tumor-associated macrophages correlate with phenomenon of epithelial-mesenchymal transition and contribute to poor prognosis in triple-negative breast cancer patients. J Surg Res. 2018;222:93–101. [DOI] [PubMed] [Google Scholar]

[CR50] 50.Lv Y, Wang H, Liu Z. The role of regulatory B cells in patients with acute myeloid leukemia. Med Sci Monit. 2019;25:3026–31. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Jiang X, et al. The role of microenvironment in tumor angiogenesis. J Exp Clin Cancer Res. 2020;39(1):204. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Vera C, et al. Role of nerve growth factor and its TRKA receptor in normal ovarian and epithelial ovarian cancer angiogenesis. J Ovarian Res. 2014;7:82. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Aoki S, et al. Placental growth factor promotes tumour desmoplasia and treatment resistance in intrahepatic cholangiocarcinoma. Gut. 2022;71(1):185–93. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Lin XC, et al. NFAT5 promotes arteriogenesis via MCP-1-dependent monocyte recruitment. J Cell Mol Med. 2020;24(2):2052–63. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Yu H, et al. Transcription factor NFAT5 promotes glioblastoma cell-driven angiogenesis via SBF2-AS1/miR-338-3p-mediated EGFL7 expression change. Front Mol Neurosci. 2017;10:301. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Mo CK, et al. Tumour evolution and microenvironment interactions in 2D and 3D space. Nature. 2024;634(8036):1178–86. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Long HS, et al. Making sense of the linear genome, gene function and TADs. Epigenetics Chromatin. 2022;15(1):4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Lieberman-Aiden E, et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science. 2009;326(5950):289–93. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Uslu C, Kapan E, Lyakhovich A. Cancer resistance and metastasis are maintained through oxidative phosphorylation. Cancer Lett. 2024;587:216705. [DOI] [PubMed] [Google Scholar]

[CR60] 60.LeBleu VS, et al. PGC-1α mediates mitochondrial biogenesis and oxidative phosphorylation in cancer cells to promote metastasis. Nat Cell Biol. 2014;16(10):992–1003 (1-15). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Araujo LF, et al. Mitochondrial transcription factor A (TFAM) shapes metabolic and invasion gene signatures in melanoma. Sci Rep. 2018;8(1):14190. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Zuo Y, et al. The HIF-1/SNHG1/miR-199a-3p/TFAM axis explains tumor angiogenesis and metastasis under hypoxic conditions in breast cancer. BioFactors. 2021;47(3):444–60. [DOI] [PubMed] [Google Scholar]

[CR63] 63.Cuppari A, et al. DNA specificities modulate the binding of human transcription factor A to mitochondrial DNA control region. Nucleic Acids Res. 2019;47(12):6519–37. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR64] 64.Li D, et al. The notch ligand JAGGED1 as a target for anti-tumor therapy. Front Oncol. 2014;4:254. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Normanno N, et al. Epidermal growth factor receptor (EGFR) signaling in cancer. Gene. 2006;366(1):2–16. [DOI] [PubMed] [Google Scholar]

[CR66] 66.Jäger B, et al. CXCR4/MIF axis amplifies tumor growth and epithelial-mesenchymal interaction in non-small cell lung cancer. Cell Signal. 2020;73:109672. [DOI] [PubMed] [Google Scholar]

[CR67] 67.Kadomatsu K, Kishida S, Tsubota S. The heparin-binding growth factor midkine: the biological activities and candidate receptors. J Biochem. 2013;153(6):511–21. [DOI] [PubMed] [Google Scholar]

[CR68] 68.Bozkaya G, et al. Cooperative interaction of MUC1 with the HGF/c-Met pathway during hepatocarcinogenesis. Mol Cancer. 2012;11:64. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Babina IS, Turner NC. Advances and challenges in targeting FGFR signalling in cancer. Nat Rev Cancer. 2017;17(5):318–32. [DOI] [PubMed] [Google Scholar]

[CR70] 70.Mankuzhy P, et al. The role of Wnt signaling in mesenchymal stromal cell-driven angiogenesis. Tissue Cell. 2023;85:102240. [DOI] [PubMed] [Google Scholar]

[CR71] 71.Liu D, et al. Dehydroepiandrosterone stimulates endothelial proliferation and angiogenesis through extracellular signal-regulated kinase 1/2-mediated mechanisms. Endocrinology. 2008;149(3):889–98. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR72] 72.Araujo AM, et al. Stromal oncostatin m cytokine promotes breast cancer progression by reprogramming the tumor microenvironment. J Clin Invest. 2022. 10.1172/JCI165107. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Tsai T, et al. Two olfactory receptors-OR2A4/7 and OR51B5-differentially affect epidermal proliferation and differentiation. Exp Dermatol. 2017;26(1):58–65. [DOI] [PubMed] [Google Scholar]

[CR74] 74.Zhao YQ, et al. Prediction of tumor microenvironment characteristics and treatment response in lung squamous cell carcinoma by pseudogene OR7E47P-related immune genes. Curr Med Sci. 2023;43(6):1133–50. [DOI] [PubMed] [Google Scholar]

[CR75] 75.Zou J, et al. Identification of key genes associated with polycystic ovary syndrome (PCOS) and ovarian cancer using an integrated bioinformatics analysis. J Ovarian Res. 2022;15(1):30. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Giwa A, et al. Identification of novel prognostic markers of survival time in high-risk neuroblastoma using gene expression profiles. Oncotarget. 2020;11(46):4293–305. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.Weber L, et al. Olfactory receptors as biomarkers in human breast carcinoma tissues. Front Oncol. 2018;8:33. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR78] 78.Kalbe B, et al. Olfactory signaling components and olfactory receptors are expressed in tubule cells of the human kidney. Arch Biochem Biophys. 2016;610:8–15. [DOI] [PubMed] [Google Scholar]

[CR79] 79.Qiao X, et al. JAG1 is associated with the prognosis and metastasis in breast cancer. Sci Rep. 2022;12(1):21986. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR80] 80.Patiño-García A, et al. Profiling of chemonaive osteosarcoma and paired-normal cells identifies EBF2 as a mediator of osteoprotegerin inhibition to tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis. Clin Cancer Res. 2009;15(16):5082–91. [DOI] [PubMed] [Google Scholar]

[CR81] 81.Mohanty S, Huang J, Basu A. Enhancement of cisplatin sensitivity of cisplatin-resistant human cervical carcinoma cells by bryostatin 1. Clin Cancer Res. 2005;11(18):6730–7. [DOI] [PubMed] [Google Scholar]

[CR82] 82.Cui T, et al. Olfactory receptor 51E1 protein as a potential novel tissue biomarker for small intestine neuroendocrine carcinomas. Eur J Endocrinol. 2013;168(2):253–61. [DOI] [PubMed] [Google Scholar]

[CR83] 83.Zhang P, et al. Dissecting the single-cell transcriptome network underlying gastric premalignant lesions and early gastric cancer. Cell Rep. 2019;27(6):1934-1947.e5. [DOI] [PubMed] [Google Scholar]

[CR84] 84.Pronin A, Slepak V. Ectopically expressed olfactory receptors OR51E1 and OR51E2 suppress proliferation and promote cell death in a prostate cancer cell line. J Biol Chem. 2021;296:100475. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR85] 85.Marquardt A, et al. Subgroup-independent mapping of renal cell carcinoma-machine learning reveals prognostic mitochondrial gene signature beyond histopathologic boundaries. Front Oncol. 2021;11:621278. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR86] 86.Wu Y, et al. Driver and novel genes correlated with metastasis of non-small cell lung cancer: a comprehensive analysis. Pathol Res Pract. 2021;224:153551. [DOI] [PubMed] [Google Scholar]

PERMALINK

Multi-omics profiling identifies ectopic olfactory receptors as putative drivers of tumor progression and prognostic indicators in clear cell renal cell carcinoma

Dong Jun Yeo

Hee Jin Cho

Abstract

Supplementary Information

Introduction

Methods

Data collection

Single-cell RNA-seq data processing

Single-cell ATAC-seq data processing

Motif and footprint analyses

Hi-C data analysis

Cell-to-cell communication analysis

Consensus clustering

Weighted gene co-expression network analysis (WGCNA)

Survival analysis

Construction of and validation of a prognostic model

Differentially expressed genes analysis

Gene set enrichment analysis (GSEA)

Gene ontology analysis

Quantification of pathway activity

Analysis of drug response

Statistical analysis

Results

Overall expression profiles of OR genes in kidney renal cell carcinoma

Fig. 1.

Construction of a prognostic model with the OR gene expression

Fig. 2.

Classification of tumor samples with OR gene expression based on consensus clustering

Fig. 3.

Overall expression profiles of OR genes at a single-cell level

Fig. 4.

OR51E1 is specifically expressed in pericytes and is associated with angiogenesis

Fig. 5.

OR2 family genes have cancer cell-specific expression and are associated with invasion and metastasis

Fig. 6.

Characterization of cell-to-cell communication networks associated with OR gene expression

Fig. 7.

Drug response profiles associated with OR gene expression

Fig. 8.

Discussion

Fig. 9.

Conclusions

Supplementary Information

Acknowledgements

Abbreviations

Authors contributions

Funding

Data availability

Declarations

Competing interests

Ethics approval

Consent to participate

Consent to publish

Footnotes

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases