Fig. 7: Single-cell transcriptomic profiling reveals clinical relevance of DSS1-driven malignant cells in ccRCC.

a Quality control for single-cell RNA-seq data from ccRCC tissues (n = 49 distinct samples). b, c Uniform Manifold Approximation and Projection (UMAP) of cells of ccRCC tissues before (b) and after batch effect correction (c). Right panel: annotation of major cell types. d, e Identification of malignant cells using inference of copy number variation (inferCNV, CNV score threshold: 0.04) of epithelial cells. Del, deletion; Amp amplification. f Comparison of DSS1 pseudo-bulk expression between malignant cells (tumor samples) and normal epithelial cells (normal samples, error bar: mean ± SD, n = 47 distinct samples with detected DSS1 RNA; two-tailed Welch’s t-test). g DSS1 amplification correlates with higher DSS1 expression (DESeq2 log₂CPM) in tumors (n = 513). h Functional module analysis for malignant cells. Upper graph: hierarchical clustering dendrogram of genes; lower graph: gene module (x-axis, different genes). i UMAP showing DSS1-driven ccRCC cells and other malignant cells. j Comparison of the epithelial-mesenchymal transition (EMT)/autophagy module scores between DSS1-driven cells and other malignant cells. KEGG v113.0, Kyoto Encyclopedia of Genes and Genomes. k Proportion of DSS1-driven cells across different stages in treatment-naive patients (I-IV: n = 1, 6, 6, 3 patients, respectively). l DSS1-driven cell signature scores in patients with different grades and pathologic stages (n = 521). m Comparison of DSS1-driven cell signature scores across molecular subtypes defined by Clinical Proteomic Tumor Analysis Consortium (CPTAC) in TCGA-KIRC dataset (n = 495). n High DSS1-driven cell signature predicts poorer survival (Log-rank test, n = 506 and 509, respectively). HR, Hazard Ratio. o–q Differences in signature scores of DSS1-driven cells between treatment-naive/response and treatment/resistance groups (VEGFR inhibitors) in ccRCC patients or patient-derived xenograft (PDX) models (left to right, TCGA-KIRC, n = 430; GSE65615, n = 122; GSE64052, n = 28; GSE76068, n = 16). MUT, mutant; WT, wildtype. g, j, l, m, o–q Two-tailed Mann-Whitney U test. Box plots (Turkey style): Center line = median; box = 25th to 75th percentiles; whiskers = ±1.5×interquartile range; Violin plots: data distribution (minima to maxima) via kernel density estimation. Statistics are provided in the source data. Source data are provided as a Source Data file.