Extended Data Fig. 3: Tumor-NAT comparisons reveal tumorigenic changes and biomarker candidates.
a. Principal component analysis for normalized transcriptome, proteome data, dots are colored based on sample types. b. Principal component analysis for normalized transcriptome data, dots are colored according to the tumor purity. c. Principal component analysis for normalized proteome data, dots are colored according the tumor purity. (In panels (a-c), n = 125 paired tumor and NAT samples.) d. Top panel: the proteome expression level of the indicated proteins in paired tumors and NAT samples without missing values (two-sided Wilcoxon signed-rank test. KLK, n = 125; AR, n = 105; AMACR, n = 125; NKX3-1, n = 110). For the boxplot, center line indicates the median value, lower and upper hinges represent the 25th and 75th percentiles, respectively and whiskers denote 1.5 × interquartile range. Bottom panel: their associations to PCa recurrence-free survival (Kaplan-Meier analysis, log-rank test, n numbers shown represent tumor samples). e. Venn plot for the upregulated (top panel) or downregulated (bottom panel) genes in transcriptome (DESeq2) and proteome (Wilcoxon signed-rank test). n = 125 paired tumor and NAT samples and the missing values were ignored. f. Expression level of three proteins in serum samples of BPH, primary tumor and metastatic PCa. (two-sided Wilcoxon rank-sum test). g. The two-sided Pearson correlation between serum concentration (ng/ml) and proteome level (log2 intensity of GOLM1 protein) in tumors, the blue shadow represents a confidence interval of 0.95. For panel (f-g), n numbers shown represent tumor samples.
