Comprehensive data analysis of genomics, epigenomics, and transcriptomics to identify specific biomolecular markers for prostate adenocarcinoma

Background: Multiomics data analysis based on high-throughput sequencing technology has become a hotspot in tumor investigation. The present study aimed to explore prognostic biomarkers via investigating DNA copy number variation (CNV) and methylation variation (MET) data in prostate cancer. Methods: We obtained the messenger RNA (mRNA) expression, CNV, and methylated data of prostate adenocarcinoma (PRAD) samples via The Cancer Genome Atlas (TCGA)-PRAD cohort. We calculated and assessed the associations between CNV and RNA sequencing (RNA-seq), and between MET and RNAseq via Pearson correlation coefficients. We then used the "iCluster" package to perform multigroup cluster analysis with CNVcor gene CNV data, METcor gene methylation data, and CNVcor and METcor gene mRNA data. The univariate Cox analysis was used to screen significant hub genes, and multivariate Cox analysis was used to construct risk a model. The nomogram was constructed based on "rms" package, and the immune infiltrating patterns were compared between high-and low-risk groups. Results: A total of 477 PRAD samples with complete CNV, methylation, mRNA, and matched clinical information were included in our study. A list of 10,073 CNVcor genes and 9841 METcor genes were confirmed with a significance level of P<0.01. We found that CNVcor is more likely to appear on chromosome (chr)8, chr17, and chr10, while METcor is more likely to appear on chr1, chr19, and chr17. Based on the core genes, we finally classified the samples into 4 subtypes, incorporating iC1 (iCluster) (92 samples), iC2 (79 samples), iC3 (165 samples), and iC4 (141 samples). Furthermore, we constructed the prognostic model for PRAD based on the 5 genes (IER3, AOX1, PRKCDBP, UBD, and FBLN5). Nomograms incorporating risk score and other clinical variables were further constructed, and these nomograms exhibited superior predictive ability. We further compared the differential immune infiltrating patterns in 2 risk groups and found significantly low levels of infiltrating cluster of differentiation (CD)8(+) T cells in high-risk samples. Conclusions: Our study integrated the multi-omics data to elucidate the molecular features of PRAD and pivotal genes for predicting prognosis.