Abstract
The pathogenic mechanisms of prostate cancer (PCa) remain to be defined. In this study, we utilized the Robust Rank Aggregation (RRA) method to integrate 10 eligible PCa microarray datasets from the GEO and identified a set of significant differentially expressed genes (DEGs) between tumor samples and normal, matched specimens. To explore potential associations between gene sets and PCa clinical features and to identify hub genes, we utilized WGCNA to construct gene co-expression networks incorporating the DEGs screened with the use of RRA. From the key module, we selected LMNB1, TK1, ZWINT, and RACGAP1 for validation. We found that these genes were up-regulated in PCa samples, and higher expression levels were associated with higher Gleason scores and tumor grades. Moreover, ROC and K-M plots indicated these genes had good diagnostic and prognostic value for PCa. On the other hand, methylation analyses suggested that the abnormal up-regulation of these four genes likely resulted from hypomethylation, while GSEA and GSVA for single hub gene revealed they all had a close association with proliferation of PCa cells. These findings provide new insight into PCa pathogenesis, and identify LMNB1, TK1, RACGAP1 and ZWINT as candidate biomarkers for diagnosis and prognosis of PCa.
Highlights
Prostate cancer (PCa) is the second most frequent cancer and the fifth leading cause of death in males worldwide [1]
Identification of robust differentially expressed gene (DEG) by the Rank Aggregation (RRA) method Figure 1 shows the workflow for identification, validation, and functional analysis of DEGs
Numerous investtigations using microarray and RNA-seq were conducted to discover novel biomarkers and therapeutic targets for PCa, inconsistencies were seen between the DEGs found in different studies [7]
Summary
Prostate cancer (PCa) is the second most frequent cancer and the fifth leading cause of death in males worldwide [1]. Prostate specific antigen (PSA) is the only circulating biomarker routinely used for early diagnosis of PCa [2]. Since the optimal PSA expression threshold for clinical samples has not been determined [4], routine PSA screening sometimes leads to overdiagnosis and over-treatment of indolent PCa [5, 6]. Small sample sizes in individual studies and use of different technological platforms create substantial inter-study variability and difficult statistical analyses. To solve this problem, integrated bioinformatics methods such as Robust Rank Aggregation (RRA) have been utilized in various cancer studies [8,9,10]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call