Combined analysis and validation for DNA methylation and gene expression profiles associated with prostate cancer

Yanqiu Tong,
Yang Song and
Shixiong DengEmail author View ORCID ID profile

Cancer Cell International201919:50

https://doi.org/10.1186/s12935-019-0753-x

Received: 16 August 2018
Accepted: 8 February 2019
Published: 4 March 2019

Abstract

Background

Prostate cancer (PCa) is a malignancy cause of cancer deaths and frequently diagnosed in male. This study aimed to identify tumor suppressor genes, hub genes and their pathways by combined bioinformatics analysis.

Methods

A combined analysis method was used for two types of microarray datasets (DNA methylation and gene expression profiles) from the Gene Expression Omnibus (GEO). Differentially methylated genes (DMGs) were identified by the R package minfi and differentially expressed genes (DEGs) were screened out via the R package limma. A total of 4451 DMGs and 1509 DEGs, identified with nine overlaps between DMGs, DEGs and tumor suppressor genes, were screened for candidate tumor suppressor genes. All these nine candidate tumor suppressor genes were validated by TCGA (The Cancer Genome Atlas) database and Oncomine database. And then, the gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes pathway (KEGG) enrichment analyses were performed by DAVID (Database for Annotation, Visualization and Integrated Discovery) database. Protein–protein interaction (PPI) network was constructed by STRING and visualized in Cytoscape. At last, Kaplan–Meier analysis was performed to validate these genes.

Results

The candidate tumor suppressor genes were IKZF1, PPM1A, FBP1, SMCHD1, ALPL, CASP5, PYHIN1, DAPK1 and CASP8. By validation in TCGA database, PPM1A, DAPK1, FBP1, PYHIN1, ALPL and SMCHD1 were significant. The hub genes were FGFR1, FGF13 and CCND1. These hub genes were identified from the PPI network, and sub-networks revealed by these genes were involved in significant pathways.

Conclusion

In summary, the study indicated that the combined analysis for identifying target genes with PCa by bioinformatics tools promote our understanding of the molecular mechanisms and underlying the development of PCa. And the hub genes might serve as molecular targets and diagnostic biomarkers for precise diagnosis and treatment of PCa.