DNA methylome and transcriptome analysis established a model of four differentially methylated positions (DMPs) as a diagnostic marker in esophageal adenocarcinoma early detection

Background: Esophageal carcinogenesis involves in alterations of DNA methylation and gene transcription. This study profiled genomic DNA methylome vs. gene expression using transcriptome data on esophageal adenocarcinoma (EAC) tissues from the online databases in order to identify methylation biomarkers in EAC early diagnosis. Materials and Methods: The DNA methylome and transcriptome data were downloaded from the UCSC Xena, Gene Expression Omnibus (GEO), and The Cancer Genome Atlas (TCGA) databases and then bioinformatically analyzed for the differentially methylated positions (DMPs) vs. gene expression between EAC and normal tissues. The highly methylated DMPs vs. reduced gene expression in EAC were selected and then stratified with those of the corresponding normal blood samples and other common human cancers to construct an EAC-specific diagnostic model. The usefulness of this model was further verified in other three GEO datasets of EAC tissues. Result: A total of 841 DMPs were associated with expression of 320 genes, some of which were aberrantly methylated in EAC tissues. Further analysis showed that four (cg07589773, cg10474350, cg13011388 and cg15208375 mapped to gene IKZF1, HOXA7, EFS and TSHZ3, respectively) of these 841 DMPs could form and establish a diagnostic model after stratified them with the corresponding normal blood samples and other common human cancers. The data were further validated in other three GEO datasets on EAC tissues in early EAC diagnosis. Conclusion: This study revealed a diagnostic model of four genes methylation to diagnose EAC early. Further study will confirm the usefulness of this model in a prospective EAC cases.

Automated summary

This study identified a diagnostic model involving methylation of four genes for early diagnosis of EAC by analyzing DNA methylome and gene expression data from EAC tissues. The model was validated in other datasets and may have potential applications in prospective EAC cases.