Identifying drug candidates for pancreatic ductal adenocarcinoma based on integrative multiomics analysis.

Background: Due to a lack of early diagnosis methods and effective drugs, pancreatic ductal adenocarcinoma (PDAC) has an extremely poor prognosis. DNA methylation, transcriptome expression and gene copy number variation (CNV) have critical relationships with development and progression of various diseases. The purpose of the study was to screen reliable early diagnostic biomarkers and potential drugs based on integrative multiomics analysis.

Methods: We used methylation, transcriptome and CNV profiles to build a diagnostic model for PDAC. The protein expression of three model-related genes were externally validated using PDAC samples. Then, potential therapeutic drugs for PDAC were identified by interaction information related to existing drugs and genes.

Results: Four significant differentially methylated regions (DMRs) were selected from 589 common DMRs to build a high-performance diagnostic model for PDAC. Then, four hub genes, , , and , were obtained. The external validation results showed that , and protein expression levels were all upregulated in tumor tissues compared with adjacent normal tissues (P<0.05). Promising candidate drugs with activity against PDAC were screened and repurposed through gene expression analysis of online datasets. The five drugs, including topotecan, PD-0325901, panobinostat, paclitaxel and 17-AAG, with the highest activity among 27 PDAC cell lines were filtered.

Conclusions: Overall, the diagnostic model built based on four significant DMRs could accurately distinguish tumor and normal tissues. The five drug candidates might be repurposed as promising therapeutics for particular PDAC patients.