Identification of pancreatic cancer type related factors by Weighted Gene Co-Expression Network Analysis

Abstract

This study aims to identify the core modules associated with pancreatic cancer (PC) types and the ncRNAs and transcription factors (TFs) that regulate core module genes by weighted gene co-expression network analysis (WGCNA). WGCNA was used to analyze the union of genes related to PC in NCBI and OMIM databases and the differentially expressed genes screened by TCGA-PAAD database. Samples were clustered according to gene expression in gene modules and Fisher exact method was performed. GO and KEGG were used for enrichment analysis to visually display module genes and screen driver genes. Hypergeometric test method was used to calculate pivot nodes among ncRNAs, TFs and mRNA based on RAID 2.0 and TRRUST v2 databases. The blue and yellow modules were identified as the core modules associated with PC types. MST1R, TMPRSS, MIR198, SULF1, COL1A1 and FAP were the core genes in the modules. Hypergeometric test results showed that ANCR, miR-3134, MT1DP, LOC154449, LOC28329 and other ncRNAs were key factors driving blue module genes, while LINC-ROR, UCA1, SNORD114-4, HEIH, SNORD114-6 and other ncRNAs were key factors driving yellow module genes. TFs with significant regulatory effect on blue module included LCOR, PIAS4, ZEB1, SNAI2, SMARCA4, etc. and on yellow module included HOXC6, PER2, HOXD3, TWIST2, VHL, etc. The core modules associated with PC types were proved as yellow and blue modules, and important ncRNAs and TFs regulating yellow and blue modules were found. This study provides relevant evidence for further identification of PC types.