Identification of Hub Genes in Gastric Cancer with High Heterogeneity Based on Weighted Gene Co-Expression Network

Intra-tumor heterogeneity (ITH) plays an important role in the therapeutic resistance and prognosis of gastric cancer (GC), but there are no effective methods to detect it. The purpose of this study was to apply the mutant-allele tumor heterogeneity (MATH) algorithm to reveal the relationship between ITH and clinical features, and to use weighted gene co-expression network analysis (WGCNA) to search hub genes. The whole exome sequencing data with tumor mutations, RNA-seq, and clinical data were obtained from The Cancer Genome Atlas database. We calculated the MATH values and further investigated their relationships with clinical features and key genes screened out from molecular classification published by Nature. The WGCNA method was applied to discover hub genes within the high ITH cases. We found that MATH values were related to grade classification (P < 0.05). Our study also showed a "high MATH" group with a higher TP53 percentage (P < 0.001), whereas PIK3CA and RHOA had the opposite results (P = 0.004; P = 0.031). Using WGCNA, we found that red module, black module, and brown module were enriched in spliceosome, ribosome, and butanoate metabolism, and their hub genes were PSMD1, RPS23, and FAM84B, respectively. Together, these results demonstrate that in the high MATH group, represented as high heterogeneity, there was a higher frequency of TP53 mutation, and RHOA and PIK3CA tended to appear in the low heterogeneity group. PSMD1, RPS23, and FAM84B were the hub genes in high heterogeneity GC. They were related to the pathology and prognosis of patients.