Identify mechanism of the prognosis of pancreatic ductal adenocarcinoma using gene expression profiles and co-expression network analysis

Abstract

Objective To detect genes expression related to the prognosis of pancreatic ductal adenocarcinoma (PDAC) and explore the potential molecular mechanism applying PDAC gene expression data,co-expression network and clinical data.Methods Univariate Cox proportional hazards model was used to find genes which were significantly correlated with patient survival data (P ＜ 0.000 1).Then calculated the correlation coefficient of these genes to establish co-expression network and found key modules of the network.Results 273 candidate genes were found related to survival data.With these genes,genes co-expression network (Pearson correlation r ＞ 0.9,P ＜ 0.05) and key modules were constructed,and then,a key module including 6 genes (PTMAP7,FTSJ3,TLK2,GOLGA3,PTTG1IP,ZBTB37) and 14 co-expression relationships between them were found.Among these genes,GOLGA3 had been known to associate with invasion and metastasis of PDAC,and the other 5 genes were all known as cancer genes.Conclusion These findings may suggest that these 6 survival genes in the key module of co-expression network are highly associated with the prognosis of PDAC,and may be biomarkers for PDAC prognosis. Key words: Neoplasms, pancreatic ductal;  Co-expression network;  Prognosis;  Molecular mechanism