Background: Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy and its mortality continues to rise globally. Because of its high heterogeneity and complex molecular landscapes, published gene signatures have shown low specificity and robustness. Functional signatures containing a group of genes involved in similar biological functions display a more robust performance.
Methods: The present study was designed to excavate potential functional signatures for PDAC by analyzing maximal number of datasets extracted from available databases with a recently developed method of FAIME (Functional Analysis of Individual Microarray Expression) in a comprehensive and integrated way.
Results: By using appropriate search strategies, we extracted 11 PDAC datasets from GEO, ICGC and TCGA databases. By systemically analyzing these datasets, we identified a robust functional signature of subpathway (path:00982_1), which belongs to the drug metabolism-cytochrome P450 pathway. The functional signature has displayed a more powerful and robust capacity in predicting prognosis, drug response and chemotherapeutic efficacy for PDAC, particularly for the classical subtype, in comparison with published gene signatures and clinically used TNM staging system. This signature was further verified by meta-analyses and validated in cell line databases and available clinical datasets.
Conclusion: This is the first functional signature for PDAC identified from the largest number of datasets by using comprehensive and integrated analyses. The novel signature warrants a further investigation, since it is like to improve the current systems for predicting the prognosis and monitoring drug response, and to serve a potential linkage to therapeutic options for combating PDAC.