Development and clinical validation of A Novel 9-gene prognostic biomarkers based on multi-omics in pancreatic adenocarcinoma

Abstract

Background: The prognosis of patients with pancreatic cancer remains poor due to the lack of biomarkers for early diagnosis and effective prognosis. Methods: RNA-Seq, SNP, CNV data were downloaded from The Cancer Genome Atlas(TCGA); Univariant cox regression was used to identify prognosis-related genes; GISTIC 2.0 was used to identify significantly amplified or deleted genes; Lasso method was used to construct risk prognosis model, which was then validated in GEO and ICGC cohorts. rms package was used to evaluate the overall predictive performance of the model by calculating and comparing the C-index values with other models. Experiments of western blot were performed to evaluate the expression of genes. Results: 54 candidate genes were obtained after integrating the genomic mutated genes and prognosis-related genes. The Lasso method was conducted to finally ascertain nine characteristic genes, including UNC13B, TSPYL4, MICAL1, KLHDC7B, KLHL32, AIM1, ARHGAP18, DCBLD1, and CACNA2D4. The 9-gene signature model can help with stratifying samples at risk in the training cohort and external validation cohort. In addition, the overall predictive performance of our model is superior to the others. We found that AIM1 and DCBLD1 were highly expressed in tumor tissues, while ARHGAP18, CACNA2D4, and TSPYL4 were lowly expressed in tumor tissues at the protein and transcription levels. Conclusion: The nine-gene signatures constructed in this study can be used as the novel prognostic markers to predict the survival of PAAD patients.