A Four Autophagy-Related Gene-Based Prognostic Signature for Pancreatic Cancer.

Abstract

This study aimed to screen autophagy-related genes (ARGs) that affect the prognosis of pancreatic cancer patients based on The Cancer Genome Atlas (TCGA) gene expression data and genotype-tissue expression (GTEx) databases. The expression data of pancreatic cancer and normal pancreas were downloaded from TCGA and GTEx databases. Human ARGs list was obtained through the Human Autophagy Database (HADB) and GeneCards database. The Wilcox test was performed to screen differentially expressed ARGs. Differentially expressed ARGs were analyzed by gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses. The CIBER-SORT algorithm was utilized to analyze immune cell infiltration in samples. A total of 21 up-regulated ARGs and 11 down-regulated ARGs were screened in the TCGA-GTEx integrated data set. The enrichment analysis of GO and KEGG showed that 32 differentially expressed ARGs were significantly enriched in autophagy-related pathways. Univariate Cox regression analysis showed that 12 candidate ARGs were significantly related to the prognosis of pancreatic cancer patients. Multivariate Cox regression analysis found that ATG16L2, GNAI3, APOL1, and PTK6 genes may be the key ARGs affecting the prognosis of pancreatic cancer patients. Based on these four key ARGs, a prognostic risk assessment model was constructed, and pancreatic cancer patients were classified into the high-risk and low-risk group according to the risk value. Survival analysis and ROC analysis confirmed that the prognostic risk assessment model can accurately predict the prognosis of patients with pancreatic cancer. Immune infiltration analysis found that B cells naive, B cells memory, plasma cells, T cells CD8, T cells CD4 memory resting, monocytes and macrophages M0 were significantly different in tissue samples of pancreatic cancer patients in the high and low risk groups. Pearson's correlation coefficient showed that the four key ARGs may affect the development of pancreatic cancer by affecting immune cell components in the tumor micro-environment. In conclusion, ATG16L2, GNAI3, APOL1, and PTK6 may be related to the prognosis of pancreatic cancer patients. The prognostic risk assessment model constructed based on these four key ARGs could accurately predict the prognosis of pancreatic cancer patients.