Abstract
Based on TCGA, GTEx, and TIMER databases and various bioinformatics analysis methods, the potential biological roles of cuprotosis-related genes in pancreatic cancer were deeply explored, and a predictive model for pancreatic cancer patients was constructed. We downloaded the RNA-Seq data and clinicopathological and predictive data of 179 pancreatic cancer tissues and 332 adjacent normal tissues from TCGA and GTEx databases. The differential expression of cuprotosis-related genes in pancreatic cancer tissue and adjacent normal tissue was analyzed, and the LASSO regression algorithm was used to construct a prediction model and verify the validity of the model prediction. Based on the LASSO regression algorithm, a predictive model composed of three genes LIPT1, LIAS, and DLAT was screened. The corresponding survival curves showed that the constructed prediction model could significantly distinguish the prognosis of pancreatic cancer patients, and the prognosis of patients in the high-risk group was worse (P = 0.00557). The ROC curve showed that the area under the curve of the predictive model for predicting the 4-, 5-, and 6-year survival rates in pancreatic cancer was 0.816, 0.836, and 0.956, respectively. The AUC value of this risk model was significantly higher than 0.7, which could more accurately predict the prognosis of pancreatic cancer patients. This study determined a risk-scoring model of cuprotosis-related genes, which can provide an essential basis for judging the prognosis of pancreatic cancer patients.