Abstract:To use a bioinformatics system to analyze the early diagnosis and prognosis prediction performance and immune infiltration correlation of Progestagenassociated endometrial protein (PAEP) in gastric cancer, this study constructed a predictive model ROC curve and deviation correction curve to test its performance accuracy, and further validate the analysis results through cell experiments. The results of bioinformatics analysis showed that the expression of PAEP in gastric cancer tissues was higher than that in normal tissues, which was statistically significant in the early stages (T1, N0, M0) (P < 0.001), and the diagnostic accuracy of gastric cancer was higher (the area under ROC curve was 0.889). The prognosis of gastric cancer patients with high expression of PAEP was poor (P ≤ 0.001). The later the TNM stage, the older the age, the higher the expression of PAEP, and the lower the survival probability of gastric cancer patients, and the deviation correction curve was close to the ideal curve (45 ° line), indicating a good predictive result. The expression of PAEP was positively correlated with the infiltration level of neutrophils and macrophages in gastric cancer, and negatively correlated with B cells, central memory T cells and T cells. qRT-PCR and Western blot results showed that the transcriptional and protein expression levels of PAEP in HGC-27, BGC-823, MKN-45, SGC-7901 and AGS cells were significantly higher than those in GES-1 cells. The results indicate that PAEP protein can be used as a biomarker in the diagnosis and prognosis of gastric cancer, which is verified by experiments, and its mechanism is related to immunity.