Abstract: Objective To explore the expression of glycolysis-related genes ( GRGs) in gastric cancer by bioinformatics, and the relationship between the established risk scoring model andprognosis of gastric cancer. Methods The genes expression profiles and clinical feature data of gastric cancer samples were downloaded from the Cancer Genome Atlas (TCGA) database. The GRGs set was obtained from the GSEA database. We used “ limma” packets to identify differentially expressed GRGs in gastric cancer tissues, and used univariate Cox regression analysis to screen for prognosis related GRGs. Then, LASSO regression analysis was used to construct a prognosis prediction model based on the GRGs. A nomogram was developed based on the independent prognostic risk factors determined by Cox regression analysis. Finally, the correlations between 22 types of infiltrating immune cells and the GRGs or risk scoring models were analyzed. Results Twenty-one differentially expressed GRGs were identified, which were mainly enriched in the alcohol metabolism andtyrosine metabolism pathways. Finally, 5 prognostic related glycolytic genes ( ADH1B, ADH4, CLDN9, VCAN, and LHX90) were selected by LASSO and Cox models and were used to constructa gene signature for gastric cancer prognosis prediction. The overall survival of gastric cancer patients with low-risk scores is significantly better than that of gastric cancer patients with high-risk scores. The ROC curve analysis showed that the values of area under the curve (AUC) of the abovemodel to predict the 1-year, 3-year, and 5-year survival for the gastric cancer patients were 0. 602,0. 680, and 0. 802, respectively. The effectiveness of this model to predict the survival of gastriccancer patients was better than that of using tumor staging and grading. Univariate and multivariate Coxanalysis showed that the prognostic model, age, gender, tumor stage and tumor grade were independent prognostic factors for gastric cancer. Based on these prognostic factors, a Nomogram was constructed to predict the survival of gastric cancer patients. Finally, we found that the proportion of CD8T cells and helper follicular T cells was significantly reduced in the high-risk group, while the proportion of M0 macrophages and M2 macrophages was significantly higher in the high-risk group than inthe low-risk group. The proportion of helper follicular T cells was negatively correlated with the expression levels of ADH1B and VCAN and the risk scores. The proportion of M2 macrophages was positivelycorrelated with the expression level of VCAN and the risk scores. Conclusion The 5 GRGs screenedout in this study are related to the prognosis of gastric cancer, which can be used for the prognosis ofgastric cancer patients and may be used as potential therapeutic targets for gastric cancer.

Key words: