Analysis of DNA Repair-Related Prognostic Function and Mechanism in Gastric Cancer

DNA repair mechanisms have been proven to be essential for cells, and abnormalities in DNA repair could cause various diseases, such as cancer. However, the diversity and complexity of DNA repair mechanisms obscure the functions of DNA repair in cancers. In addition, the relationships between DNA repair, the tumor mutational burden (TMB), and immune infiltration are still ambiguous. In the present study, we evaluated the prognostic values of various types of DNA repair mechanisms and found that double-strand break repair through single-strand annealing (SSA) and nonhomologous end-joining (NHEJ) was the most prognostic DNA repair processes in gastric cancer (GC) patients. Based on the activity of these two approaches and expression profiles, we constructed a HR-LR model, which could accurately divide patients into high-risk and low-risk groups with different probabilities of survival and recurrence. Similarly, we also constructed a cancer-normal model to estimate whether an individual had GC or normal health status. The prognostic value of the HR-LR model and the accuracy of the cancer-normal model were validated in several independent datasets. Notably, low-risk samples, which had higher SSA and NHEJ activities, had more somatic mutations and less immune infiltration. Furthermore, the analysis found that low-risk samples had higher and lower methylation levels in CpG islands (CGIs) and open sea regions respectively, and had higher expression levels of programmed death-ligand 1 (PD-L1) and lower methylation levels in the promoter of the gene encoding PD-L1. Moreover, low-risk samples were characterized primarily by higher levels of CD4(+) memory T cells, CD8(+) naive T cells, and CD8(+) TEM cells than those in high-risk samples. Finally, we proposed a decision tree and nomogram to help predict the clinical outcome of an individual. These results provide an improved understanding of the complexity of DNA repair, the TMB, and immune infiltration in GC, and present an accurate prognostic model for use in GC patients.