The HP-related prognostic genes were identified based on HP infection-related gene markers and HP infection sample datasets by risk method and NMF algorithm. Principal component analysis (PCA) algorithm was used to constructed the HPscore system. The “limma” R package was employed to determine differentially expressed genes. In addition, the R packages, such as “xCell” and “GSVA”, was used to analyze the relationship between the HPscore and tumor microenvironment. Finally, quantitative real-time polymerase chain reaction (qRT-PCR) was conducted to verify the expression levels of 28 HP-related prognostic genes in tissues.
We successfully identified 28 HP-related prognostic genes that accurately classified the GC population. There are significant differences in survival between different subgroups (high-, low-risk and cluster_1,2). Thereafter, the HPscore system was constructed to evaluate the signatures of the 28 HP-related prognostic genes. The overall survival rate in the high-HPscore group was poor and immunological surveillance was reduced, whereas the low-HPscore group had a survival advantage and was related to the inflammatory response. HPscore was also strongly correlated with the tumour stage, TME cell infiltration and stemness. The qRT-PCR results showed that DOCK4 expression level of 28 HP-related prognostic genes was higher in gastric cancer tissues than in adjacent tissues.
HP signatures play a crucial role in the TME and tumourigenesis. HPscore evaluation of a single tumour sample can help identify the TME characteristics and the carcinogenic mechanism of GC patients infected with HP, based on which personalized treatment can be administered.