Objective To explore differential expression of mitophagy-related genes (MRGs) in stomach adenocarcinoma(STAD) and to investigate related core genes mining.

Methods Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases to obtain clinical sample information of patients with STAD. A total of 26 MRGs with statistical significance were obtained based on the Gene Set Enrichment Analysis (GSEA) website and GeneCards database. A prognostic model of MRGs in STAD was constructed based on Least Absolute Shrinkage and Selection Operator (LASSO) regression, and mitochondrial autophagy prognostic genes (MPRGs) were screened out. Through gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, key genes involved in biological processes and pathways were obtained. Protein-protein interaction(PPI) network was established by means of STRING database, and the Cytoscape software was utilized for visualization.

Results Fifteen MRGs were selected in the LASSO regression model: ATG12, CSNK2A2, CSNK2B, FUNDC1, MAP1LC3A, MAP1LC3B, PGAM5, PINK1, SQSTM1, TOMM20, TOMM22, TOMM40, TOMM5, UBA52 and UBC, which were risk factors for STAD. Among 15 MPRGs, UBC and UBA52 genes had more impact on progression and prognosis of STAD, and the expression of PGAM5 was significantly correlated with the occurrence of STAD. ATG12 gene had the highest functional similarity score with other genes. PPI network analysis showed that PINK1 protein had the most interactions with other proteins.

Conclusion A total of 15 MPRGs play important roles in the occurrence and development of STAD, and may be used as targets for STAD gene detection, treatment and independent prognostic tools for STAD.