Abstract
OBJECTIVE: To identify key mitochondrial autophagy-related genes (MARGs) in fetal growth restriction (FGR)and evaluate their diagnostic potential through bioinformatics and machine learning approaches. METHODS: The GSE24192 dataset were obtained from Gene Expression Omnibus data base (GEO). Differentially expressed genes (DEGs) were identified using differentially expressed analysis. Mitochondrial autophagy-related genes (MARGs) were identified using GeneCards. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed with the clusterProfiler package. Protein-protein interaction (PPI) network was constructed using STRING, and key genes were selected using machine learning. Receiver operating characteristic (ROC) curves assessed diagnostic performance of key genes. Immune infiltration analysis was used to evaluated immune microenvironment. The miRNAs were predicted in TargetScan website. RESULTS: A total of 42 MARGs were identified in FGR samples, and three key genes (THBS1, RAB15, LMO7) were selected through machine learning methods. These genes showed high diagnostic potential with area under the curve (AUC) values of 0.97, 0.95, and 0.92, respectively. Immune infiltration analysis revealed significant increase of CD8+ T cells, endothelial cells, and macrophages in FGR samples. Correlation analysis indicated THBS1 was positively related to several immune cells, while RAB15 and LMO7 were negatively related to several immune cells. The miRNA-mRNA regulatory network revealed four miRNAs potentially regulating these key genes. CONCLUSION: In conclusion, our study identified THBS1, RAB15, and LMO7 as key mitochondrial autophagy-related genes in FGR, with potential as diagnostic biomarkers.