Abstract: Objective To screen hub genes for gestational diabetes mellitus（GDM） by bioinformatics. Methods Microarray data related to GDM were downloaded from gene expression omnibus（GEO）database（GSE103552）. Limma package was used to screen the differentially expressed genes of control group and GDM group. Clusterprofiler was used for Gene Ontology（GO） and Kyoto Encyclopedia of Genes and Genomes（KEGG） enrichment analysis. WGCNA outputed module genes were used，and the intersection with modular genes and differentially expressed genes was used to determine the characteristic difference genes. The online website STRING and Cytoscape software were used to construct the protein interaction network for the characteristic difference genes，and the hub genes were screened. R software was used to analyze the expression and diagnostic value of hub genes. Receiver operating characteristic（ROC） curve was used to evaluate the diagnostic performance of hub genes. Results The GDM-related microarray matrix GSE103552 included 17 controls and 20 GDM samples. There were 118 differentially expressed genes，including 65 up-regulated cases and 53 down-regulated cases. Differentially expressed genes were enriched in extracellular matrix tissues，extracellular structural tissues and DNA conformation change response，and they regulated RNA transport and taste transduction. Totally，33 characteristic difference genes. Protein interaction network construction identified 5 hub genes（ANKRD36C，CLK1，LUC7L3，NKTR and RSRP1），the 5 hub genes were highly expressed in differentially expressed genes，and the area under curve for diagnosing GDM was >0.8. Conclusions Based on GEO database，the 5 hub genes are highly expressed in differentially expressed genes，providing potential diagnostic，prognostic markers and therapeutic targets for GDM.

Key words: Gestational diabetes mellitus, Gene expression omnibus database, Hub genes, Diagnostic marker