Predicting of disease genes for gestational diabetes mellitus based on network and functional consistency

Gestational diabetes mellitus (GDM) is a world-widely prevalent disease with adverse outcomes. This study aims to identify its disease genes through bioinformatics analysis.

The raw gene expression profiling (ID: GSE19649) was downloaded from Gene Expression Omnibus database, including 3 GDM and 2 healthy control specimens. Then limma package in R was utilized to identify differentially expressed genes (DEGs, criteria: p value <0.05 and |log2 FC|>1). Simultaneously, known disease genes of GDM were downloaded from Online Mendelian Inheritance in Man database. Then, DEGs and known disease genes were uploaded to STRING to investigate their protein-protein interactions (PPIs). Gene pairs with confidence score >0.8 were utilized to construct PPI network. Furthermore, pathway and functional enrichment analyses were performed through KOBAS (criterion: p value <0.05) and DAVID (The Database for Annotation, Visualization and Integrated Discovery) software (criterion: false discovery rate <0.05), respectively.

A total of 404 DEGs were identified, including 273 up-regulated and 131 down-regulated DEGs. Moreover, 68 known disease genes of GDM were obtained. Then, 190 gene pairs were identified to significantly interact with each other. After deleting PPIs between DEGs, PPI network was constructed, consisting of 115 gene pairs. Furthermore, genes in PPI network were significantly enriched in 10 functions and 8 pathways.

Based on PPI network and functional consistency, 6 candidate genes of GDM were considered to be candidate disease genes of GDM, including CYP1A1, LEPR, ESR1, GYS2, AGRP, and CACNA1G. However, further studies are required to validate these results.