Identification and validation of an angiogenesis-related signature associated with preeclampsia by bioinformatic analysis

Preeclampsia (PE) is a pregnancy disorder with high morbidity and mortality rates for both mothers and newborns. This study explores potential diagnostic indicators of PE. We downloaded the messenger ribonucleic acid profiles of the GSE75010 dataset from the Gene Expression Omnibus database, and used placenta samples to carry out different analyses including differential expression, Gene Ontology, and Kyoto Encyclopedia of Genes and Genomes analyses. Least absolute shrinkage and selection operator regression was constructed and the receiver operating characteristic curve was drawn to evaluate the accuracy of the model. An external validation was conducted to prove the stability of the risk model. We found 140 angiogenesis-related genes and identified 29 angiogenesis-related genes between the 2 groups, including 12 upregulated genes and 17 downregulated genes. In addition, we established a 12-gene risk signature, which has a high accuracy in predicting PE during pregnancy (area under curve = 0.90). The immune infiltration characteristics are differentially distributed in the 2 groups, which may be the cause of hypertension during pregnancy. The external validation with the GSE25906 dataset confirmed the high accuracy of our model (area under curve = 0.87). Our results outline the characteristics of a set of genes potentially involved in PE and its subgroups, contributing to a better understanding of the molecular mechanisms of PE.