Establishment and Verification of a Predictive Model for Preeclampsia Based on Bioinformatics Analysis

Preeclampsia (PE) has an increasing incidence worldwide, and there is no gold standard for prediction. Recent progress has shown that abnormal decidualization and impaired vascular remodeling are the keys to the pathogenesis of PE. Therefore, it is of great significance to analyze the decidua basalis and blood changes of PE to explore new methods. Here, we performed weighted gene co-expression network analysis (WGCNA) based on 9553 differentially expressed genes of decidua basalis data from Gene Expression Omnibus to generate and screen relevant module-eigengenes (MEs). Among them, MEblue and the MEgrey were the most correlated with PE characteristics, which contained 371 core genes. Subsequently, we applied the logistic least absolute shrinkage and selection operator (LASSO) regression, screened 43 genes most relevant to prediction from the intersections of the 371 genes and blood sample genes (training set), and built a predictive model. The specificity and sensitivity of the model were illustrated by receiver operating characteristic curves, and the stability was verified by the other two sets of samples (validation sets). The results demonstrated that our predictive model showed good predictions, with an area under the curve of 0.991 [94.4%, 100.0%] for the training set, 0.874[80.9%,78.7%] and 0.986 [97.9%,91.7%] for the validation sets. Finally, we found the 43 key marker genes in the model were closely associated with the clinically accepted predictive molecules, including FLT1 , PIGF , ENG and VEGF , although their AUCs were not very high. Therefore, this predictive model provides a new potential approach for the diagnosis and treatment of PE.