Identification of key extracellular proteins as biomarkers for prediction and assessment of preeclampsia

Background Preeclampsia (PE) is a major cause of maternal and neonatal death, and its pathogenesis is related to extracellular proteins (EPs) secreted by the placenta. However, there are no objective indicators for the diagnosis and treatment of PE. We hope to contribute to the clinical work by studying the role of extracellular proteins in PE. Methods Differential expression analysis and WGCNA were used to preliminary screening the extracellular proteins and differential expression genes (EP-DEGs). Machine learning algorithms were used to further identify key EP-DEGs. GO and KEGG were used to analyze the function and pathway of EP-DEGs. Immune infiltration, ROC curve and correlation analyses were performed to assess EP-DEGs and immune cells, diagnostic and prognostic abilities respectively. Results 245 up-regulated and 233 down-regulated DEGs in GSE75010 were found and four thousand and six EPs were gained from HPA and Uniprot. 172 EP-DEGs were selected from the intersection of DEGs and EP. 5 genes (FSTL3, FLNB, P4HA1, CST6, EFNB1) were defined as the hub EP-DEGs by taking intersection of 6 candidate genes from Lasso regression algorithm and the top 10 most important genes from RF algorithm. Conclusion This study identified 5 key genes in PE and performed ROC curve and correlation analyses, indicating that EPs play important roles in signaling, cell adhesion, inflammation and immune response in PE.