Integrated Protein Network Analysis of Whole Exome Sequencing of Severe Preeclampsia

Abstract

To identify clusters of patients with shared networks of genes associated with early onset severe preeclampsia from whole exome sequencing data through novel bioinformatic analysis.We performed a case-control study using whole exome sequencing (WES) on early onset preeclamptic mothers with severe features delivering < 34 weeks and mothers who delivered ≥ 37 weeks. Genotype testing identified variants that were differentially abundant between cases and controls. A Protein-Protein interaction (PPI) analysis and visualization tool, Proteinarium, was implemented to identify clusters of patients with shared networks associated with severe preeclampsia.A total of 61 early onset preeclamptic women with severe features and 82 race and ethnicity matched control women at term were sequenced. We identified 8,867 predicted deleterious variants. 21 of these variants were nominally associated with preeclampsia by genotype testing. Using Proteinarium129 out of the 143 sequenced patients were assigned to statistically significant clusters, Cluster A and B (p< 0.0001). Case dominated Cluster A contained 47 of the 61 case subjects. There were 13 unique genes in the PPI network of Cluster A compared to control dominated Cluster B. Amongst these unique genes, LAMB2, PTK2, RAC1, QSOX1, FN1, and VCAM1 have known associations with the pathogenic mechanisms of preeclampsia.Our network analysis identified genes that were unique to a large cluster of patients with shared networks that provide insights for severe preeclampsia. We also identified genes imputed from the interactome that may otherwise have not been identified by conventional analysis. Strict phenotyping of both cases and controls improved the likelihood of identifying these otherwise difficult to find genetic associations. These uniquely identified genes and their associated variants are potential candidates for developing polygenic risk scores for severe preeclampsia. These results support our hypothesis on the genetic architecture of complex diseases and are generalizable to other phenotypes.