Genetic co-occurrence networks identify polymorphisms within ontologies highly associated with preeclampsia.

Polygenic diseases require the co-occurrence of multiple risk variants to initiate a pathology. An example is preeclampsia, a hypertensive disease of pregnancy with no known cure or therapy other than the often-preterm delivery of the neonate and placenta. Preeclampsia is challenging to predict due to symptomatic and outcome heterogeneity. Transcriptomic and genetic analysis suggests that this is a multi-syndromic and multigenic disease. Previous research applications of traditional GWAS methods to preeclampsia identified only a few alleles with marginal differences between cases and controls. We seek to identify genetic networks related to the pathophysiology of preeclampsia as potential avenues of therapeutic investigation and early genetic testing. We created a novel systems biology-based method that identifies networks of co-occurring SNPs associated with a trait or disease. The co-occurring pairs are assembled into higher-order associations using network graphs. We validated our method using simulation modelling and tested it against maternal genetic data of a previously assessed preeclampsia cohort. The genetic co-occurrence network identified SNPs in or near genes with ontological enrichment for VEGF, immunological and hormonal pathways, among others with known physiological disruption in preeclampsia. Our findings suggests that preeclampsia is caused by relatively common alleles (<5%) that accumulate in unfavorable combinations.