Enhancing insight into placental dysfunction in common obstetric disorders using placental multiomics

Project Summary Adverse obstetrics disorders, including preeclampsia, preterm birth, and fetal growth restriction, are common worldwide. However, they are difficult to study given multifactorial etiologies and co-occurrence of disorders. Placental dysfunction has a role in these diseases but the underlying molecular

mechanisms are not well understood. Integrated multiomic analysis can address this gap by identifying underlying molecular mechanisms associated with individual disorders and combinations of disorders. The goal of this study is to advance mechanistic understanding of the molecular networks associated

with obstetric disorders, to identify and prioritize new directions for future research for improving maternal-fetal health. We propose to find patterns at the individual, network, and system levels that shed light on mechanisms of disease associated with normal placental physiology and placental dysfunction.

The Sadovsky lab has provided de-identified metabolomic, proteomic, and transcriptomic placental data paired with histopathology reports and clinical data including demographics, routine clinical labs, maternal comorbidities, and delivery records. These data come from 333 placentas from people with singleton pregnancies, including uncomplicated term pregnancy, fetal growth restriction, preeclampsia,

fetal growth restriction with a hypertensive disorder, and spontaneous preterm birth. The analyses in this proposal are complementary to, and distinct from, those in a separate K99. In Aim 1, we will define placental physiological association networks to determine placental regulation. We have preliminary data

for pairwise associations between analytes and placental histopathological features using generalized linear models. We will conduct community structure analysis on these associations to identify subnetworks of placental regulation and evaluate differences in network identity and structure. For each

obstetric disorder, we will also evaluate the contribution of top interconnected community networks to regulation of placental physiology by evaluating model robustness and multicollinearity. This will provide domain-agnostic detection of patterns of systems-level placental regulation. In Aim 2, we will determine

molecular network differences among obstetric disorders. We have evaluated pairwise differences between individual analytes between phenotypes, and we will build on this work by evaluating placental network differences between different obstetric outcomes in two ways: determining the network structure

differences between phenotypes, and building and evaluating classification models for obstetric disorders. We will then perform case-study outlier analysis to provide potential molecular insight into placental dysfunction on the individual level. This will provide molecular network understanding of placental regulation and how it’s disrupted in obstetric disorders at the population and individual level.