Pregnancy

Understanding pregnancy trajectories: from early biological changes to mechanism and prevention

Pregnancy orchestrates a rare physiological transformation across vascular, immune, and metabolic systems. Pregnancy complications do not suddenly appear when they are diagnosed; they may reflect biological trajectories that begin to diverge weeks or months earlier.

We develop computational and AI approaches to identify these deviations, understand the biology that drives them, and ultimately translate early molecular and physiological signals into opportunities for prediction and prevention.

From Trajectory to Prevention

Publications
Pregnancy as a Dynamic Biological Trajectory

Rather than treating pregnancy as a series of isolated measurements, we model it as a continuously evolving biological system.Using longitudinal proteomic, metabolomic, genomic, microbial, imaging, and clinical data, we ask when a pregnancy first departs from a healthy trajectory, which systems change first, and whether different adverse outcomes arise through shared or distinct molecular pathways.

In our large longitudinal pregnancy cohorts, repeated sampling allows us to observe molecular biology months before the clinical outcome is known.

Visionary AI - The Eye as a Window Into Vascular Health

Human milk contains thousands of molecular components, including metabolites, proteins, micronutrients, immune factors, microbes, and signaling molecules. Yet we still understand surprisingly little about how these components are regulated or how they reflect mammary gland physiology.

Across diverse populations worldwide, we study how milk composition varies with:maternal nutrition and metabolismstage of lactationmammary gland functioninfant growth and developmentearly-life microbiome establishment

• Maternal nutrition and metabolism

• Stage of lactation

• Mammary gland function

• Infant growth and development

• Early-life microbiome establishment

From Imaging and Multi-Omics to Mechanism

Prediction is only the beginning. Once we identify an abnormal molecular or physiological trajectory, we want to understand why it occurs.

We are developing APOLLO — Adverse Pregnancy Outcome Learning through causaL multi-Omic Search — to integrate empirical multi-omic signals with biological prior knowledge and reconstruct candidate mechanistic pathways underlying pregnancy complications. APOLLO combines proteomics, metabolomics, genomics, molecular interaction networks, and pregnancy-specific biological knowledge to generate interpretable, testable mechanistic hypotheses.  

APOLLO uses network-based optimization to connect observed molecular changes into parsimonious mechanistic models, producing predicted pathway activity, candidate interactions, and supporting biological evidence. This allows us to move beyond lists of biomarkers toward questions such as:

• Which molecular pathways drive the earliest divergence from healthy pregnancy?

• Do clinically similar outcomes arise through biologically distinct mechanisms?

• Which proteins and metabolites are upstream drivers rather than downstream consequences?

• Which components of an abnormal trajectory may be therapeutically modifiable?

Toward Computationally Guided Intervention

Beyond identifying biomarkers and mechanisms, we are beginning to explore how computational models can help identify potentially modifiable pathways and prioritize candidate interventions.

By integrating disease-associated molecular signatures with perturbation data and biological knowledge, we aim to generate mechanistically grounded hypotheses that can be tested experimentally in relevant placental models.

From Pregnancy to Future Health

Pregnancy is not only a window into immediate maternal and fetal health. The vascular and metabolic adaptations of pregnancy may also reveal biological vulnerability that persists long after delivery.

By connecting pregnancy trajectories with postpartum and long-term outcomes, we aim to understand how early vascular, molecular, and metabolic signatures relate to later cardiovascular and metabolic health.