PhD Thesis Defense: Rachel Huang

ZOOM LINK
Meeting ID: 997 5379 5294
Passcode: 161899

"From Spatial Omics to Drug Target Discovery"

Abstract

The discovery of novel drugs relies on the identification of mechanistically relevant molecular targets. While target identification and associated hallmark molecular signatures have been accelerated through the use of traditional transcriptomics, spatial transcriptomics promotes target discovery in a unique manner by preserving the spatial context for determining the true pharmacological relevance of a molecular target.

However‚ to date, there are no systematic computational methods for transforming complex omics data into drug targets‚ and this is especially the case in reproductive biology.

To bridge this translational gap, this dissertation establishes an end-to-end pipeline that advances spatial biology from purely descriptive mapping to actionable drug discovery, with a focus on non-hormonal contraception and ovulation. Recognizing that high-fidelity data is a prerequisite for accurate target discovery, we first built a robust foundation by systematically benchmarking spatial transcriptomics platforms. This enabled the construction of the first high-resolution spatiotemporal atlas of the ovulatory cascade, detailing the complex cell-cell communication networks preceding follicle rupture. Then, to overcome the field’s computational bottleneck, we developed a scoring framework, an innovative algorithm that rigorously integrates spatial specificity, temporal dynamics, and tissue-wide safety profiles. This framework successfully translated omics datasets and available databases into a prioritized hierarchy of therapeutic targets, pinpointing and validating CHST15 as a highly promising candidate for non-hormonal contraception. Completing this translational loop, we advanced CHST15 into a computer-aided drug design workflow. High-throughput virtual screening and molecular dynamics simulations confirmed the target’s structural druggability and yielded a select group of stable, high-affinity small-molecule inhibitors. These in silico findings serve as definitive proof of concept for target translation, providing robust hit compounds ready for downstream DNA-encoded library screening and in vitro functional assays.

Ultimately, this dissertation establishes a cohesive pipeline bridging spatial transcriptomics, algorithmic target prioritization, and virtual drug design, demonstrating that spatial data can be rationally mined to guide tangible drug development. This work not only delivers concrete candidate molecules for non-hormonal contraception but also provides a scalable blueprint for leveraging spatial omics to address critical, unmet medical needs in women’s health and beyond.

Thesis Committee

• Emme Burgin (Chair)
• Margaret Ackerman
• Aaron McKenna
• Francesca Duncan (Northwestern University)

Contact

For more information, contact Thayer Registrar at thayer.registrar@dartmouth.edu.