Abstract

The tumor microenvironment (TME) plays a critical role in driving cancer progression. However, the complexity and heterogeneity of the TME—in terms of cell types and spatial context—have posed significant challenges for cancer characterization. In this dissertation, I present several tools and methods for spatial and molecular quantification of the TME across several cancer types. First, I introduce Pollock, a deep learning-based classifier for single-cell data that offers high accuracy, broad cross-platform compatibility, and biologically interpretable outputs. I then apply single-cell and bulk transcriptomic analyses to pancreatic ductal adenocarcinoma (PDAC), identifying tumor ecotypes and a continuum of developmental stemness in malignant and stromal cells that predict patient survival. Finally, I spatially extend this approach to breast and prostate cancer, integrating spatial transcriptomics, multiplexed imaging, and 3D volumetric reconstruction to uncover genes that characterize the precancer-to-cancer transition.

Committee Chair

Li Ding

Committee Members

Aadel Chaudhuri; Kooresh Shoghi; Stephen Oh; Tau Ju

Degree

Doctor of Philosophy (PhD)

Author's Department

Biology & Biomedical Sciences (Computational & Systems Biology)

Author's School

Graduate School of Arts and Sciences

Document Type

Dissertation

Date of Award

5-6-2025

Language

English (en)

Author's ORCID

https://orcid.org/0000-0002-8041-0864

Available for download on Wednesday, May 05, 2027

Included in

Biology Commons

Share

COinS