Author ORCID Identifier

0009-0003-5844-4755

Date of Graduation

8-2025

Document Type

Dissertation (PhD)

Program Affiliation

Quantitative Sciences

Degree Name

Doctor of Philosophy (PhD)

Advisor/Committee Chair

Ken Chen

Committee Member

Traver Hart

Committee Member

Christine Peterson

Committee Member

Hind Rafei

Committee Member

Katy Rezvani

Abstract

Recent advances in immunotherapy, including immune checkpoint blockade (ICB) and adoptive cell therapy, face challenges such as resistance and immune-related adverse events, partly due to our limited understanding of the immune signaling pathways. While high- throughput genomic data provide unprecedented resolution into these immune pathways, their full potential is limited by the lack of well-annotated, context-specific immune gene sets. To address this need, I developed a workflow to construct immune gene sets by integrating RNA- seq datasets and performing decomposition. Using this approach, I constructed 28 immune- specific gene sets from 83 bulk RNA-seq datasets and 12 Natural Killer (NK) cell-specific gene sets from 55 scRNA-seq datasets. I have demonstrated their utilities in refining pan- cancer immune subtypes, improving ICB response prediction and cancer survival, annotating spatial niches, and guiding therapeutic strategies to guide NK engineering. To further enhance immune gene set annotations, I have built Immune Cell Knowledge Graphs (ICKGs) for T cells, B cells, NK cells and Macrophages by integrating over 24,000 published abstracts using large language models (LLMs) and Natural Language Processing (NLP). Validated through independent functional omics data, ICKGs were shown to capture context-specific immune information, enabling granular annotations for both experimentally-validated and data-derived immune gene sets, including ones widely used in clinical settings. Our interactive platform (https://kchen-lab.github.io/immune-knowledgegraph.github.io/) facilitates ICKG-based pathway annotations, supporting advancements in immune research and cancer immunotherapy. Together, these resources bridge critical gaps in immune pathway discovery and interpretation, offering powerful tools to understand immune-genomics data, enhance biomarker discovery, and accelerate translational research in cancer immunotherapy.

Keywords

Immunology, Bioinformatics, Gene Programs, Knowledge Graphs

Available for download on Friday, May 22, 2026

Share

COinS