Author ORCID Identifier
Document Type
Dissertation
Date of Award
2025
Degree Name
Doctor of Philosophy (PhD)
Department
Biomedical Engineering
First Advisor
Etienne Z. Gnimpieba
Abstract
Heterogeneity and rarity can make cancer difficult to study and treat. Cancer's heterogeneity, evident in tumor locations, tumor cell types, subtypes, and microbiota within the tumor microenvironment, complicates diagnosis, prognosis, and treatment. Research into the tumor microenvironment and its microbiota is an evolving area of oncology that may advance cancer drug-patient response. Cancer-omics signatures are crucial to drug response, yet there is no tool that physicians use as a gold standard. Research indicates that microbial presence can alter drug metabolism and immune response, affecting cancer-drug efficacy on a patient-specific basis. This work highlights the need for training readiness in transdisciplinary research. It utilizes the research problem of combining oncology and biology, along with data science and systems biology, to provide a convergence research framework addressing this issue. Utilizing a problem-based approach, this dissertation develops bioinformatics workflows for transcriptomics, RNA sequencing and single-cell RNA sequencing, and metagenomics, which are key components of this work. The resulting analyses of three use cases within biomedical research informed future decisions regarding workflow development. For initial model training, RNA sequencing and computationally derived microbial taxonomic abundances were utilized. While each of these data modalities addresses only a portion of the systems biology challenge, their integration led to the development of the initial version of the Drug and Immunotherapy Response Prediction System, achieving 85% accuracy. This system aims to predict cancer treatment responses within a precision biomedicine framework. Ultimately, this work establishes the foundation for Artificial Intelligence and Machine Learning-based predictions of the cell-microbe impact on tumor therapy response, paving the way for enhanced treatment in precision oncology and potentially transforming clinical decision-making in cancer care.
Subject Categories
Artificial Intelligence and Robotics | Bioinformatics | Education
Keywords
Convergence Research Drug Response Prediction Intratumoral Microbiome Genomic Heterogeneity Molecular Signature Predictive Modeling TME Drug Response
Number of Pages
232
Publisher
University of South Dakota
Recommended Citation
Zylla, Jessica, "CONVERGENCE IN CANCER: INTEGRATIVE MULTI-MODAL MODELING FOR PAN-CANCER DRUG RESPONSE PREDICTION" (2025). Dissertations and Theses. 375.
https://red.library.usd.edu/diss-thesis/375