Developing Interpretable Data Mining Frameworks for Addressing Biomedical Challenges in Genomics and Imaging

The global burden of disease has evolved significantly in recent decades. Since the early 2000s, we have witnessed multiple widespread outbreaks of respiratory viruses including Influenza A and SARS-CoV-2, alongside a marked increase in chronic conditions such as diabetic retinopathy (DR). These diseases pose serious health threats – respiratory infections can progress to fatal multi-organ failure, while untreated DR may result in vision impairment or complete blindness. Given the impacts of these conditions on individual quality of life and healthcare systems worldwide, it is imperative to develop a comprehensive set of computational methods for studying and detecting these diseases. This dissertation highlights the need for accessible and interpretable data visualization techniques and models in studying respiratory viruses and DR. My work extends previous approaches that have been developed for studying SARS-CoV-2 host-pathogen interactions, diagnostic assays for respiratory virus detection, and analysis of spatiotemporal trends in SARS-CoV-2 variant transmission. My work also focuses on developing explainable deep learning models that extract latent and explicit information from retina fundus images to detect DR and identify DR stages. The methods discussed in this dissertation contribute to the research community by consolidating information and extracting hidden insights from publicly available data sources through interpretable data visualization approaches and models. Moreover, my efforts in DR detection and grading demonstrate the utility of using clinically grounded data in both model design and post-model decision analysis. Taken together, these approaches show how novel computational methods and models can be used to provide valuable insights and drive innovation in diverse biomedical domains.

Author dc:creator

• Gill, Hunter Mathias

Advisor dc:contributor.advisor

• Janga, Sarath Chandra

• Bioinformatics
• Diabetic Retinopathy
• Explainable AI
• Interpretability
• Respiratory Viruses
• SARS-CoV-2

Language dc:language.iso

en_US

Identifier URI

https://doi.org/10.7912/MGBG-W385

OAI identifier oai:identifier

oai:scholarworks.indianapolis.iu.edu:1805/52102