Katte, P. (BMEB) – Interactive and Scalable Frameworks for Pathogen Surveillance and Ancestral Recombination Graph

The explosive growth of genomic data, driven by advances in sequencing and inference technologies, presents both an opportunity and a challenge for evolutionary biology and public health. Existing visualization and analysis tools often fall short in handling the scale, complexity, and uncertainty of modern genomic datasets—especially in the areas of pathogen surveillance and ancestral recombination inference. This thesis introduces new tools that provide scalable visualization and analysis to bridge these gaps and enable more interpretable and actionable genomic insights.

First, I develop an interactive dashboard within a tool called WEPP for wastewater-based pathogen surveillance. It combines phylogenetic placement with intuitive web-based visualization, allowing public health officials to track variant spread at high resolution. Second, I build Lorax, a browser-based platform for visualizing Ancestral Recombination Graphs (ARGs) at biobank scale. Lorax incorporates a multi-agent system that supports natural language querying, code generation, and interactive tree exploration. Finally, I introduce a novel inference framework based on Generative Flow Networks to sample from posterior distributions over ARGs, addressing key limitations in uncertainty quantification and scalability found in existing methods. Together, these tools aim to make the study of evolution and disease more accessible and effective, helping researchers and public health teams draw clearer conclusions from complex genetic data.

Event Host: Pratik Katte, PhD Student, Biomolecular Engineering & Bioinformatics

Advisor: Russ Corbett-Detig