Optimized Workflow for Microbiome-Based Survival Analysis

Microbiome-based survival analysis is used to study the association between microbiome and clinically relevant time-to-event outcomes, such as disease onset, relapse, or treatment response. However, microbiome sequencing data are inherently compositional, high-dimensional, and sparse. This thesis develops and evaluates an optimized workflow for microbiome-based survival modeling, focusing on how alternative microbiome data structures and abundance transformations affect Cox proportional hazards model behavior, and predictive performance. Using a publicly available dataset of non-obese diabetic (NOD) mice originally analyzed by (Pujolassos et al., 2024).