The Weight of History: Stochastic Trajectories in Mutation, Longevity, and Sport

The organization of complex systems is not random. Components interact through networks whose topology shapes how perturbations accumulate, how states evolve, and how dysfunction emerges over time. Understanding such systems therefore requires understanding network structure -- not just the identity of individual components, but how they are connected, and what those connections make possible or inevitable. The first project investigates somatic mutation accumulation across the human genome during aging. By comparing empirical mutation burdens to rigorous null models, I identify organized vulnerability across genes, tissues, and pathways, revealing how the topology of gene interaction networks determines which components are exposed to damage over a lifetime. The second project applies network medicine to aging biology, integrating genetic and pharmacologic datasets to map longevity mechanisms onto the human interactome and support mechanism-based drug repurposing through network proximity and transcriptional reversal of age-associated signatures. The third project models sequential event dynamics in professional soccer using transformer-based architectures, where the learned attention structure implicitly encodes a dynamic network over prior events, forming the context that is essential for predicting what comes next and understanding how early states shape downstream outcomes. Together, these projects illustrate how network organization, whether explicit in biological systems or emergent in learned models, constrains the trajectories available to complex systems. The frameworks developed here, spanning network analysis, multi-omic integration, and sequence modeling, offer complementary approaches to understanding systems where structure and history jointly determine function.