Developing computational models to understand aging

Abstract

Aging in biological organisms is a complex process, involving changes at all levels of functioning. No single pathway or mechanism is responsible for aging, leading to the current understanding that aging is due to a number of interacting biological factors. To understand this interconnected complex process, this thesis develops complex computational models of aging. Using human data I develop network models of aging, which model the aging process as a network of interacting components. These models are used to understand the network structure of different aspects of health, as well as make quantitative predictions of aging health outcomes and mortality. Using worm data I develop an aging trajectory clustering model, which describes the dynamics of worm aging with a low-dimensional latent space that exhibits simple dynamics and clear clustering. This model is used to infer distinct worm aging phenotypes. Using mice and human data, I develop a method to extract damage and repair processes in aging. This approach is used to study the effects of age and interventions on the processes of damage and repair.

This work is an attempt to build computational models of aging, and demonstrates the potential of these types models in the study of aging in the future.