Preview

PDF, English (Thesis dissertation) - main document Download (7MB) | Lizenz: Rights reserved - Free Access

Abstract

In this thesis, I study how biological prior knowledge and high throughput biological data can be systematically integrated to yield mechanistic biological insights. I focused the scope of my work mainly on signalling pathways and metabolism, especially how these two biological functions interact and control each other. The overall goal of this work is to better characterise the molecular driver of complex diseases and chronic health conditions such as cancer, metabolic syndromes and fibrosis. Indeed, if we can better and more systematically understand these conditions, we may be able to design better, more targeted treatments and even prevent them more efficiently. In the first chapter, I draw a state of the art of multi-omic data generation and how to analyze them in mechanistic contexts. What we call omic data are datasets where the abundance of hundred to thousand unique biological molecules are measured in parallel. Then, in the second chapter, I present a collection of scientific studies where I could learn and apply the principles detailed in the first chapter. In the third chapter, I present my attempt at developing a way to systematically analyse and integrate multiple types of omic data together. The resulting tool, named COSMOS, is presented in the context of a kidney cancer study using multiple types of omic data generated from a cohort of patients. In the final chapter, I present a tool called ocEAn, which aims at estimating metabolic enzyme activity changes from metabolomic data.