From Genomic to Functional models
Résumé
This document encapsulates the research that I have pursued since 2003 and recaptures the mainresults during this period that were obtained by myself and the students that I have supervised. Ithas benefited from strong collaboration with fellow researchers, as can be seen in the complete listof publications (see page 105). It is organized in four chapters having distinct objectives, and thusvarying in form and style.The first chapter – bearing a quite original title Introduction – has in mind an ingenuous readerwho is not necessarily familiar with the general research domain of bioinformatics and systemsbiology. There is absolutely no claim to provide an introduction to these fields, but rather to set somecommon reference points that determine the boundaries of my own work.The second chapter treats a classical but essential topic in bioinformatics, that of uncovering thecommon points between related genomes. Precisely, here we study the question of building highqualityprotein families in a computationally efficient way. We then address the question of detectinggene fusion and fission, a case where alignment-based computation of families will fail.The third chapter concerns a slightly less classical approach for studying evolution by means ofrearrangement analysis. Here we first revisit the theory of computation of rearrangement analysis,second propose a new method of analysis of ancestral architectures that reconciles both breakpointand rearrangement analyses, and third introduce optimization algorithms that can address this problemin practice.In the last chapter I describe the work in construction of dynamic models of living organisms thatexhibits my enduring interest in systems biology. I do not consider the boundary between bioinformaticsand systems biology by any means a rigid one: both disciplines essentially being concernedwith building models of living organisms. We propose here a novel high-level modeling frameworkintegrating discrete and continuous multi-scale dynamics within the same semantics domain. I concludewith experimental results using a software implementation of this framework.
Domaines
Informatique [cs]
Loading...