Aspects algorithmiques de la prédiction des structures secondaires d'ARN

Abstract : In this work we study two algorithmic problems: triangular forms in boolean matrices and spatial structure search problems in molecular biology. We first consider the following problem: Given a square boolean matrix, can we permute its lines to yield a triangular matrix? This problem arises when we compute in place a system of equations. One natural way to solve the problem is to use graph theory, especially bipartite graphs. We show several NP-completeness results and introduce some extensions of the problem. At the RNA level, biological signals are defined by a combination of spatial structures and sequence patterns. Until now, few attempts have been made to develop general purpose search programs that take into account both sequence and structure criteria. Therefore, there is a growing demand for search programs combining both descriptive and programming approaches. With this aim in view, we study combinatorial algorithmic problems on 2-intervals intersection graphs. Recent research considering the topic of intersection graphs of geometric objects shows the relevance of the 2-interval notion. We show that pattern matching in 2-intervals set is an NP-complete problem even when restricted to constrained spatial structures. One important goal of this thesis is to use techniques for the prediction of structural biological patterns, and especially auto-catalytic group 1 introns. Our ORANGE algorithm for predicting auto-catalytic group 1 introns is an improved version of the CITRON algorithm of Lisacek and Michel. It is both time efficient and has good biological results. Moreover, an implementation of the ORANGE algorithm is available on the Internet.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00628623
Contributor : Stéphane Vialette <>
Submitted on : Wednesday, October 5, 2011 - 11:22:28 AM
Last modification on : Wednesday, May 23, 2018 - 3:44:02 PM
Long-term archiving on : Friday, January 6, 2012 - 2:20:56 AM

Identifiers

  • HAL Id : tel-00628623, version 1

Collections

Citation

Stéphane Vialette. Aspects algorithmiques de la prédiction des structures secondaires d'ARN. Informatique [cs]. Université Paris-Diderot - Paris VII, 2001. Français. ⟨tel-00628623⟩

Share

Metrics

Record views

411

Files downloads

171