Skip to Main content Skip to Navigation
Theses

Heuristiques pour la résolution du problème d'alignement multiple

Abstract : Multiple alignment is one of the basic and central tasks in Bioinformatics which tries to higlight similarities between sequences. It is a prior to phylogeny reconstruction, pattern matching and protein structure prediction. This thesis aims to develop new algorithms to tackle this problem, and it introduces two new progressive algorithms that align alignments instead of profiles.
The first algorithm, Plasma I, uses a descent where each iteration consists in gap insertions in one of the two multiple alignments. The second algorithm, Plasma II is based on the Dynamic Programming principle. We propose with this algorithm a generalization of the pairwise sequence alignment algorithm, and we extend the Dynamic Programming framework to align two multiple alignments. The performances of Plasma II are assessed on the well-known Balibase benchmarks, and compared with several algorithms. Plasma II provides results of quality with fast computation time.
Document type :
Theses
Complete list of metadatas

Cited literature [34 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00352784
Contributor : Anne-Marie Plé <>
Submitted on : Tuesday, January 13, 2009 - 5:09:37 PM
Last modification on : Monday, June 1, 2020 - 4:54:02 PM
Long-term archiving on: : Tuesday, June 8, 2010 - 5:17:41 PM

File

Identifiers

  • HAL Id : tel-00352784, version 1

Citation

Vincent Derrien. Heuristiques pour la résolution du problème d'alignement multiple. Informatique [cs]. Université d'Angers, 2008. Français. ⟨tel-00352784⟩

Share

Metrics

Record views

322

Files downloads

2703