Skip to Main content Skip to Navigation
Theses

Prédiction markovienne in silico des régions constantes et variables des lentivirus

Abstract : Lentiviruses exhibit a considerable plasticity of the env gene, particularly in the region which encodes the surface (SU) glycoprotein. Interestingly, the SU mutations are clustered in restricted areas, called variable (V) regions and interspaced by more stable regions, called constant regions (C). To determine if the C and V regions are characterized by specific signatures, we developed HMM (Hidden Markov Models), based on the statistical oligonuclotide composition of each type of region, aimed at differentiating the C and V regions of the lentiviruses. We trained hidden Markov models on SU sequences of the equine, human, simian and small ruminant lentiviruses. We obtained a clear delimitation of the C and V regions of all these lentiviruses and of bovine and feline lentiviruses which were not used to define the models. Our results suggest that the C and V regions of the lentiviruses have statistical differences in their composition in words of nucleotides or amino-acids. Specific signatures of the C and V regions have been extracted from our models.
Document type :
Theses
Complete list of metadatas

Cited literature [131 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00124142
Contributor : Aurélia Quillon <>
Submitted on : Friday, January 12, 2007 - 2:14:13 PM
Last modification on : Wednesday, October 14, 2020 - 3:44:18 AM
Long-term archiving on: : Tuesday, April 6, 2010 - 10:00:41 PM

Identifiers

  • HAL Id : tel-00124142, version 1
  • PRODINRA : 252032

Collections

Citation

Aurélia Quillon. Prédiction markovienne in silico des régions constantes et variables des lentivirus. Mathématiques [math]. Université Claude Bernard - Lyon I, 2006. Français. ⟨tel-00124142⟩

Share

Metrics

Record views

534

Files downloads

1615