Skip to Main content Skip to Navigation
Master Thesis

Régression fonctionnelle de Poisson pour l'analyse de données de séquençage haut-débit

Abstract : Next Generation Sequencing (NGS) is now widespread for the analysis of genome-wide molecular phenomenon, but also raises methodologicals issues, since reads counts are following a Poisson distribution, and are spatially arranged along the genome. This work introduces a new framework to denoise such data, based on the discrete wavelet transform. It consists in an investigation of the associated statiscal issues, especially heteroscedastical high dimension regression, and the application of this method to experimental data, through the use of cycle spinning and the integration of replicates. This work shows the value of wavelets for analysing NGS data, and may lead to the developpementof a new method for peak calling in Chip-Seq experiments.
Document type :
Master Thesis
Complete list of metadatas

Cited literature [35 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-01296665
Contributor : Agrocampus Ouest <>
Submitted on : Friday, April 1, 2016 - 1:43:03 PM
Last modification on : Monday, July 6, 2020 - 3:38:21 PM
Long-term archiving on: : Monday, November 14, 2016 - 11:59:25 AM

Identifiers

  • HAL Id : dumas-01296665, version 1

Collections

Citation

Cervin Guyomar. Régression fonctionnelle de Poisson pour l'analyse de données de séquençage haut-débit. Sciences du Vivant [q-bio]. 2015. ⟨dumas-01296665⟩

Share

Metrics

Record views

207

Files downloads

188