Development of statistical methods to detect epistasis in Genome-Wide Association Studies.

Co-direction : Christophe Ambroise et Cyril Dalmasso

manuscrit.pdf

slides_soutenance.pdf

• Genome-Wide association studies.

• Epistasis, gene-gene interaction.

• Structured penalized regression models

• Group Lasso Gene-Gene Eigen Epistasis (GGEE) :

The GGEE package implements the group lasso Gene-Gene Eigen Epistasis (G-GEE) method for detecting epistasis at the gene level. The proposed functions allow to compute interaction variables for each pair of gene and to fit a general model with a Group Lasso penalty. The package allows to generate gene structured genotype data and continuous phenotype. The package is available from github and can be installed as follows from within R:

# install.packages("devtools")
devtools::install_github("vstanislas/GGEE")

• Annual Conference of the International Society for Clinical Biostatistics (ISCB), Vigo, Spain, July 2017

• Statistical Methods for Post Genomic Data (SMPGD), London, England, January 2017

• Journées Ouvertes en Biologie, Informatique et Mathématiques (JOBIM), Lyon, June 2016

• Journées de Statistique (JdS), Montpellier, June 2016

• DAGStat 2016 - “Statistics under one umbrella”, Göttingen, Germany, March 2016

• Statistical Analysis of Massive Genomic Data (SAMGD), Évry, November 2015

• Statistique Mathématique et Applications, Fréjus, September 2015

• L2 Bio : TD Statistique

• L1 Math-Eco : TD Analyse Réelle