Séminaire de Probabilités et Statistique :

Le 17 mars 2014 à 15:00 - CIRAD, campus de Lavalette, amphi Jacques Alliot

Présentée par Tadesse Mahlet G. - Geogetown University, Washington DC

Bayesian methods for integrative genomic data analysis

There is a growing interest in relating data sets from various genomic technologies with the hope of gaining a better understanding of the complex biological processes underlying various phenotypes. I will present Bayesian methods we have proposed to address this problem. In the first part of the talk, I will present a stochastic partitioning method that combines ideas of mixtures of regression models and variable selection to relate two high-dimensional data sets. This provides a unified framework to uncover correlated expression profiles and identify their associated subset of markers. In the second part of the talk, I will discuss an extension of the method in the context of copy number variant (CNV) analysis, where we address the CNV detection and gene expression association analysis in a unified manner. This is accomplished by specifying a measurement error model that relates the gene expression levels to latent copy number states, which in turn are related to the continuous surrogate CNV measurements via a hidden Markov model. I will illustrate the methods with applications to genomic studies.