Mixture models for analysing transcriptome and ChIP-chip data – Marie Laure Martin Magniette

Mixture models are useful for identifying underlying structures. In such models, the density of the observations is modelled by a weighted sum of parametric density (e.g. each component is a Gaussian distribution) and each one represents a subpopulation composed of observations sharing common characteristics. The first part of my talk will be dedicated to a presentation of the mixture models. I will explain the concept and the outputs of an analysis based on a mixture through easy examples. In the second part of my talk, I will show how mixture models can be applied to analyze transcriptomic (co‐expression analysis of Arabidopsis thaliana genes) and chIP‐chip data (detection of enriched regions and of differentially methylated regions).


Comments are closed.