Probabilistic models and Bayesian inference
Author(s): Magnus Rattray
Affiliations: The University of Manchester
Keywords: probablistics models; tutorial
Probabilistic models are models of data in which the parameters and the data are both treated as random variables. They have already been applied extensively in genomics and functional genomics, and more recently have also been applied to problems in systems biology. Bayesian inference methods provide a useful approach for estimating the model parameters, assessing different models and making predictions or inferences from data. Bayesian methods are often computationally demanding and approximations or sampling-based approaches are required in practical applications. In this tutorial I will describe some of the methods that can be used for practical Bayesian inference and describe some recent applications to gene expression data analysis and to models of transcriptional regulation.