A Mixture Model-Based Approach to the Clustering of Microarray Expression Data
Clusters tissues using all group means
Clusters genes using mixtures of normal distributions
Clusters tissues
EMMIXgene:
Heat maps
Plot a single gene expression histogram with best fitted mixture of t-...
Selects genes using the EMMIXgene algorithm.
Cluster tissues
Provides unsupervised selection and clustering of microarray data using mixture models. Following the methods described in McLachlan, Bean and Peel (2002) <doi:10.1093/bioinformatics/18.3.413> a subset of genes are selected based one the likelihood ratio statistic for the test of one versus two components when fitting mixtures of t-distributions to the expression data for each gene. The dimensionality of this gene subset is further reduced through the use of mixtures of factor analyzers, allowing the tissue samples to be clustered by fitting mixtures of normal distributions.