DIRECT-package {DIRECT} | R Documentation |
This package implements the Bayesian clustering method in Fu et al. (2013). It also contains a simulation function to generate data under the random-effects mixture model presented in this paper, as well as summary and plotting functions to process MCMC samples and display the clustering results. Replicated time-course microarray gene expression data analyzed in this paper are in tc.data
.
This package three sets of functions.
Functions DIRECT
and others for clustering data. They estimate the number of clusters and infers the partition for multivariate data, e.g., replicated time-course microarray gene expression data. The clustering method involves a random-effects mixture model that decomposes the total variability in the data into within-cluster variability, variability across experimental conditions (e.g., time points), and variability in replicates (i.e., residual variability). The clustering method uses a Dirichlet-process prior to induce a distribution on the number of clusters as well as clustering. It uses Metropolis-Hastings Markov chain Monte Carlo for parameter estimation. To estimate the posterior allocation probability matrix while dealing with the label-switching problem, there is a two-step posterior inference procedure involving resampling and relabeling.
Functions for processing MCMC samples and plotting the clustering results.
Functions for simulating data under the random-effects mixture model.
Package: | DIRECT |
Type: | Package |
Version: | 1.0 |
Date: | 2011-10-19 |
License: | GPL (>=2) |
LazyLoad: | yes |
See DIRECT
for details on using the function for clustering.
See summaryDIRECT
, which points to other related plotting functions, for details on how to process MCMC samples and display clustering results.
See simuDataREM
, which points to other related functions, for simulating data under the random-effects mixture model.
Audrey Qiuyan Fu
Maintainer: Audrey Q. Fu <audreyqyfu@gmail.com>
Fu, A. Q., Russell, S., Bray, S. and Tavare, S. (2013) Bayesian clustering of replicated time-course gene expression data with weak signals. The Annals of Applied Statistics. 7(3) 1334-1361.
DIRECT
for the clustering method.
summaryDIRECT
for processing MCMC estimates for clustering.
simuDataREM
for simulating data under the mixture random-effects model.
## See examples in DIRECT and simuDataREM.