\name{emfit}
\alias{emfit}
\alias{emfit,ExpressionSet,character,ebarraysPatterns-method}
\alias{emfit,ExpressionSet,ebarraysFamily,ebarraysPatterns-method}
\alias{emfit,matrix,character,ebarraysPatterns-method}
\alias{emfit,matrix,ebarraysFamily,ebarraysPatterns-method}
\alias{show,ebarraysEMfit-method}
\alias{ebarraysEMfit-class}
\title{ Implements EM algorithm for gene expression mixture model }
\description{
  Implements the EM algorithm for gene expression mixture model
}
\synopsis{
emfit(data,
      family,
      hypotheses,
      ...)
}
\usage{
emfit(data,
      family,
      hypotheses,
      cluster,
      type=2,
      criterion="BIC",
      cluster.init = NULL,
      num.iter = 20,
      verbose = getOption("verbose"),
      optim.control = list(),...)
}
\arguments{
  \item{data}{ a matrix}
  \item{family}{
    an object of class ``ebarraysFamily'' or a character string which can
    be coerced to one. Currently, only the characters "GG" and "LNN", and
    "LNNMV" are valid. For LNNMV, a \code{groupid} is required. See below.
    Other families can be supplied by constructing them explicitly.
  }
  \item{hypotheses}{
    an object of class ``ebarraysPatterns'' representing the hypotheses
    of interest. Such patterns can be generated by the function
    \code{\link{ebPatterns}}
  }
  \item{cluster}{if \code{type}=1, \code{cluster} is a vector specifying the fixed cluster
    membership for each gene; if \code{type}=2, \code{cluster} specifies
    the number of clusters to be fitted}
  \item{type}{if \code{type}=1, the cluster membership is fixed as input
    \code{cluster}; if \code{type}=2, fit the data with a fixed number
    of clusters}
  \item{criterion}{only used when \code{type}=2 and \code{cluster}
    contains more than one integers. All numbers of clusters provided in
    \code{cluster} will be fitted and the one that minimizes
    \code{criterion} will be returned. Possible values now are
    "BIC", "AIC" and "HQ"}
  \item{cluster.init}{only used when \code{type}=2. Specify the initial
    clustering membership.}
  \item{num.iter}{ number of EM iterations}
  \item{verbose}{ logical or numeric (0,1,2) indicating desired level of
    information printed for the user}
  \item{optim.control}{ list passed unchanged to \code{\link{optim}} for
    finer control}
  \item{...}{\code{groupid}: an integer vector indicating which group each sample belongs
    to, required  in the ``LNNMV'' model. It does not depend on ``hypotheses''.}
}
\value{
  an object of class ``ebarraysEMfit'', that can be summarized by
  \code{show()} and used to generate posterior probabilities using
  \code{\link{postprob}}
}
\author{Ming Yuan, Ping Wang, Deepayan Sarkar, Michael Newton, and
  Christina Kendziorski}
\references{
  
  Newton, M.A., Kendziorski, C.M., Richmond, C.S., Blattner, F.R. (2001).
  On differential variability of expression ratios: Improving
  statistical inference about gene expression changes from microarray data.
  Journal of Computational Biology 8:37-52.

  Kendziorski, C.M., Newton, M.A., Lan, H., Gould, M.N. (2003).
  On parametric empirical Bayes methods for comparing multiple groups
  using replicated gene expression profiles.
  Statistics in Medicine 22:3899-3914.

  Newton, M.A. and Kendziorski, C.M.
  Parametric Empirical Bayes Methods for Microarrays in
  The analysis of gene expression data: methods and software. Eds.
  G. Parmigiani, E.S. Garrett, R. Irizarry and S.L. Zeger,
  New York: Springer Verlag, 2003.

  Newton, M.A., Noueiry, A., Sarkar, D., and Ahlquist, P. (2004).
  Detecting differential gene expression with a semiparametric
  hierarchical mixture model. Biostatistics 5: 155-176.

  Yuan, M. and Kendziorski, C. (2006). A unified approach for simultaneous
  gene clustering and differential expression identification.
  Biometrics 62(4): 1089-1098.

}
\seealso{
  \code{\link{ebPatterns}}, \code{\link{ebarraysFamily-class}}
}
\examples{
data(sample.ExpressionSet) ## from Biobase
eset <- exprs(sample.ExpressionSet)
patterns <- ebPatterns(c("1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1",
                         "1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2"))
gg.fit <- emfit(data = eset, family = "GG", hypotheses = patterns, verbose = TRUE)
show(gg.fit)
}
\keyword{models}