\name{hmmBeadStudioSet}
\alias{hmmBeadStudioSet}
\alias{hmmOligoSnpSet}
%- Also NEED an '\alias' for EACH other topic documented here.
\title{
   HMM functions for oligoSnpSet and BeadStudioSet containers
}
\description{

   HMM functions for \code{oligoSnpSet} and \code{BeadStudioSet}
   containers. These functions are exported in the package's namespace
   to provide documentation of arguments that can be passed from the
   \code{hmm} method for these containers.  The \code{hmmBeadStudioSet}
   function is always called when the object passed to the \code{hmm}
   method is a \code{BeadStudioSet}.  By contrast, the \code{hmm} method
   for \code{oligoSnpSet} objects will only call the
   \code{hmmOligoSnpSet} function if B allele frequences (assay data
   element "baf") is not included in the list of assay data elements.
   Specifically, if assay data element "baf" is in the list of assay
   data elements of a \code{oligoSnpSet} container, the \code{hmm}
   method for the \code{oligoSnpSet} class calls the
   \code{hmmBeadStudioSet} function.

}
\usage{
hmmBeadStudioSet(object, cnStates = logCnStates(), normalIndex = 3L,rohIndex = normalIndex + 1L, prOutlierCN = 0.01, prOutlierBAF = 0.001,p.hom = 0.05, TAUP = 1e+08, is.log, initialProb =rep(1/length(cnStates), length(cnStates)), center = TRUE, reestimation =TRUE, nupdates = 10, tolerance = 1, ...)
hmmOligoSnpSet(object, cnStates = c(0, 1, 2, 2, 3, 4), normalIndex = 3L, rohIndex = normalIndex + 1L, prOutlierCN = 0.01, prOutlierBAF = 0.001, p.hom = 0.05, TAUP = 1e+08, is.log, initialProb = rep(1/length(cnStates), length(cnStates)), center = TRUE, reestimation = TRUE, nupdates = 10, tolerance = 1, ...)
}
%- maybe also 'usage' for other objects documented here.
\arguments{
  \item{object}{
    A \code{oligoSnpSet} or \code{BeadStudioSet}.
}
\item{cnStates}{

  A vector of starting values (numeric) specifying the means of the
  Normal distribution assumed for latent copy numbers. The means must be
  specified for states homozygous deletion (zero copies), hemizygous
  deletion (1 copy), normal (2 copies), normal and no heterozygotes (2
  copies), single copy duplication (3 copies), and two+ copy duplication
  (4+ copies). The starting values are updated via EM.

}

\item{normalIndex}{

  Integer indicating the state index for diploid copy number. This
    should nearly always be '3' if the 6-state HMM (see \code{cnStates})
    is fit as recommended.

  }

  \item{rohIndex}{

    The state index (integer) for diploid copy number without
    heterozygotes. This should be '4'.  It is important to include this
    state even if homozygous regions with diploid copy number are not of
    interest.

}

\item{prOutlierCN}{

    The probability that a copy number estimate is an outlier.  This is
    an initial estimate that is updated for each copy number state via
    EM.

}

  \item{prOutlierBAF}{

        The probability that a B allele frequency is an outlier.  This
    is an initial estimate that is updated for each copy number state
    via EM.

  }

  \item{p.hom}{


  }


  \item{TAUP}{

    Scalar for the transition probability matrix.  Larger values
    discourage transitions from the normal state. (The transition
    probabilities are a function of the distance between adjacent
    markers.  These probabilities are not updated as part of the EM
    step.)

  }

  \item{is.log}{

    A \code{logical} indicating whether the copy number estimates are on
    the log scale.  Note that the assay data elements in
    \code{oligoSnpSet} and \code{BeadStudioSet} should be represented as
    integers (copy number or relative copy number * 100). If
    \code{is.log} is TRUE, we assume that after division by 100 the
    assay data element containing the copy numbers (or relative copy
    numbers) is on the log-scale.  The scale has implications on what is
    considered to be extreme.

  }

  \item{initialProb}{

    Vector of initial state probabilities.  This is required to be the
    same length as \code{cnStates}.

  }

  \item{center}{

    Whether to center the copy number for each chromosomal arm at the
    theoretical mean for the diploid copy number state.  This may not be
    appropriate for some datasets (e.g., trisomy 21, cancer
    applications).  A safer approach is to set this argument to
    \code{FALSE} and center all autosomes at the theoretical mean for
    two copies prior to fitting the HMM.

  }

  \item{reestimation}{ Logical, but currently ignored.  }

  \item{nupdates}{

    The maximum number of reestimation steps for updating the mean,
    variance, and outlier probabilities of the Gaussian-Uniform mixture
    for each copy number state.

  }

  \item{tolerance}{

    If the difference in the log likelihood between successive EM
    updates is less than tolerance, the number of updates can be less
    than \code{nupdates}.

  }

  \item{\dots}{
    Presently ignored
  }
}
\value{
  A \code{RangedData}-derived object.
}

\author{
R. Scharpf
}

\seealso{
  \code{\link{hmmSnpSet}}
}
\keyword{smooth}