%\VignetteIndexEntry{2. XPS Vignette: Classes}
%\VignetteDepends{xps}
%\VignetteKeywords{Expression, Affymetrix, Oligonucleotide Arrays}
%\VignettePackage{xps}
\documentclass{article}

\usepackage{hyperref}
\usepackage[authoryear,round]{natbib}

\textwidth=6.2in
\textheight=8.5in
%\parskip=.3cm
\oddsidemargin=.1in
\evensidemargin=.1in
\headheight=-.3in

\newcommand{\Rfunction}[1]{{\texttt{#1}}}
\newcommand{\Rmethod}[1]{{\texttt{#1}}}
\newcommand{\Rcode}[1]{{\texttt{#1}}}
\newcommand{\Robject}[1]{{\texttt{#1}}}
\newcommand{\Rpackage}[1]{{\textsf{#1}}}
\newcommand{\Rclass}[1]{{\textit{#1}}}
\newcommand{\Cclass}[1]{{\textit{#1}}}
\newcommand{\Rexten}[1]{{\textit{#1}}}
\newcommand{\xps}{\Rpackage{xps}}
\newcommand{\ROOT}{\Robject{ROOT}}

\begin{document}

\bibliographystyle{plainnat}

\title{Introduction to the xps Package: Classes structure}
\date{October, 2009}
\author{Christian Stratowa}
\maketitle

\tableofcontents

\section{Introduction}

 This document provides a tutorial on the class structures used in the \xps\ package. 
 The \xps\ package provides S4 class definitions and associated methods for raw
 intensity data, preprocessed intensity data, expression measures, and detection call data
 for batches of Affymetrix expression or exon arrays. Detailed information on functions,
 classes and methods can be obtained in the help files.

 Since the main purpose of the S4 classes is to provide a link to and allow access to
 \ROOT\ files and the \ROOT\ trees contained therein, it is important to understand the
 meaning of \ROOT\ files and the \ROOT\ trees first.


\section{XPS and ROOT}

 Package \xps\ is based on two powerful frameworks, namely \Robject{R} and \ROOT. 
 \ROOT (\url{http://root.cern.ch}) is an object-oriented C++ framework that has been 
 developed at CERN for distributed data warehousing and data mining of particle 
 data in the petabyte range. Data are stored as sets of objects in machine-independent 
 files, and a specialized class, called a \ROOT\ tree, has been designed to store large
 quantities of data, see The ROOT \cite{root:2009}. 


\subsection{ROOT File}

 A ROOT file, C++ class \Cclass{TFile}, is like a UNIX file directory. It can contain
 directories and objects organized in unlimited number of levels. It is also stored in
 machine independent format. Thus it can be considered to be the \ROOT\ equivalent of an 
 \Robject{R} \Robject{environment}. Just like an \Robject{R} \Robject{environment}
 a \ROOT\ file can be copied to any system directory and used from within any OS.

 A \ROOT\ file can store any C++ object like \ROOT\ trees, but also any other object such
 as a text file. For our purpose it is only important to know that \ROOT\ trees can be
 stored in \ROOT\ files.


\subsection{ROOT Tree}

 \ROOT\ trees, C++ class \Cclass{TTree}, have been especially designed to store large 
 quantities of same-class objects. As simplest case, a \ROOT\ tree can store data imported
 from tables, and thus can be considered to be the \ROOT\ equivalent of an \Robject{R}
 \Robject{data.frame}. A \ROOT\ tree consists of branches and leaves, with each leaf
 containing the data from one column of the imported data table.

 The \Cclass{TTree} class is optimized to reduce disk space and enhance access speed. When
 using a TTree, you fill its branch buffers with leaf data and the buffers are written to
 disk when it is full. Branches, buffers, and leaves are explained in: The ROOT \cite{root:2009};
 here it is important to realize that each object is not written individually, but rather
 collected and written a bunch at a time. This is where \Cclass{TTree} takes advantage of
 compression and will produce a much smaller file than if the objects were written individually.


\section{XPS S4 classes}

 Taking advantage of the features described above, package \xps\ stores all data as \ROOT\
 trees in portable \ROOT\ files. The four main classes for Affymetrix oligonucleotide
 array data are:

\begin{description}
\item
{{\tt SchemeTreeSet}:} Data describing microarray layout, probe information and metadata
 for genes are stored as \ROOT\ trees in \ROOT\ \Rclass{scheme} files, accessible
 from \Robject{R} as S4 class \Rclass{SchemeTreeSet} objects. 

\item
{{\tt DataTreeSet}:} This class provides the link to \ROOT\ data files and the \ROOT\
 trees contained therein. Data files contain the content of the raw CEL-files as
 \ROOT\ trees. Additional data files can contain the background-corrected intensities
 as well as the computed background intensities.

\item
{{\tt ExprTreeSet}:} Preprocessing of raw data, i.e. summarization and normalization, results
 in conversion of probe level data to expression values. These data are stored as \ROOT\ 
 trees in \ROOT\ expression files. Class \Rclass{ExprTreeSet} provides the link to the \ROOT\
 expression file and the \ROOT\ trees contained therein. 

\item
{{\tt CallTreeSet}:} Results of MAS5 detection calls or DABG calls, respectively, and the
 corresponding p-values are stored as \ROOT\ trees in \ROOT\ call files, accessible via
 class \Rclass{CallTreeSet}.
\end{description}

 All of these S4 classes are derived from the same virtual base class \Rclass{TreeSet}. \\


 The following classes are relevant for filtering and analysis of expression values:

\begin{description}
\item
{{\tt PreFilter}:} This class allows the initialization of different non--specific filters
 and stores the relevant parameters of each filter. 

\item
{{\tt UniFilter}:} This class allows the initialization of different univariate filters
 and stores the relevant parameters of each filter.. 

\item
{{\tt FilterTreeSet}:} This class provides a link between class \Rclass{ExprTreeSet}, the
 filter class applied to this \Rclass{ExprTreeSet} and the resulting \ROOT\ filter file and 
 the \ROOT\ mask tree contained therein. 

\item
{{\tt AnalysisTreeSet}:} Univariate analysis allows the selection of differentially
 expressed genes. This class provides a link between class \Rclass{FilterTreeSet}
 containing the relevant filtering information, and the resulting \ROOT\ filter file and 
 the \ROOT\ trees contained therein.
\end{description}

 The S4 classes \Rclass{PreFilter} and \Rclass{UniFilter} are derived from base class 
 \Rclass{Filter} while S4 classes \Rclass{FilterTreeSet} and \Rclass{AnalysisTreeSet} 
 are derived from the virtual base class \Rclass{TreeSet}. \\


 An optional S4 class, called \Rclass{ProjectInfo}, can be used to describe the relevant
 phenotypic and MIAME-like project information.


\subsection{{\tt TreeSet} class} 

 This is the virtual base class for all other S4 classes and provides the link to the
 \ROOT\ file and the \ROOT\ trees contained therein for the corresponding subclass.

 Function \Rfunction{getClassDef} can be used to get information on the slots and the
 inheritance information:

<<results=hide>>=
library(xps)
@

<<R>>=
getClassDef("TreeSet")
@

 The meaning of the slots is as follows:
\begin{description}
  \item{{\tt setname}:}{ Object of class "character", representing the name to the \ROOT\
   file subdirectory where the \ROOT\ trees are stored, usually one of \Cclass{DataTreeSet},
   \Cclass{PreprocesSet}, \Cclass{CallTreeSet}.}
  \item{{\tt settype}:}{ Object of class "character", describing the type of treeset
   stored in \Robject{setname}, usually one of \Robject{scheme}, \Robject{rawdata},
   \Robject{preprocess}.}
  \item{{\tt rootfile}:}{ Object of class "character", representing the name of the
   \ROOT\ file, including the full path.} 
  \item{{\tt filedir}:}{ Object of class "character", describing the full path to the
   system directory where \Robject{rootfile} is stored.}
  \item{{\tt numtrees}:}{ Object of class "numeric", representing the number of \ROOT\ 
   trees stored in subdirectoy \Robject{setname}.} 
  \item{{\tt treenames}:}{ Object of class "list", representing the names of the \ROOT\ 
   trees stored in subdirectoy \Robject{setname}.}
\end{description}


\subsection{{\tt SchemeTreeSet} class} 

 This class extends class \Rclass{TreeSet} and allows access to the \ROOT\ scheme trees,
 probe trees and annotation trees. Furthermore, it contains certain array information 
 (slot \Robject{probeinfo}) and allows access to the array mask (slot \Robject{mask}),
 describing the probe type. 

<<R>>=
getClassDef("SchemeTreeSet")
@

 In addition to the slots of \Rclass{TreeSet} it contains the following slots:
\begin{description}
  \item{{\tt chipname}:}{ Object of class "character", representing the Affymetrix
   chip name.}
  \item{{\tt chiptype}:}{ Object of class "character", representing the chip tpye, 
   either "GeneChip" or "ExonChip".}
  \item{{\tt probeinfo}:}{ Object of class "list", representing chip information, 
   including nrows, ncols, number of probes, etc.}
  \item{{\tt mask}:}{ Object of class "data.frame". The data.frame can contain the mask 
   used to identify the probes as e.g. PM, MM, or control probes.}
\end{description}

 As an example, \Robject{scheme.test3}, described in Vignette "xps.pdf" contains the
 information for the Affymetrix  `Test3' expression array:

<<>>=
scheme.test3 <- root.scheme(paste(.path.package("xps"),"schemes/SchemeTest3.root",sep="/"))
str(scheme.test3)
@ 

 As this example demonstrates, the \ROOT\ scheme file contains 4 \ROOT\ trees, storing
 all necessary GeneChip information, including the annotation. All trees can be exported
 as tables, and optionally also imported as \Robject{data.frame}.

 The `Unit' tree with extension \Rexten{idx} contains information about the processing
 units, i.e. the Affymetrix probesets, such as an internal `UNIT\_ID', probeset names and
 number of cells:

<<>>=
idx <- export(scheme.test3, treetype="idx", outfile="Test3_idx.txt", as.dataframe=TRUE, verbose=FALSE)
idx[10:16,]
@ 

 The `Scheme' tree with extension \Rexten{scm} provides the chip layout for the different
 probesets, i.e. the (x,y)-coordinates for each probeset with a defined `UNIT\_ID':

<<>>=
scm <- export(scheme.test3, treetype="scm", outfile="Test3_scm.txt", as.dataframe=TRUE, verbose=FALSE)
scm[1843:1848,]
@ 

 As shown, this tree contains also a column (leaf) `Mask', which assigns the probe type
 to each probe on the array. For expression arrays, probe types are PM with \Rcode{msk=1},
 MM with \Rcode{msk=0}, and control probes with \Rcode{msk=-1}. `Mask' for exon arrays
 will be described elsewhere.

 As default, slot \Rcode{mask} is empty to save memory (especially when using exon arrays).
 It is easy to add the `Mask' to slot \Rcode{mask} (and remove it, too):

<<>>=
scheme.test3 <- attachMask(scheme.test3)
msk <- chipMask(scheme.test3)
scheme.test3 <- removeMask(scheme.test3)
msk[1843:1848,]
@

 The `Probe' tree with extension \Rexten{prb} contains the nucleotide sequence for each
 PM oligonucleotide, as well as the GC-content and the calculated melting temperature Tm:

<<>>=
prb <- export(scheme.test3, treetype="prb", outfile="Test3_prb.txt",as.dataframe=TRUE, verbose=FALSE)
prb[1843:1848,]
@ 

 Finally, the transcript `Annotation' tree with extension \Rexten{ann} contains the
 probeset annotation, such as gene name and symbol:

<<>>=
ann <- export(scheme.test3, treetype="ann", outfile="Test3_ann.txt",as.dataframe=TRUE, verbose=FALSE)
head(ann)
@ 

 The \ROOT\ \Rclass{scheme} files for exon arrays contain additional tree types, see 
 the help file \Rcode{?validTreeType}.


\subsection{{\tt ProcesSet} class} 

 This is the common base class for subclasses \Rclass{DataTreeSet}, \Rclass{ExprTreeSet}, and
 \Rclass{CallTreeSet}, respectively, and contains the following slots:

<<R>>=
getClassDef("ProcesSet")
@

 The additional slots are:
\begin{description}
  \item{{\tt scheme}:}{ Object of class "SchemeTreeSet", providing access to the \ROOT\
   scheme file.}
  \item{{\tt data}:}{ Object of class "data.frame". The data.frame can contain 
   the data stored in \ROOT\ data trees.}
  \item{{\tt params}:}{ Object of class "list" representing relevant parameters.}
\end{description}

 The content of slot \Rcode{data} depends on the type of subclass:

 For subclass \Rclass{DataTreeSet} slot \Rcode{data} is reserved for the probe data,
 however, it is empty by default to avoid potential memory problems with large
 datasets. For subclasses \Rclass{ExprTreeSet} and \Rclass{CallTreeSet} slot \Rcode{data}
 contains expression levels and detection p-values, respectively. 


\subsection{{\tt DataTreeSet} class}
 
 Class \Rclass{DataTreeSet} is usually created when importing raw CEL-files as \ROOT\
 trees into a \ROOT\ data file, or accessing an existing \ROOT\ data file, using
 functions \Rcode{import.data} or \Rcode{root.data}, respectively. However, certain
 preprocessing functions such as \Rcode{bgcorrect} can also result in the creation
 of class \Rclass{DataTreeSet}. The slots are:

<<R>>=
getClassDef("DataTreeSet")
@

 Following additional slots are defined for class \Rclass{DataTreeSet}:
\begin{description}
  \item{{\tt bgtreenames}:}{ Object of class "list", representing the names of  
   optional \ROOT\ background trees.}
  \item{{\tt bgrd}:}{ Object of class "data.frame". The data.frame can contain 
   background intensities stored in \ROOT\ background trees.}
  \item{{\tt projectinfo}:}{ Object of class "ProjectInfo", containing information 
   about the project.}
\end{description}

 Slots \Rcode{bgtreenames} and \Rcode{bgrd} are reserved for optional background
 tree names and data, respectively.

 Slot \Rcode{projectinfo} allows to save optional project information defined in
 class \Rclass{ProjectInfo} described below. The project information will not only
 be stored in class \Rclass{DataTreeSet}, but also in the \ROOT\ data file containing
 the raw data from the CEL-files, when using functions \Rcode{import.data} or
 \Rcode{addData}, respectively. \\


 As an example, \Robject{data.test3}, described in Vignette "xps.pdf" contains the
 information to access the CEL-files from the `Test3' expression array, which are
 stored as \ROOT\ trees in \ROOT\ data file \Rclass{DataTest3\_cel.root}:

<<>>=
data.test3 <- root.data(scheme.test3, paste(.path.package("xps"),"rootdata/DataTest3_cel.root",sep="/"))
str(data.test3)
@ 

 As mentioned, slot \Rcode{data} is initially empty to avoid memory problems with
 large datasets, especially exon array datasets. Functions \Rcode{attachInten} and
 \Rcode{removeInten} allow to attach/remove these data, respectively, as described
 in Vignette "xps.pdf".


\subsection{{\tt ExprTreeSet} class} 

 Class \Rclass{ExprTreeSet} is obtained as result of preprocessing the raw data, i.e.
 summarization, or as result of normalizing preprocessed data, respectively. It
 contains the expression levels in slot \Rcode{data}. It has the following slots:

<<R>>=
getClassDef("ExprTreeSet")
@

 The additional slots are:
\begin{description}
  \item{{\tt exprtype}:}{ Object of class "character", representing the exression type, 
   i.e. \Rcode{rma}, \Rcode{mas5}, \Rcode{mas4} or \Rcode{custom}.}
  \item{{\tt normtype}:}{ Object of class "character", representing the normalization 
   type, i.e. \Rcode{none}, \Rcode{mean}, \Rcode{median}, \Rcode{lowess},\Rcode{supsmu}.}
\end{description}


\subsection{{\tt CallTreeSet} class} 

 Class \Rclass{CallTreeSet} allows to access the Affymetrix detection calls and corresponding
 p-values, either the MAS5 detection calls or the detection above background (DABG) calls.
 It is obtained as result of functions \Rcode{mas5.call} or  \Rcode{dabg.call}, respectively.
 It has the following slots:

<<R>>=
getClassDef("CallTreeSet")
@

 The additional slots are:
\begin{description}
  \item{{\tt calltype}:}{ Object of class "character", representing the call type, 
     i.e. "mas5" or "dabg".}
  \item{{\tt detcall}:}{ Object of class "data.frame". The data.frame can contain 
     the detection calls stored in \ROOT\ call trees.}
\end{description}

 Slot \Rcode{data} contains the detection call p-values while slot \Rcode{detcall}
 contains the detection calls.


\subsection{{\tt ProjectInfo} class} 

 This class is a stand-alone supporting S4 class and allows to store phenotypic and
 MIAME-like project information. It contains the following slots:

<<R>>=
getClassDef("ProjectInfo")
@

\begin{description}
  \item{{\tt submitter}:}{ Object of class "character", representing the name of the
   submitter.}
  \item{{\tt laboratory}:}{ Object of class "character", representing the laboratory 
   of the submitter.}
  \item{{\tt contact}:}{ Object of class "character", representing the contact address
   of the submitter.}
  \item{{\tt project}:}{ Object of class "list", representing the project information.}
  \item{{\tt author}:}{ Object of class "list", representing the author information.}
  \item{{\tt dataset}:}{ Object of class "list", representing the dataset information.}
  \item{{\tt source}:}{ Object of class "list", representing the sample source information.}
  \item{{\tt sample}:}{ Object of class "list", representing the sample information.}
  \item{{\tt celline}:}{ Object of class "list", representing the sample information for cell lines.}
  \item{{\tt primarycell}:}{ Object of class "list", representing the sample information for primary cells.}
  \item{{\tt tissue}:}{ Object of class "list", representing the sample information for tissues.}
  \item{{\tt biopsy}:}{ Object of class "list", representing the sample information for biopsies.}
  \item{{\tt arraytype}:}{ Object of class "list", representing the array information.}
  \item{{\tt hybridizations}:}{ Object of class "data.frame", representing the hybridization information for each hybridization.}
  \item{{\tt treatments}:}{ Object of class "data.frame", representing the treatment information for each hybridization.}
\end{description}

 This class must be created explicitely first using the constructor function \Rcode{ProjectInfo},
 before it can be added to class \Rclass{DataTreeSet}. Here is a simple example for
 \Robject{data.test3}:

<<>>=
project <- ProjectInfo(submitter="Christian", laboratory="home",contact="email")
projectInfo(project)    <- c("TestProject","20060106","Project Type","use Test3 data for testing","my comment")
authorInfo(project)     <- c("Stratowa","Christian","Project Leader","Company","Dept","cstrato.at.aon.at","++43-1-1234","my comment")
datasetInfo(project)    <- c("Test3Set","MC","Tissue","Stratowa","20060106","description","my comment")
sourceInfo(project)     <- c("Unknown","source type","Homo sapiens","caucasian","description","my comment")
primcellInfo(project)   <- c("Mel31","primary cell",20071123,"extracted from patient","male","my pheno","my genotype","RNA extraction",TRUE,"NMRI","female",7.0,"months", "my comment")
arrayInfo(project)      <- c("Test3","GeneChip","description","my comment")
hybridizInfo(project)   <- c(c("TestA1","hyb type","TestA1.CEL",20071117,"my prep1","standard protocol","A1",1,"my comment"),
                             c("TestA2","hyb type","TestA2.CEL",20071117,"my prep2","standard protocol","A2",1,"my comment"),
                             c("TestB1","hyb type","TestB1.CEL",20071117,"my prep1","standard protocol","B1",2,"my comment"),
                             c("TestB2","hyb type","TestB2.CEL",20071117,"my prep2","standard protocol","B2",2,"my comment"))
treatmentInfo(project)  <- c(c("TestA1","DMSO",4.3,"mM",1.0,"hours","intravenous","my comment"),
                             c("TestA2","DMSO",4.3,"mM",8.0,"hours","intravenous","my comment"),
                             c("TestB1","DrugA2",4.3,"mM",1.0,"hours","intravenous","my comment"),
                             c("TestB2","DrugA2",4.3,"mM",8.0,"hours","intravenous","my comment"))
show(project)
@ 


\subsection{{\tt Filter} class} 

 This is the base class for the S4 classes \Rclass{PreFilter} and \Rclass{UniFilter}.

<<R>>=
getClassDef("Filter")
@

 The meaning of the slots is as follows:
\begin{description}
  \item{{\tt numfilters}:}{ Object of class "numeric" giving the number of filters applied.}
\end{description}


\subsection{{\tt PreFilter} class} 

 Class \Rclass{PreFilter} allows to apply different filters to class \Rcode{ExprTreeSet},
 i.e. to the expression level data.frame \Rcode{data}..

<<R>>=
getClassDef("PreFilter")
@

 The meaning of the slots is as follows:
\begin{description}
  \item{{\tt mad}:}{Object of class "list" describing parameters for \Rcode{madFilter}.}
  \item{{\tt cv}:}{Object of class "list" describing parameters for \Rcode{cvFilter}.}
  \item{{\tt variance}:}{Object of class "list" describing parameters for \Rcode{varFilter}.}
  \item{{\tt difference}:}{Object of class "list" describing parameters for \Rcode{diffFilter}.}
  \item{{\tt ratio}:}{Object of class "list" describing parameters for \Rcode{RratioFilter}.}
  \item{{\tt gap}:}{Object of class "list" describing parameters for \Rcode{RgapFilter}.}
  \item{{\tt hithreshold}:}{Object of class "list" describing parameters for \Rcode{highFilter}.}
  \item{{\tt lothreshold}:}{Object of class "list" describing parameters for \Rcode{lowFilter}.}
  \item{{\tt quantile}:}{Object of class "list" describing parameters for \Rcode{quantileFilter}.}
  \item{{\tt prescall}:}{Object of class "list" describing parameters for \Rcode{callFilter}.}
\end{description}

 This class must be created explicitely using the constructor function \Rcode{PreFilter},
 and is added as parameter \Rcode{filter} to function \Rfunction{prefilter}. Here is an example,
 where all filters are initialized for demonstration purposes only:

<<>>=
prefltr <- PreFilter()
madFilter(prefltr) <- c(0.5,0.01)
cvFilter(prefltr) <- c(0.3,0.0,0.01)
varFilter(prefltr) <- c(0.6,0.02,0.01)
diffFilter(prefltr) <- c(2.2,0.0,0.01)
ratioFilter(prefltr) <- c(1.5)
gapFilter(prefltr) <- c(0.3,0.05,0.0,0.01)
lowFilter(prefltr) <- c(4.0,3,"samples")
highFilter(prefltr) <- c(14.5,75.0,"percent")
quantileFilter(prefltr) <- c(3.0, 0.05, 0.95)
callFilter(prefltr) <- c(0.02,80.0,"percent")
str(prefltr)
@ 


\subsection{{\tt UniFilter} class} 

 Class \Rclass{UniFilter} allows to apply different unitest filters to class \Rcode{ExprTreeSet},
 i.e. to the expression level data.frame \Rcode{data}.

<<R>>=
getClassDef("UniFilter")
@

 The meaning of the slots is as follows:
\begin{description}
  \item{{\tt foldchange}:}{Object of class "list" describing parameters for \Rcode{fcFilter}.}
  \item{{\tt prescall}:}{Object of class "list" describing parameters for \Rcode{callFilter}.}
  \item{{\tt unifilter}:}{Object of class "list" describing parameters for \Rcode{unitestFilter}.}
  \item{{\tt unitest}:}{Object of class "list" describing parameters for \Rcode{uniTest}.}
\end{description}

 This class must be created explicitely using the constructor function \Rcode{UniFilter},
 and is added as parameter \Rcode{filter} to function \Rfunction{unifilter}. Here is an example,
 where all filters are initialized for demonstration purposes only:

<<>>=
unifltr <- UniFilter()
fcFilter(unifltr) <- c(1.5,"both")
callFilter(unifltr) <- c(0.02,80.0,"percent")
unitestFilter(unifltr) <- c(0.01,"pval")
uniTest(unifltr) <- c("t.test","two.sided","wy",5000,0.0,FALSE,0.98,TRUE)
str(unifltr)
@ 


\subsection{{\tt FilterTreeSet} class} 

 Class \Rclass{FilterTreeSet} is obtained as result of filtering an instance of class
 \Rclass{ExprTreeSet}. It contains the filter mask in slot \Rcode{data}. 

<<R>>=
getClassDef("FilterTreeSet")
@

 In addition to the slots of classes \Rclass{ProcesSet} and \Rclass{TreeSet} it contains
 the following slots:
\begin{description}
  \item{{\tt filter}:}{Object of class \Rcode{"Filter"} currently providing access to 
     the \Rcode{PreFilter} settings.}
  \item{{\tt exprset}:}{Object of class \Rcode{"ExprTreeSet"} providing direct access to 
     the \Rcode{ExprTreeSet} used for filtering.}
  \item{{\tt callset}:}{Object of class \Rcode{"CallTreeSet"} providing direct access to 
     the optional \Rcode{CallTreeSet} used for filtering.}
\end{description}


\subsection{{\tt AnalysisTreeSet} class} 

 Class \Rclass{AnalysisTreeSet} is obtained as result of filtering an instance of class
 \Rclass{ExprTreeSet} using function \Rfunction{unifilter}. It contains currently the results
 of the univariate analysis in slot \Rcode{data}. 

<<R>>=
getClassDef("AnalysisTreeSet")
@

 In addition to the slots of classes \Rclass{ProcesSet} and \Rclass{TreeSet} it contains
 the following slots:
\begin{description}
  \item{{\tt fltrset}:}{Object of class \Rcode{"FilterTreeSet"} providing indirect access to 
     the \Rcode{ExprTreeSet} used and the \Rcode{UniFilter} settings.}
\end{description}


\bibliography{xps}

\end{document}