Template-class: Template: An S4 class representing a template of a group of FC Samples.

Description

An object of class "Template" summarizes a group of samples belonging to same biological-class with a class-template. A template is represented by a collection of meta-clusters (MetaCluster) created from samples of same class. An object of class "Template" therefore stores a list of MetaCluster objects and other necessary parameters.

Arguments

Creating Object

An object of class "Template" can be created using the function create.template : create.template(clustSamples, dist.type = "Mahalanobis", unmatch.penalty=999999, template.id = NA_integer_). The arguments to the create.template function is described below:

clustSamples: A list of ClusteredSample objects from which the template is created. The working examples describe how this objects are created by clustering FC samples.
dist.type: character, indicating the method with which the dissimilarity between a pair of clusters is computed. Supported dissimilarity measures are: 'Mahalanobis', 'KL' and 'Euclidean'. If this argument is not passed then 'Mahalanobis' distance is used by default.
unmatch.penalty: A numeric value denoting the penalty for leaving a cluster unmatched. This parameter should be already known or be estimated empirically estimated from data (see the reference for a discussion). Default is set to a very high value so that no cluster remains unmatched.
template.id: integer, denoting the index of the template (relative to other template). Default is NA_integer_

Slots

num.metaclusters:: The number of meta-clusters in the template.
metaClusters:: A list of length num.metaclusters storing the meta-clusters. Each meta-cluster is stored as an object of class MetaCluster.
dimension:: Dimensionality of the samples from which the template is created.
size:: Number of cells in the template (summation of all meta-cluster sizes).
tree:: A list (similar to an hclust object) storing the hierarchy of the samples in a template.
template.id:: integer, denoting the index of the template (relative to other templates). Default is NA_integer_

Accessors

get.size:: Number of cells in the template (summation of all meta-cluster sizes). Usage: get.size(object) here object is a Template object.
get.num.metaclusters:: Returns the number of meta-clusters in the template.
get.metaClusters:: Returns a list of length num.metaclusters storing the meta-clusters. Each meta-cluster is stored as an object of class MetaCluster.
get.dimension:: Returns the dimensionality of the samples from which the template is created.
get.tree:: Returns a hclust object storing the hierarchy of the samples in a template.
get.template.id:: Returns the index of the template (relative to other templates).

Methods

show

Display details about the Template object.

summary

Return descriptive summary for each MetaCluster of a Template. Usage: summary(Template)

plot

We plot a template as a collection of bivariate contour plots of its meta-clusters. To plot each meta-cluster we consider the clusters within the meta-cluster normally distributed and represent each cluster with an ellipsoid. The axes of an ellipsoid is estimated from the eigen values and eigen vectors of the covariance matrix of a cluster ("Applied Multivariate Statistical Analysis" by R. Johnson and D. Wichern, 5th edition, Prentice hall). We then plot the bivariate projection of the ellipsoid as 2-D ellipses. Usage: plot(template, alpha=.05, plot.mc=FALSE, color.mc=NULL, colorbysample=FALSE, ...) the arguments of the plot function are:

template: An object of class Template for which the plot function is invoked.
alpha: (1-alpha)*100% quantile of the distribution of the clusters or meta-cluster is plotted.
plot.mc: TRUE/FALSE, when TRUE the functions draws contour of the combined meta-cluster and when FALSE the function draws the contours of the individual clusters.
color.mc: A character vector of length num.metaclusters denoting the colors to be used to draw the contours. The ith color of this vector is used to draw the ellipses denoting clusters in the ith meta-cluster or the combined ith meta-cluster (depending on the argument plot.mc). By default an empty vector is passed and then an arbitrary color is used to draw each meta-cluster.
colorbysample: TRUE/FALSE, when TRUE the functions draws clusters from same samples in a single color and when FALSE the function draws meta-clusters in a single color.
... : Other usual plotting related parameters.

template.tree

Plot the hierarchy of samples established while creating the template-tree. See template.tree

References

Azad, Ariful and Pyne, Saumyadipta and Pothen, Alex (2012), Matching phosphorylation response patterns of antigen-receptor-stimulated T cells via flow cytometry; BMC Bioinformatics, 13 (Suppl 2), S10.

Examples

Run this code

## ------------------------------------------------
## load data
## ------------------------------------------------

library(healthyFlowData)
data(hd)

## ------------------------------------------------
## Retrieve each sample, clsuter it and store the
## clustered samples in a list
## ------------------------------------------------
set.seed(1234) # for reproducable clustering 
cat('Clustering samples: ')
clustSamples = list()
for(i in 1:length(hd.flowSet))
{
  cat(i, ' ')
  sample1 = exprs(hd.flowSet[[i]])
  clust1 = kmeans(sample1, centers=4, nstart=20)
  cluster.labels1 = clust1$cluster
  clustSample1 = ClusteredSample(labels=cluster.labels1, sample=sample1)
  clustSamples = c(clustSamples, clustSample1)
}

## ------------------------------------------------
## Create a template from the list of clustered samples and plot functions
## ------------------------------------------------

template = create.template(clustSamples)
summary(template)

## plot the tree denoting the hierarchy of the samples in a template
tree = template.tree(template)

## plot the template in terms of the meta-clusters
## option-1 (default): plot contours of each cluster of the meta-clusters
plot(template)

## option-2: plot contours of each cluster of the meta-clusters with defined color
plot(template, color.mc=c('blue','black','green3','red'))

## option-3: plot contours of the meta-clusters with defined color
plot(template, plot.mc=TRUE, color.mc=c('blue','black','green3','red'))


## option-4: plot contours of each cluster of the meta-clusters with different colors for different samples
plot(template, colorbysample=TRUE)

Run the code above in your browser using DataLab

Last chance! 50% off unlimited learning