ReorderToReference: Order the outputs of the clustering methods against a...
In IntClust: Integration of Multiple Data Sets with Clustering Techniques

Description Usage Arguments Details Value Note Examples

When multiple methods are performed on a data set, it is interesting to compare their results. However, a comparison is not easily done since different methods leads to a different ordering of the objects. The ReorderToReference rearranges the cluster to a reference method.

1 2	ReorderToReference(List, nrclusters = NULL, fusionsLog = FALSE, weightclust = FALSE, names = NULL)

`List`	A list of clustering outputs to be compared. The first element of the list will be used as the reference.
`nrclusters`	The number of clusters to cut the dendrogram in. Default is NULL.
`fusionsLog`	Logical. Indicator for the fusion of clusters. Default is FALSE.
`weightclust`	Logical. To be handed to `ReorderToReference`: to be used for the outputs of CEC, WeightedClust or WeightedSimClust. If TRUE, only the result of the Clust element is considered. Default is TRUE.
`names`	Optional. A character vector with the names of the methods.

It is interesting to compare the results of the methods described in the methodology. All methods result in a dendrogram which is cut into a specific number of clusters with the cutree function. This results in an numbering of cluster based on the ordering of the names in the data and not on the order in which they are grouped into clusters. However, different methods lead to different clusters and it is possible that cluster $1$ of one method will not be the cluster that has the most in common with cluster 1 of another method. This makes comparisons rather difficult. Therefore the ReorderToReference function was written which takes one method as a reference and rearranges the cluster numbers of the other methods to this reference such that clusters are appointed to that cluster they have the most in common with. The result of this function is a matrix of which the columns are in the order of the clustering of the objects of the referenced method and the rows represent the methods. Each cell contains the number of the cluster the compound is in for that method compared to the method used as a reference. This function is applied in the functions SimilarityMeasure, DiffGenes, Pathways and ComparePlot. It is a possibility that 2 or more clusters are fused together compared to the reference method. If this is true, the function will alert the user and will ask to put the parameter fusionsLog to true. Since ReorderToReference is often used as an internal function, also for visualization, it will print out how many more colors should be specified for those clusters that did not find a suitable match. This can be due to fusion or complete segregation of its objects into other clusters.

A matrix of which the cells indicate to what cluster the objects belong to according to the rearranged methods.

The ReorderToReference function was optimized for the situations presented by the data sets at hand. It is noted that the function might fail in a particular situation which results in a infinite loop.

data(fingerprintMat)
data(targetMat)

MCF7_F = Cluster(fingerprintMat,type="data",distmeasure="tanimoto",normalize=FALSE,
method=NULL,clust="agnes",linkage="flexible",gap=FALSE,maxK=55,StopRange=FALSE)
MCF7_T = Cluster(targetMat,type="data",distmeasure="tanimoto",normalize=FALSE,
method=NULL,clust="agnes",linkage="flexible",gap=FALSE,maxK=55,StopRange=FALSE)
MCF7_ADC=ADC(list(fingerprintMat,targetMat),distmeasure="tanimoto",normalize=FALSE,
method=NULL,clust="agnes",linkage="flexible")

L=list(MCF7_F,MCF7_ADC,MCF7_T)
names=c("FP","ADC","TP")

MCF7_Matrix=ReorderToReference(List=L,nrclusters = 7, fusionsLog = FALSE, weightclust = 
FALSE, names = names)