ROBUST & CONSENSUS FUNCTIONS IN R
Allan Tucker, IDA Group, Brunel University

GeneImport.R
Remarks:	This function reads in data in the Eisenlab format
Input:		Filename , number of variables (n), number of experiments (l)
Output:	 	A list containing: GeneIDs, Expts, a matrix containing the expression data

Agreemat.R 
Remarks:	This function generates an n*n agreement matrix from the results of k clustering 
		methods
Input: 		k*n matrix representing the results of k clustering algorithms 
		Each column in the matrix represents a set of clusters in vector format
Output:		An n*n matrix representing the agreement between cluster methods for each variable
		pair

Robust.R
Remarks:	This function generates a list of lists which represent the robust clusters from an 
		agreement matrix
Input:		Vector of geneIDs, An agreement matrix, the number of clustering methods used to generate the 
		agreement matrix
Output:		A list of lists representing the robust clusters

Consensus.R
Remarks: 	This makes use of ConsensusFn.R which scores a set of clusters given an agreement
	 	matrix
Input:		An agreement matrix, the maximum number of consensus clusters, starting temperature
		the number of iterations
Output:		A vector representing the consensus clusters

GenerateRCs.R
Remarks: 	This is a script that uses R implementations of PAM, Average Link Hierarchical
		and K-Means to generate robust clusters which it returns
Input:		Filename, n, l, number of clusters for PAM and kmeans and hierarchical
Output:		A vector representing the robust clusters

GenerateCCs.R
Remarks: 	This is a script that uses R implementations of PAM, Average Link Hierarchical
		and K-Means to generate consensus clusters which it returns
Input:		Filename, n, l, number of clusters for PAM and kmeans and hierarchical
Output:		A vector representing the consensus clusters

ConsensusFn.R
Remarks:	This Function takes a possible set of consensus clusters and scores them according
		to the agreement matrix, it requires a parameter alpha which represents the
		'Minimum Agreement' usually set to maximum agreement - minimum agreement in the
		agreement matrix.
Input:		Possible set of clusters in vector format, the agreement matrix, alpha (min agreement)
Output:		Score for the input Consensus Clusters

WKappa.R
Remarks:	Uses the Weighted Kappa metric to compare the similarity of two sets of clusters
Input:	Two sets of clusters in vector form
Output:	The Weighted Kappa of the two sets of clusters input