Instructions for finding domains that are linked together by successive domain fusion 
events by clustering an adjacency matrix of proteins versus domains.


1. Create an adjacency matrix of proteins versus domains where proteins constitute
   the rows of the matrix and domains constitute the columns. If a protein contains 
   a domain then matrix entry is one and zero otherwise. The protein versus domain 
   matrix used in the study is the file Protein_Domain_Saccharomyces_cerevisiae.dat.

2. Execute the following commands in Matlab version 13 or higher with the statistics 
   toolbox installed.

   a. svdAnalysis 
   b. SVDclust
    
   Note that other *.m files contain subroutines used by these matlab scripts.

3. The PERL script ConvertIndices.pl can be used extract the groups of domains and
   the correponding groups of proteins which contributed the domains to each group.
   The script requires the output from the clustering, files containing the lists 
   of proteins and domains used in the study. The lists should in the order as they
   appear in the adjacency matrix. For the current study the lists are for the list 
   of ORF names (ProteinList) and the list of PFam designations (DomainList).



 

