buildSinglenode

It showed up that an attempt where the modalities are set as vertices in the graph and taking all of the available edges leads to the circumstance of a heavily extended run time. For this reason the [buildSinglenode()] function has been implemented. As first argument it takes the generated matrix from calcscorematrix(). The second parameter is a list which contains first the parameter "pernode" or "global". In the "pernode" mode the top n edges of each node are used, which can be determined by the second parameter in this list. The "global" mode performs also a reduction of edges, but here a global cutoff is used so that only the top n edges of the graph are used (here it is possible that some vertices are not used for the analyses). By selecting 0 in this mode the cutoff is set to the top (20 * number of vertices) edges.

singlenode_list <- buildSinglenode(matrix = matrix, cutoff = c("pernode", 20))

The generated list is saved to the working directory as adjMatrix_singlenode.csv and contains in the first two columns the gene names and in the third the strength of the edge between them.

This file can now be given to the graph analysis which is performed by Graph.py