NCIS: A Network-Assisted Co-clustering Algorithm to Discover Cancer Subtypes based on Gene Expression

The user can download the "ncis.m" function file and put it in the folder where the datasets are. The user can then read the data using importdata() and call the function to get the results (following the steps in the instruction file).
The input files for both BRCA and GBM datasets can be downloaded from here (expression data downloaded from TCGA data portal.
For each dataset (BRCA or GBM), the user can first read "expressioninfo.txt" and "networkinfo.txt" into MATLAB (using function 'importdata()' would work).
Call function 'ncis()' following NCIS User's Guide to get the co-clustering results. The file "ncis.m" should be in the same folder and the working directory should be set there.
Results will be in the output files "sample_group.txt", "gene_group.txt", and "gene_weight.txt", which represent the information of each sample's group, each gene's group, and each gene's weight, respectively.
To get Figure 3 in the paper, the user can run code network.m. Gene expression file has name "expressioninfo.txt". Network file has name "networkinfo.txt". Sample group file has name "sample_group.txt". Gene weight file has name "gene_weight.txt".

Liu Y, Gu Q, Hou JP, Han J, and Ma J. A Network-assisted Co-clustering Algorithm to Discover Cancer Subtypes based on Gene Expression. BMC Bioinformatics, in press.