The method, NSRGRN, was created and tested using MATLAB® 2016b. Please refer to our manuscript termed “NSRGRN: A network structure refinement method for gene regulaotry network inference” for more detailed information about this algorithm.
main.m and main_sos.m are the running script of NSRGRN.
getLinkList.m gets the preliminary ranking list of gene regulations.
getSOSLinkList.m gets the preliminary ranking list of gene regulations.
get_link_list.m gets the final ranking list of gene regulations.
getDataSet.m is a function for getting the specific dataset in DREAM3 or DREAM4 challenges.
getGlodNet.m is a function for getting the corresponding standard network of the current dataset.
getPIR.m is a function for finding the potentially indirect regulation in the feedforward loop.
getBalanceConcentration.m calculates the balance concentration of all genes under a specific network.
edgePruning.m verifies the potentially indirect regulation using CMID.
cmi.m is a function for calculating the CMI. This code is derived from a paper termed "Inferring gene regulatory networks from gene expression data by path consistency algorithm based on conditional mutual information" by Zhang et al., 2012.
The folder "comparing_method " contains the prediction results of six comparing methods (PCA-CMI, NIMEFI, MMFGRN, PLSNET, GENIE3 and dynGENIE3). This folder also contains the code and prediction results of the NSR algorithm when acting as a post-processing step.
The folder "data10 " contains datasets of 10-gene networks in DREAM3.
The folder "data50 " contains datasets of 50-gene networks in DREAM3.
The folder "data100 " contains datasets of 100-gene networks in DREAM3.
The folder "gold" contains the standard networks of all datasets in DREAM3.
The folder "DREAM4_InSilico_Size10" contains datasets of 10-gene networks in DREAM4.
The folder "DREAM4_InSilico_Size100" contains datasets of 100-gene networks in DREAM4.
The folder "DREAM4_InSilicoNetworks_GoldStandard" contains the standard networks of all datasets in DREAM4.
The folder "detailed_information" contains the detailed information about DREAM3 and DREAM4 challenges.
You can find the entire dataset of DREAM3 in here and the entire dataset of DREAM4 in here.