Soybean Functional Network

        A database of Soybean Functional Gene Network (SoyFGN) and miRNA Functional Network (SoymiRFN)

                         Functional Gene Network (FGN) | miRNA Network (miRFN)
                                          miRNA-Target datasets | 3'-UTR | Gene Annotations| KEGG Pathway
                                                                               Functional Similarity | miRNA Network | Statistics | Site map | Visit Log
                                                                                                                               Contact | Privacy policy



Pseudo-3D Clustering Algorithm



Citation: Yungang Xu*, Maozu Guo*, Xiaoyan Liu, Chunyu Wang, Yang Liu, Guojun Liu. Identify bilayer modules via Pseudo-3D Clustering Algorithm.( Under review )


Try Psedo-3D Clustering on your bilayer networks ? Clik HERE to download the latest runable JAR file, which can run on any operating system equipped with the JAVA. Please using above citaiton when you use the pseudo-3D clustering algorithm. Refer to below quick start guide as follows.

*Start program

      using following command:

             java   -jar   P3DClustering15.02.01.jar    -net1   netfile1    -net2   netfile2    -net3   netfile3   [-Xmx2g]   [-mail   your_email]

       -Using'-net1' specify the first network layer, '-net2' the second network layer, and '-net3' the inter-connections between two layers. All network files should including their paths.

       -Using '-Xmx' to maximize the heap space for a large bilayer network (optional).

       -Using '-mail' to specify a E-mail address to receive the results notification when your job was completed. NOTE: The notification mail may be put into your spam, please add to your whitelist.

       -All networks should be formated as a tab-separated plain text, of which each line corresponds to an edge. As follows or down load sample file.

                    node1 node2 edge-weight

                    node2 node3 edge-weight

       -For unwighed networks, the algorithm will set the edge weight as "1".

*Output. All results will be output into a newlly created folder at the same path with JAR file. The folder name is of the formate of "Run_date,time".Such as

the 'date and time' is the time when you start the program.

There are two folders and several files in the outputing folder. Including as follows (some of them could be different according to different bilayer network you input.)












File leagands:

[1] The coheisiveness matrix of each level. The modularity (Q value) is provide at the end of the each file.

[2] The resulting bilayer modules at each level. The bilayer modules of the same level are put into a single folder named as "L1", "L2" and so on.

[3] The merged net files of three inputing net files.

[4] The adjacency block matrix of the bilayer network.

Note: the following [5]~[11] files are not useful for users, all nodes are recoded as the column (or row) index in the ajacency matarix in file [4].

[5] The initial biclusters (modules) for miRNA?gene direction. 1st column is the module index starting from 1; 2nd column the number of miRNAs in each module; 3rd column the number of genes in each module; 4th column the miRNA names of each module; 5th column the gene names of each module. [6] is the corresponding initial bicliques for inital biclusters identification.

[7] and [8] are the same as [5] and [6] , but for the gene→miRNA direction.

[9]~[11] display the process of integrating module from two directions, and the tuning processing.

[12]~[19] are the resulting levels at which the bilayer network is partitioned. Each is a tab-seperated 6-column plain text with digital in file name as the modularity of this level. The number of these files depends on the number of levels that your input bilayer network will be partioned. 1st column is the row index; 2nd column the gene number of each module; 3rd column the miRNA number of each module; 4th the clustering coefficient of each module; 5th column the gene names of this module; 6th the miRNA name of this module.

[20] The most important resulting file. It provies the statistics of this runing. The max modularity value refers to the optimal level of module partition.

[21] A program runing log.




About Us | Site Map | Privacy Policy | Contact Us | ©2013 Natural Computation Lab.
Support major browsers, IE 7.0, Chrome 30 or higher are recommended.