A novel network regularized matrix decomposition method to detect mutated cancer genes in tumour samples with inter-patient heterogeneity
=================================================
Developer: Jianing Xi [email protected] from Health Informatics Lab, School of Information Science and Technology, University of Science and Technology of China
- 4GB memory
- MATLAB R2015a or later
The default If you want to use the default gene interaction network [iRefIndex 9] (http://irefindex.org), just ignore this step. Otherwise, if you want to use a user-defined network, replace the two files below with the files of the user-defined network:
./network/index_genes.txt./network/edge_list.txt
The first text file index_genes.txt is a table of the connection between the node id and the related gene names. The second text file edge_list.txt is the edges table that assigns the source nodes to the target nodes.
We provide an example data file of somatic mutations in breast invasive carcinoma (BRCA) samples from [UCSC Cancer Genomics Browser] (https://genome-cancer.soe.ucsc.edu/proj/site/hgHeatmap/) in ./data/example_data.zip. To analyze this data, please extract the .txt file somatic_data_BRCA.txt from ./example_data.zip and put the .txt file in the ./data folder. If you want to analyze the example data file with the default configurations, please run ./demo_mCGfinder.m and then the result file will be automatically saved in the directory ./output/.
If you want to analyze a user-specific data, a .txt file of the mutation binary table (samples x genes) of the sample names and the gene symbols must be provided as the format of the example data file. Put the txt file of the user-specific data in the ./data folder and run ./demo_mCGfinder.m.
The configurations of mCGfinder can be changed in script file ./demo_mCGfinder.m, and the descriptions of these parameters are provided below:
=================================================================================================
| PARAMETER NAME | DESCRIPTION |
=================================================================================================
|CompLeastProportion |Least sample proportion included in each components, which represents |
| |minimum proportion of the samples in every components given by the |
| |mCGfinder. The default proportion is set to 15%. |
-------------------------------------------------------------------------------------------------
|maxCompoent |Maximum number of components, which denotes the number of components |
| |given components given by mCGfinder at most. The default number is 5. |
-------------------------------------------------------------------------------------------------
|NetConf.lambda_T |The tuning parameter of network regularization, which is used to balance|
| |the fitness of the model (first term) and the smoothness of the scores |
| |of connected genes (second term). The default number is 0.1. |
-------------------------------------------------------------------------------------------------
The descriptions of output variables of mCGfinder are provided below:
=================================================================================================
| VARIABLE NAME | DESCRIPTION |
=================================================================================================
|detected_genes |Genes detected by mCGfinder as significantly mutated cancer genes. |
-------------------------------------------------------------------------------------------------
|S_sample_indicator |The sample indicator vectors of all component, which indicates the |
| |assignment of tumour samples to the every components. The i-th |
| |coefficient being 1 represents that the i-th samples are included in the|
| |component, and 0 otherwise. |
-------------------------------------------------------------------------------------------------
|Symbol_Net |The investigated gene list in the gene interaction network. |
-------------------------------------------------------------------------------------------------
|G_gene_score |The gene score vectors of all components, of which the coefficients are |
| |related to the gene lists variable 'Symbol_Net', and a higher value of |
| |a certain coefficient presents a larger potential of the gene to be |
| |cancer gene candidate. |
-------------------------------------------------------------------------------------------------
|Q_values |The q-values of all investigated genes in variable 'Symbol_Net', which |
| |are obtained by Benjamini-Hochberg false discovery rates control of the |
| |p-values of the investigated genes. |
-------------------------------------------------------------------------------------------------
For users without MATLAB licenses, we also offer mCGfinder standalone version for Windows ./mCGfinder_standalone.zip.
-
Step 1: MATLAB runtime installer: verify the MATLAB runtime is installed and ensure you have installed version 8.5 (R2015a). If the MATLAB runtime is not installed, please download the Windows 64-bit version of the MATLAB runtime for R2015a from the MathWorks Web site by navigating to
[http://www.mathworks.com/products/compiler/mcr/index.html] (http://www.mathworks.com/products/compiler/mcr/index.html)
Or run the installation file
./mCGfinder_standalone/MyAppInstaller_web.exeprovided in the .zip file. -
Step 2: Extract the
./mCGfinder_standalone.zipfile, and locate the executable file./mCGfinder_standalone/mCGfinder.exeand the two folders./mCGfinder_standalone/dataand./mCGfinder_standalone/networkin the same directory./mCGfinder_standalone/. Put the .txt files of mutation data in folder./mCGfinder_standalone/data/and network filesindex_genes.txtandedge_list.txtin folder./mCGfinder_standalone/network/. -
Step 3: Run
./mCGfinder_standalone/mCGfinder.exe. Please wait until the current program is finished, and the output variables are automatically saved as .txt files in folder./mCGfinder_standalone/output.
Please feel free to contact us if you need any help: [email protected].
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent