Multiview Robust Graph-Based Clustering for Cancer Subtype Identification
Our method first learns robust latent representations from the raw omics data to alleviate the influences of the experimental and biological noise, where a set of similarity matrices are then adaptively learned based on these new representations. Finally, a global similarity graph is obtained by exploiting the consensus structure from the graphs of each view. As a result, the three parts in our method can reinforce each other in a mutual iterative manner.
MRGC was developed in MATLAB 2019b
We provided a demo for users. To run this demo, please load the script demo.m into your MATLAB programming environment and click run.
All the cancer datasets used can be downloaded at http://acgt.cs.tau.ac.il/multi_omic_benchmark/download.html.
There are three parameters in our method, i.e., alpha, beta, and the dictionary size base. The default value is 0.01, 0.001 and 10, repectively. Users can change their value in demo.m.
Users can change the input file directory and output file directory by changing the dataDir variable and the outDir variable in demo.m, respectively.
@ARTICLE{9685002,
author={Shi, Xiaofeng and Liang, Cheng and Wang, Hong},
journal={IEEE/ACM Transactions on Computational Biology and Bioinformatics},
title={Multiview Robust Graph-Based Clustering for Cancer Subtype Identification},
year={2023},
volume={20},
number={1},
pages={544-556},
doi={10.1109/TCBB.2022.3143897}
}