Prediction of potential miRNA-disease associations based on adaptive multi-view multi-label learning
Please refer to the Readme.pdf file for more details.
Method Description
AMVML is a computational framework that can effectively and reliably uncover potential disease-related miRNAs. It learns a new affinity graph adaptively for both diseases and miRNAs from multiple biological data sources. It also simultaneously update the miRNA-disease association predicted from both spaces based on multiple learning. In particular, the convergence of AMVML has been proved theoretically and the corresponding analysis indicates that it has a fast convergence rate. It is also worth mentioning that AMVML can be easily extended if there are more biological datasets available.
Requirements
AMVML was developed in MATLAB 2014b environment, but it should be working in all MATLAB versions.
Usage
We provided two functions, case study and global leave-one-out cross-validation(LOOCV), for users. To run the case study, please load the script 'caseStudy.m' into your MATLAB programming environment and click 'run'. To run global LOOCV, please load the script 'global_loocv.m' accordingly and other operations are the same as that of case study. Users can also run the scripts in standard command-line mode, where you should input the following commands for each function, respectively:
matlab -nodisplay -nodesktop -nosplash -r "global_loocv;exit;"
matlab -nodisplay -nodesktop -nosplash -r "caseStudy;exit;"
All the datasets used in the code, i.e. miRNA sequence similarity, miRNA functional similarity, miRNA semantic similarity, disease semantic similarity and miRNA-disease associations are all provided in the corresponding 'datasets/*.mat'.
Parameters
There are two parameters alpha and beta in AMVML. The default value for both parameters are 1e-4. Users can change their value in "MultiViewPrediction.m" file.
Output
The default output directory of AMVML is under the same directory where the scripts locate and it can be changed in the 'caseStudy.m' and 'global_loocv.m' file accordingly. All the results are stored in 'mat' file for convenience.
Contact
For any questions regarding our work, please feel free to contact us: alcs417@sdnu.edu.cn.