Multi-Metric and Multi-Substructure Biclustering Analysis for Gene Expression Data, in IEEE Computational Systems Bioinformatics Conference, 2005.

Authors: S. Y. Kung, M.W. Mak, and I. Tagkopoulos

The paper can be download here.

 

Machine Learning Approach to DNA Microarray Biclustering Analysis, in IEEE Workshop on Machine Learning for Signal Processing, 2005.

Authors: S. Y. Kung, and M.W. Mak

The paper can be download here.

 

Machine Learning for Multi-Modality Genomic Signal Processing, in IEEE Signal Processing Magazine, 2006.

Authors: S. Y. Kung and M.W. Mak

The paper can be download here.

 

Symmetric and Asymmetric Multi-Modality Biclustering Analysis for Microarry Data Matrix, J. of Bioinformatics and Computational Biology, 4(3), June 2006.

Authors: S. Y. Kung, M.W. Mak, and I. Tagkopoulos

The paper can be download here.

 

The Matlab program (Bicluster) for these papers can be download here.

Installation Procedures:

  1. Unzip bicluster.zip to your working directory, e.g. c:\work.

  2. Download Schwaighofer's SVM package from www.igi.tugraz.at/aschwaig. Put the package in c:\work\svm. If you do not have Matlab Optimization toolbox, you will need the PRLOQO solver.

  3. If you do not have Matlab statistic toolbox, you may need Bishop's Netlab for the K-means algorithm. Put the package in c:\work\netlab.

  4. cd to "data-files" directory.

  5. Add the path "..\m-files" to Matlab by typing "addpath ..\m-files".

  6. To produce the results in the paper in CSB'2005, run the script files "aveRocSearchMultiPreProc.m" and "aveRocSearchPreProc.m". These script files will call the respective bicluster functions to perform biclustering.

  7. To produce the results in the paper in MLSP'2005, run the script files "aveRocProjectMultiPreProc.m" and "aveDbnnRocProjectPreProc.m". These script files will call the respective bicluster functions to perform biclustering.

Note: If you do not want to install SVM and Netlab, you may still be able to use the Bicluster package by setting the variables para.classifierType = 'dbnn', para.nPosCtrs = 1, and para.nNegCtrs = 1 in the Matlab scripts (e.g. in aveRocSearchPreProc.m). However, without these package, only DBNN fusion with one center per class can be performed.

The expression matrix of yeast can be obtained from Cheung and Church Webpage.

For bug reports and enquiry, please contact Manwai MAK at enmwmak@polyu.edu.hk

M.W. Mak homepage

Last update: Friday, 02 June 2006