Gene expression data is usually arranged in a matrix such that each row represents a gene
and each column corresponds to an experimental condition. The knowledge which genes
co-regulate with one another under a particular set of experimental conditions is demanded
because this provides information for scientists to determine the genes which participate
in the same biological processes. The classical clustering methods fail to extract these
relations because they are only able to either partition genes using the whole column data
or cluster columns using the whole row data. Recently, biclustering is investigated to
discover the underlined relations. The biclustering approaches can group a subset of
genes in a subset of experimental conditions with respect to some similarity scores. In
, we propose a new efficient algorithm based on parallel coordinate (PC) formulation for
biclustering. By using different similarity criteria, it is possible to detect biclusters of
the following types:
constant, constant rows, constant columns, additive and
An explanation of the proposed biclustering algorithm can be found in the
while some experimental results for yeast S. cerevisiae are given in the
BiVisu is a software tool which implements our proposed biclustering algorithm. It is also a
software for visualizing the detected biclusters in a 2D setting using PC plots. Currently,
it is available in a Matlab program.
The biclustering algorithm and visualization of BiVisu are both based on
parallel coordinate (PC) plots.
The BiVisu has an interactive graphical user interface (GUI) that users can analyze
and refine their biclustering results in a convenient way. Main features of BiVisu
are listed below,
Detection of biclusters of constant, constant rows, constant columns, additive and
- Preprocessing gene expression data using logarithm
Filtering detected biclusters according to specified requirements such as minimum
number of rows, minimum number of columns, maximum number of biclusters and maximum
PC plots of expression values
PC plots of difference and ratio matrices for biclusters of additive and
multiplicative types respectively
Display of coherence measures including mean square residue score (MSRS)  and
average correlation value (ACV)  together with other useful biclusters information
in the main panel
Other common functions available in biclustering software tools such as heat map
display and exporting biclustering results to text files
Click here for an overview. A step-by-step example can be
found through this link.
Matlab version of BiVisu is available here.
The artificial and real gene expression data used in  can be downloaded through the
- Artificial data for additive models
- Artificial data for multiplicative models
- Yeast Saccharomyces cerevisiae datasets :
- Arabidopsis thaliana datasets :
K.-O. Cheng, N.-F. Law, and W.-C. Siu and A. W.-C. Liew, Identification of Coherent
Patterns in Gene Expression Data Using an Efficient Biclustering Algorithm and
Parallel Coordinate Visualization, BMC Bioinformatics , vol. 9,
article 210, April 2008, doi: 10.1186/1471-2105-9-210.
Y. Cheng and G.M. Church, Biclustering of Expression Data, Proceedings,
Conference on Intelligent Systems for Molecular Biology , pp. 93 - 103, 2000.
L. Teng and L.-W. Chan, Biclustering Gene Expression Profiles by Alternately Sorting
with Weighted Correlated Coefficient, Proceedings of IEEE International
Workshop on Machine Learning for Signal Processing , pp. 289 - 294, 2006.
A. Prelic et al. , A Systematic Comparison and Evaluation of Biclustering
Methods for Gene Expression Data, Bioinformatics , vol. 22, no. 9,
pp. 1122-1129, 2006.
S.C. Madeira and A.L. Oliveira, Biclustering Algorithms for Biological Data
Analysis: A Survey, IEEE/ACM Trans. Computational Biology and Bioinformatics
, vol. 1, no. 1, pp. 24-45, 2004.
C.I. Castillo-Davis and D.L. Hartl, GeneMerge - post-genomic analysis, data mining,
and hypothesis testing, Bioinformatics , vol.19, no.7, pp.891-892, 2003.
Dr. Bonnie Law: email@example.com