VIMIDA

VIMIDA - is a Java-applet for VIsualization of MIcroarray DAta using linear and non-linear principal components analysis (PCA)
keywords: visualization; microarrays; principal components analysis; elastic maps; visual clustering

Description

VIMIDA allows two-dimensional visualization of a data file with vectorial data (tab- or space- delimited table with a simple header, look at a sample). The basic features of the applet are the following:

Linear Principal Components Analysis
Specifying classes of datapoints with different colors
Non-linear principal manifold analysis (elmap algorithm)
Point density estimation and visualization (visual inspection of clusters)
Visualization of class densities, relative densities
Visualization of functions defined in the data space (a field value, an average of several field values)
Possibility to automatically transpose the data table and visualize it
Feedback to other programs via HTTP GET interface
Possibility to deal with a remote data file

For further details go to the TUTORIAL

Examples

You can test VIMIDA immediately with a sample data file

Screenshot of visualization of CGH array data

Refer for other examples to the TUTORIAL

References

Other software implementing elastic maps algorithm:

ViDaExpert - Windows application for data visualization, with hundreds of features including interactive construction of elastic maps
elmap - C++ implementation of elastic maps algorithm
Visual DAO - Java package for data analysis

Publications:

Gorban A., Zinovyev A. Elastic Principal Graphs and Manifolds and their Practical Applications. Computing, 2005 (PDF)
Gorban A, Zinovyev A. and Wunsch D. Application of the method of elastic maps in analysis of genetic texts. Proc. International Joint Conference on Neural Networks (IJCNN), 2003 (PDF)
Gorban A, Zinovyev A. Visualization of Data by Method of Elastic Maps and Its Applications in Genomics, Economics and Sociology. IHES Preprint, 2001 (PDF)

Web-pages:

Implementation

The applet is a part of VDAOEngine package which is implemented using the core java classes. For solving systems of algebraic linear equations with big sparse matrices in the algorithm of principal manifold construction, COLT library is used (it is included in the applet jar file). The applet is signed by Andrei Zinvoyev and Bioinformatics Service of Institut Curie. This is done to allow the applet to have access to the local filesystem.

Downloads

Contacts

Bioinformatics service of Instiute of Curie
Andrei Zinovyev