Category Cross-Omics>Pathway Analysis/Tools

Abstract The VANTED (Visualization and Analysis of NeTworks containing Experimental Data) software system for transcriptomics, proteomics and metabolomics analysis is a platform-independent system which enables researchers to evaluate extensive biochemical data in an easy way.

VANTED supports the integrated analysis of data for different growth conditions or transgenic lines from optionally different time points.

It uses the KEGG Pathway database, which includes a comprehensive set of 'pathway maps' representing knowledge about metabolism, and the Gene Ontology (GO), which contains a hierarchy of controlled vocabulary to describe gene and gene products in organisms.

Such networks and hierarchies, either imported, modified or newly created by the user, serve as the basis for different methods for mapping, visualization and analysis of data.

Not all imaginable use-cases may be supported by a software system, and therefore the possibility of enhancing an application with newly developed analysis, visualization and data exchange methods is of great importance for many researchers.

A Java and Ruby script-interface and interpreter allows the user to dynamically extend the VANTED system for such purposes.

VANTED Network Generation and Data Mapping --

For the analysis of experimental data it is sensible to apply an integrated view of the measured data and its related background information, such as ‘metabolic pathways’ or regulative processes.

This approach corresponds to the general idea of systems biology, where a biological system is analyzed Not only by studying a single phenomenon, but by considering a broader view which includes all elements of a biological system.

To achieve this, three (3) aspects were important for the design of the VANTED software system:

1) VANTED supports 'dynamic networks'.

Networks can be imported from databases (e.g. KEGG) or may be loaded from files in different formats - Graph Modeling Language (GML), Systems Biology Markup Language (SBML), Systems Biology Graphical Notation (SBGN), and Pajek-.NET [Pajek (Slovene word for Spider) is a program, for Windows, for the analysis and visualization of large networks].

It is also possible to create networks by hand with an integrated graphical editor. A big advantage of dynamic networks is the possibility of customizing them easily for different requirements. For instance, networks can be easily extended when more substances are measured.

2) The integration of measured data and relevant network elements is supported. An automatic mapping of data onto relevant network elements occurs if the measured data and the network nodes have common identifiers.

Also during this mapping procedure synonyms are used as long as they are included in one of the supported databases (e.g. the SIB Enzyme nomenclature database from the Swiss Institute of Bioinformatics).

If an automated integration is Not possible, a new ‘graph node’ is generated, which is then used for data mapping. Additionally, data may be assigned manually to user-given network elements.

(3) VANTED supports the display of 'multiple values' on a single network element.

While some approaches often support only the coloring of network elements based on single values (e.g. directly measured data or a computed factor, such as comparison of two different datasets), the inclusion of diagrams in the network representation allows the visualization of more complicated data.

An additional advantage which arises from the use of line charts or bar charts is the easy interpretation of such a representation.

VANTED Statistical Tests --

The measured data of a sample varies around a mean value because of measuring inaccuracies and biological variability. When a wild-type of a plant is compared to different other lines, or the plant is exposed to environmental stress, it is of interest whether the sample means differ significantly or Not.

For normally distributed data two variations of the t-test can be applied. Depending on the assumption of equality of variances, Student’s unpaired t-test or the Welch-Satterthwaite t-test can be carried out.

Whether a sample is normally distributed can be checked within VANTED with the built-in Davidquick test. The measurements which do Not fulfill this criterion are marked and can then be examined separately.

As an alternative to the t-test, the U-test is provided, which may also be used for Not normally distributed data. Another phenomenon is outliers in the dataset which can be identified in VANTED based on the Grubbs test.

VANTED Correlation Coefficients --

Relations between different measured substances can be recognized with XY diagrams. Here VANTED allows the selection of a number of 'substance nodes' from the network view. These substances are pair- wise related to each other and displayed in an array of XY diagrams.

The correlation factor between each two substances is computed and visualized using different colors of the diagram borders.

In a similar way, the correlation of a user-selected substance to all other substances can be determined. A positive or negative correlation between the selected substance and another substance becomes immediately visible by different node background colors.

Furthermore it is possible to determine statistically significant correlations between all measured substances which can be visualized by new graph edges connecting significantly correlated graph nodes.

VANTED Automated Data Clustering --

To recognize typical patterns in the temporal courses of the substance concentrations, VANTED includes a neural network (NN) algorithm, the Self-Organizing Map (SOM).

In the first phase of this algorithm a given number of typical profiles of substance concentrations over time are determined.

For instance, the substance concentration may increase in one group of substances during the time and decrease in another.

In the second step every measured substance is assigned to the best suitable pattern. Each substance afterwards belongs to a specific group.

In the 'graphical network' view the grouped data sets can then be separated and individually analyzed.

System Requirements

Contact manufacturer.


VANTED is mainly developed and designed by Christian Klukas, working in the Network Analysis Group of Dr. Falk Schreiber at the IPK- Gatersleben.

Manufacturer Web Site VANTED

Price Contact manufacturer.

G6G Abstract Number 20463

G6G Manufacturer Number 104090