Home

ViaComplex FAQ

Frequently Asked Questions about the ViaComplex application.


Last-modified: 02-Mar-2009
Version: 1.0
Contact: Mauro Castro (mauro@ufrgs.br) or José Rybarczyk (jose.rybarczyk@gmail.com)


Introduction

This FAQ is an attempt to answer some frequently asked questions about the ViaComplex application.

Contents

General Questions

What is ViaComplex?

Where can I get ViaComplex?

What is the current version of ViaComplex?

In which language is ViaComplex written?

Is the source code of ViaComplex available?

 

Technical Questions

Can ViaComplex handle large networks or there is a restriction in size?

What is the limit on the total number of genes that ViaComplex can analyze and display?

Can I use an adjacency matrix as input for the network?

In what form has the annotation of the genes been given?

Although the current version of the software is described as limited to human, is the tool able to run the same analysis for other species?

What is the acceptable format of gene expression data?

What is the novelty of ViaComplex comparing to available tools for high-throughput data analysis such as R packages or commercial tools such as Ingenuity Pathway Analysis (IPA)?

Are there future plans of releasing ViaComplex as a Matlab or R/Bioconductor package?

Are there numerical examples available covering the possibilities of the program?

Some considerations about network size (regard release 09_2009)!!

 

 


What is ViaComplex?

ViaComplex is open-source software that builds and analyzes landscape maps of genome maintenance mechanisms (GMM).

The software allows investigators to mine and connect their individual high-throughput data with GMM network-based model.

 

 

 

Top  Home


Where can I get ViaComplex?

Software and manuals are available from the ViaComplex Home Page

http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/

 

 

 

Top  Home


What is the current version of ViaComplex?

The current version of ViaComplex is 1.0 (September 2009)

 

 

 

Top  Home


In which language is ViaComplex written?

ViaComplex is written in FORTRAN and compiled in Intel FORTRAN Compiler 10 or higher.

 

 

 

Top  Home


Is the source code of ViaComplex available?

Yes. The source code is available at the ViaComplex Home Page.

It is distributed under GNU General Public License and is completely free for any use.

 

 

 

Top  Home


Can ViaComplex handle large networks or there is a restriction in size?

Although the software has been designed to deal with the human GMM model, ViaComplex algorithms can handle different network models, including large networks. This function is available at the “Custom Model” options where the user can input its own networks. Originally, we designed this module for testing purposes only to help readers to understand the flow of GMM model. However, it would be fantastic to see the application extended to other network models. We provided several numerical examples (available for download at http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/examples.htm) illustrating the performance of the software for different network sizes, including a network with 10000 nodes.

 

 

 

Top  Home


What is the limit on the total number of genes that ViaComplex can analyze and display?

We provide several examples (available for download at http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/examples.htm) illustrating the performance of the software for different network sizes.

 

 

 

Top  Home


Can I use an adjacency matrix as input for the network?

Unfortunately the software can not use this data format. Given that the software projects gene expression data onto network graphs, it requires node coordinates which are not available in the adjacency matrix.

 

 

 

Top  Home


In what form has the annotation of the genes been given?

Please, see the panel exemplifying the flow of the analysis together with acceptable formats:

http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/input_format.pdf

 

 

 

Top  Home


Although the current version of the software is described as limited to human, is the tool able to run the same analysis for other species?

Yes, it is possible. Although the software has been designed to dell with a specific network model, the algorithm can handle different networks from different species. Such use is possible because the “Custom Model” module can project different expression data types onto different gene networks to build functional maps. The only condition is that node_ID has to be compatible with gene_ID. We included two interactomes in ViaComplex home page illustrating this possibility (E. coli and S. cerevisiae gene networks: http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/examples.htm)

 

 

 

Top  Home


What is the acceptable format of gene expression data?

ViaComplex can read the common gene/protein identifiers (EMBL, ENTREZ, UniProt, HGNC, RefSeq, and UniGene) and when the data input contains inconsistent IDs or lacks any gene ID necessary to execute the landscape analysis then it is able to recognize and save such occurrence in a log file (file_rec.log). This procedure alerts user to the problem and indicates the specific point of his data that needs to be revised.

 

Please, see the panel exemplifying the flow of the analysis together with acceptable formats:

http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/input_format.pdf

 

 

 

Top  Home


What is the novelty of ViaComplex comparing to available tools for high-throughput data analysis such as R packages or commercial tools such as Ingenuity Pathway Analysis (IPA)?

We have compared ViaComplex with other related commercial and open source softwares. To facilitate this task, we present a comparative able where different tools are organized in five classes:

(i) Visualization of network topology and/or interaction maps: Here we list softwares able to visualize the network structure, whether in details or not.

(ii) Layout algorithms and/or network manipulation: Here we list softwares able to change the network structure, whether via algorithmic methods or via user interactive manipulation.

(iii) 3D representations: Here we list softwares able to show networks in both 2D and 3D representations. A pseudo-3D classification was added according to Pavlopoulos et al (BioData Mining, 1:12, 2008).

(iv) Node-based visualization of high-throughput data: Here we list softwares able to integrate high-throughput data to the network nodes (e.g. plot quantitative data in the network vertices only). This capability is mainly associated to the analysis large scale gene expression data.

(iv) Landscape visualization of high-throughput data: Here we considered an essential classification to contrast ViaComplex with the other softwares, that is, the ability to project high-throughput data onto the network. This capability differs from the node-based visualization, since it deals with the entire network topology, plotting quantitative data as a third axis of the network coordinates, which is able to create a gene expression landscape.

It is important to mention that all these softwares can provide many other functions, for example: “Build developmental gene-regulatory network models”, “Simulate and analyze network behavior”, “Visualize process diagrams”, “Conduct graph theoretical analysis”, “Integrate data from different sources”, etc. However, we focused in general aspects to give a good overview and with more emphasis in those aspects necessary to evaluate ViaComplex. If there are other related tools not listed, or if we have compared in less details than expected to give a good overview, please, contact us in order to review this list.

 

The comparative table is available at http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/softwares.html

 

 

 

Top  Home


Are there future plans of releasing ViaComplex as a Matlab or R/Bioconductor package?

We have concluded this first version of the software as a natural extension of our previous works published at NAR (PMID:17332015 and PMID:18832373). The conceptual design of the software certainly was an essential step to develop the code and the previous works contributed a lot in this direction. Now, the extension of this concept will demand a different strategy and R is an excellent approach to explore the algorithm. However, our next idea is to use ViaComplex to expand our knowledge about the systemic properties of the genome maintenance mechanisms. This is a two-way strategy, since it will further test the algorithm under different conditions. Next, we are planning a second version of ViaComplex incorporating any new concept, including a release as R package.

 

 

 

Top  Home


Are there numerical examples available covering the possibilities of the program and demonstrating its current capabilities?

We provide several numerical examples at http://lief.if.ufrgs.br/pub/biosoftwares/viacomplex/examples.htm). Also, the install package contains many other examples illustrating the landscape analysis of our human GMM model.

 

 

 

Top  Home


Some considerations about network size (regard release 09_2009).

 

Provided that ViaComplex is able to map certain quantities onto a given gene/protein network, it requires a visual representation of an area, comprising the features and relationships of the network elements. Like many types of maps, this strategy sacrifices a certain amount of accuracy to deliver a greater visual usefulness to the user. However, in order to avoid misinterpretations, it is important to recognize that maps - in general - are drawn to a scale. For instance, small scale maps are good to cover large regions, such as continents and the whole globe, and there is no sense trying to find some kinds of details in these types of maps (e.g. buildings, streets, people, etc.).

Likewise, ViaComplex is designed to deal with a range of networks. Although there is no exact dividing line between large and small maps, if you are trying to analyze small networks it would be better using a different strategy. You would be disappointed using ViaComplex to analyze the relationship among few genes, say 20 nodes. In these cases, there are other softwares that can be more helpful. The comparative table at ViaComplex Homepage may help you finding another tool.

Even so, we implemented some improvements in order to better scale the analysis according to the network size, as follows:

1)   Up to 50 nodes: 9 points between nodes

2)   51 to 500 nodes: 3 points between nodes

3)   More than 500 nodes: 1 point between nodes

...where each point corresponds to a new coordinate (onto an edge) used to distribute a given quantity (e.g. microarray signal),

weighed according to the node-pair distance.

 

Top  Home