Gene Network Construction... Web Based Tool

Question

[Dataminer]

Hi!

I have 10 genes linked to a particular disease (for the sake of example say cancer).

I want to build a gene network for these 10 genes.

Any web based tools available which can do the job?

Thank you

[David Quigley]

There are many different ways you could construct a network. Gene co-expression, protein-protein interaction, genetic interaction, co-appearance in literature etc. Your question would be improved if it were more specific about the kind of network you're looking to build.

[Dataminer]

Thank you, for your comment. I already got my answer from qdjm and I believe I have framed my question correctly. The GeneMania incorporates all the aspects before constructing gene network.

Answer 1

[Qdjm]

Try GeneMANIA. If you open up the advanced options, there's hundreds of networks to choose from. You can save the network from the file menu.

Then try STRING. This is the oldest and probably most widely used web tool for this purpose.

Other web-based network-construction tools for human include:

• FunCoup -- one of the few tools with regulatory interactions. Also has some nice display and clustering options
• FunctionalNet -- by one of the founders of the field of functional interaction networks.

It's worth trying all the tools because they use different network data sources and use a different algorithm to add nodes to your initial list to fill out the network.

The tools above combine network data from multiple sources (e.g. co-expression, protein interactions, sequence similarity, etc). There's a few other tools that focus only on one type of network data.

These include:

• BioGRID (protein and genetic interactions)
• IRefWeb (protein interactions)
• GEMMA (co-expression)

Edit:

Looks like you are know about Cytoscape plugins. If you are already using Cytoscape, then it's worth checking out the Reactome FI plugin. There's a GeneMANIA Cytoscape plugin as well.

[Treylathe]

just wanted to point out, if it's not out-of-line, GeneMania has a (publicly-available) tutorial (GeneMania funds tutorial, we provide): http://www.openhelix.com/genemania

[Dataminer]

WoW... Thank's a lot for such a detailed answer.

Answer 2

[Lyco]

You might want to check http://www.genemania.org/. It won't do the whole work for you, but it can be a great help.

[Lyco]

No problem. I like qdjm's answer better then mine, too :-)

[Dataminer]

Thank you for the answer, GeneMania is the only solution if one wants to construct a gene network. But qdjm gave a very descriptive answer, hence I had to accept his answer and not yours. Thank you again.

Answer 3

[Pierre Lindenbaum]

I found the following link by searching for "Online graphviz dot":

http://ashitani.jp/gv/

there's also: http://cytoscapeweb.cytoscape.org/

you can also create a 'fake' RDF file and use the W3C RDF validation service ( http://www.w3.org/RDF/Validator/ ) with the 'Graph' option

[Lyco]

Graphviz is only helpful for the visualization, if you already know the connections. The same is true for Cytoscape, but the latter has some plugins that assist with the network construction.

Answer 4

[Israel Barrantes]

Respect to Cytoscape plugins, I would recommend Cytoprophet, that reconstructs networks based on protein and domain interaction databases. You just have to load a list of SwissProt homologs, and then the plugin does the rest.

Answer 5

[Robin Haw]

I would definitely try the Reactome Functional Interaction (FI) Network plugin. It has been developed for analyzing cancer and other disease datasets.

[Dataminer]

Hi Robin! I have one more question, I have a set of TF and set of genes regulated by these TF's. How can I represent them as a network or what is the best wasy to do it? Any suggestions?

Answer 6

[favrin]

Hi

Try looking into http://www.esyn.org

It is a VERY easy network building, publishing and sharing tool - allows you to analyse your network - find what drugs interact with the proteins in your network and you can even import models from the biomodel database (https://www.ebi.ac.uk/biomodels-main/) and visualise them as Petri Nets

Hope it helps