Guiyuan Lei - iNotes

The use of Cytoscape:

Cytoscape is an open source bioinformatics software platform for visualizing molecular interaction networks and integrating these interactions with gene expression profiles and other state data.

• Install cytoscape

Before install Cytoscape, need to install Java 1.4.2 first. Download the self-extracting file like j2sdk-1_4_2_12-linux-i586.bin for Linux and run this command, java 1.4.2 will be installed in your linux. Then set the directory of this 1.4.2 version java (use "which java" will show the directory of your java) in cytoscape.sh file, run cytoscape.sh, Cytoscape is launched. "Load" network file from "File" menu, "creat view" (click right mouse button when hightlight the network) for large network. Select "Layout" style, usually I choose "yFile->Circular", it is good for viewing sub-network.

• Set Visualization Style

The simple interactionn format is .sif, each line specify a source node, a relationship type (or edge type), and one or more target nodes. If you need a undirect graph, use "pp" as edge type.

There are node attribute (.na) and edge attribute (.eda) files which can be mapped to node label or edge label by setting visulaization sytle from "Visualization" menu. For example, if you use gene IDs in your .sif file, but you'd like the node label be shown as gene symbols (be mind that usually a couple of gene IDs are corresponding to one same gene symbol, use gene IDs in .sif file and use gene Symbols in .sif file will generate different views of same network), you can map one .na file to node label. Each line in the .na file specify which gene symbol is corresponding to which gene ID, like "1553402_a_at = HFE". Don't forget the first line of .na file is the name of the attribute, like "Official Gene Symbols". Steps for visualizatio style to map gene IDs to gene symbols

After load the network file, load corresponding node attribute file.
Go to menu "Visualization", "Set Visual Style", choose "default" sytle, click "Duplicate" button, name the visual style.
Click "Define" button, in the new open window, select "Node Label", select "BasicPassThrough" for "Mapping" method, select "Official" for "Map Attribute". Click "Apply to Network" and "Close" to finish the setting. Important: .na should be loaded so that you can set your own node label attribute, the "Official" is the name of attribute in the first line of your .na file.
You can flip through different visual sytle from "Set Visual Style" menu or from the Visual Style pull down menu.

An alternative to see node attribute is through "Attribute Browser", you need not pass through gene symbols to all gene IDs. Select "Node Attribute Brower" (at the bottown of the network), click "Select Attributes", select "Official". Then when you select nodes, the Official gene symbols of these genes are shown in node attribute browser (use setting visual sytle, all node labels are gene symbols instead of gene IDs).

Setting edge attribute is also useful. For example, if you need to have an idea of how strength of the interaction between two genes, you can set visual style to show the interaction strength.

After load the network file, load edge attribute (.eda) file. The first line of the file is the name of the attribute, like "InteractionStrength".
Go to menu "Visualization", "Set Visual Style", choose "default" sytle, click "Duplicate" button, name the visual style.
Click "Define" button, in the new open window, select "Edge Label", select "TestPassThrough" for "Mapping" method, select "InteractionStrength" for "Map Attribute". Click "Apply to Network" and "Close" to finish the setting. Important: .eda should be loaded so that you can set your own edge label attribute, the "InteractionStrength" is the name of attribute in the first line of your .eda file. You can also select visual style for edge color to view the interaction strength.

• Extract sub-network

The ealiest way to extract sub-network is to select nodes by name from "Select" menu, then select first neighbours of selected nodes, do "select first neighbours of selected nodes" a couple of times, finally "To NewNetwork" for "Selecte nodes, All edges".

• Plugin: Agilent Literature Search

The Agilent Literature Search plugin in Cytoscape provides a flexible, interactive platform for mining text, and assessing the results in a network context. You can also very your constructed network by gathering evidence from literatures:

Set java environment. In tcsh shell (put the following command in .cshrc):
```
setenv PATH /usr/local/java/j2re1.4.2/bin:$PATH
```
(put the added path in the beginning of old path) or
```
setenv PATH ${PATH}:/usr/local/java/j2re1.4.2/bin
```
In bash shell
```
export PATH=/home/ngl9/j2re1.4.2_12/bin:$PATH
```
Then run install.bin (for linux), install the directory of plugin under your installed cytoscaple directory.
In the "Plugins" menu of Cytoscaple, click the "Agilent Literature Search" to lanuch the application. Set http proxy if necessary. In math department, host:wwwcache.ncl.ac.uk, port:8080. In gerontology building, host:pac.ncl.ac.uk, port:8080.
Search term by giving gene symbols, combile context like "cancer". Can search multiple terms, also selece different engines like PubMed. Articles related to the query will be shown in the window of "Query Match", there are also links to these articles. A network (if there is network) will be shown in Cytoscape interface. Click the edge in the network will show "Sentences" (from these articles) related this interaction. There are some "patameters" for search, like the "Max Engine Matches", "Use Aliases", "use Context","Organism", "Interaction Lexicon", etc.
Varify network, gather evidence from literatures. Right click the edge, click "Gather evidence from literatures", after search, if there are articles which support this interaction, you can have a look at these articutes by click "Show evidence from literatures".

• Other Features

There are many other features for Cytoscape, like "Filters & Editor", "Fetching External Data", "Expression Analysis", "Gene Onotology" and "Modules and Complexes". It's very worth reading Cytoscape Online Tutorial

Last modified:
27 June, 2006