MCODE User's Manual

TableOfContents()

Installation

To use the MCODE PlugIn, you must first obtain a copy of Cytoscape. The compatible MCODE and Cytoscape versions are outlined in the downloads section on the [:Software/MCODE: MCODE website]. The lastest MCODE, version 1.2, requires Cytoscape 2.3.2 or later.

• You can download a copy of Cytoscape from: http://www.cytsoscape.org.

Once you have downloaded Cytoscape and verified that it works, proceed with the next steps:

1. Copy the MCODE.jar file to your [Cytoscape_Home]/plugins directory.
2. Start Cytoscape. For example
   • On Unix/Linux or MacOS X, run: cytoscape.sh
   • On Windows, run: cytoscape.bat
3. Check that MCODE appears in the Plugins menu of Cytoscape
   • If it does not, then you likely placed the MCODE.jar file in the wrong directory. Repeat step one. You will have to restart Cytoscape to reload the plugin.

Running MCODE

MCODE is an extension for Cytoscape and can only be accessed through Cytoscape.

1. Start Cytoscape
2. Go to the Plugins Menu
3. Move the mouse over MCODE

4. Click Start MCODE

   • The main MCODE interface will appear as a tab in the left-hand panel of Cytoscape

Note: If MCODE does not appear in the Plugins Menu, then the installation of MCODE was not successfull. Please refer to the previous section of the User's Manual.

This menu provides two additional options:

• About MCODE

  • A quick reference to MCODE acknowledgements, citation information, and a link to the MCODE paper.

• Help

  • A link that will open the MCODE website at [:Home: www.baderlab.org] in your default browser for quick access to downloads, contact information, and the User's Manual with help and tutorials.

New to Version 1.2: MCODE can now be seen as a console for the underlying algorithm. It can be started independent of network loading. This provides for a more easily accessible, repeatable and modifyable analysis.

Getting Results

The MCODE Main Panel is the starting point for analysis. It contains two main sections: Scope and Advanced Options. The latter is intended for fine-tuning of results by experienced users who are familiar with the MCODE Paper. It will be discussed in the next secton. This section will cover some of the basic steps of running MCODE on a network.

1. LOAD YOUR NETWORK. To begin, make sure the network to be analyzed is loaded into Cytoscape. You can load as many networks as your computer system can handle, large or small. MCODE will recognize which network you wish to analyze either by noting which network view is on top or by your selection of the network in the Network Tab on the left-hand panel.

2. CHOOSE THE SCOPE. Cluster results can be reported in two fundamental ways with MCODE. This is referred to as the Scope of the process.

   1. Find Clusters in Whole Network
      • MCODE will find and report all clusters in the entire network.
   2. Find Clusters from Selection
      • Only those clusters including the selected node(s) as one of their members will be reported.
      • Selections can be made either in the view directly or Cytoscape's handy search tool.
   3. The choice of scope is dependent on the your familiarity with the network in question and the desired outcome. Having a particular protein of interest within a network, for example, it may be appropriate to search for only those clusters involving such a protein. On the other hand, uniformed, exploratory work will most benefit from a whole network scope.

3. ANALYZE YOUR NETWORK. Next, press Analyze. This will display a task monitor reporing the progress of the scoring, finding, and drawing steps, provided that the task is not too quick.

   • You may see several different messages at this step:
     1. The "No network" message
        • This means that MCODE failed to detect a loaded network for analysis. You must load a network, and make sure it is selected, before you can anaylze it.
     2. The "No selection" message
        • This message appears when the Selection Scope is used without an actual selection. You must select the desired node(s) before you MCODE can attempt to find clusters.
     3. The "Parameters unchanged" message
        • Parameters are discussed in detail in the following section. For now, you should know that if you attempt to analyze a network twice without changing any of the settings, such as scope, MCODE will let you know that this analysis was previously conducted and will consequently display the previously attained results.
4. BROWSE YOUR RESULTS. If everything goes to plan, a new tab will appear in the right-hand panel displaying the results as "Results 1"

   • Cluster Browser:

     • On the left side is a graphical representation of the cluster.
       • Cluster members are coloured red.

       • The highest scoring node in the cluster is called the Seed. It is the node from which the cluster was derived and is represented by a square shape.

       • Other cluster members are circular.
       • Edges, representing interactions for example, are blue.
       • Edge directionality is represented by cyan arrows.
     • On the right side is a statisical summary of the cluster.

       • Rank is based on the cluster's computed Score and is used to ID the clusters within each result set.

         • For example, Cluster 1 is the highest ranked cluster in a given result set, and thus, at the top of the list.

       • Nodes and Edges is a simple enumeration of the clusters members and their interconnections.

     • These results can be discarded at any time by pressing the Close button at the bottom of the results panel.

   • Network View:

     • If the network being analyzed has a view, MCODE will apply a custom visual style utilizing two MCODE generated node attributes.

       • MCODE_Node_Status: Node shapes indicate the cluster status of the nodes.

         • Square: seed (highest score in the cluster)
         • Circle: clustered
         • Diamond: unclustered

       • MCODE_Score: Node colors represent the node score.

         • A range from black to red indicates the MCODE computed node score (lowest to highest, respectively).
         • White indicates a score of zero.

       • MCODE_Cluster: This is an additional list type attribute that indicates which cluster the node belongs to. The MCODE visual style does not make use of it, but it is there in case it may be of some use. Note that if the Fluff parameter (discussed in the following section) is turned on, some nodes may belong to more than one cluster.

   • Cluster Selection:

     • The clusters in the cluster browser table are selectable and will automatically select the corresponding nodes in the network view (if it exists). If no network view is available, the selected nodes can be reviewed in the Cytoscape native node attribute browser.

     • Secondly, a new Cluster Exploration Panel will appear below the Cluster Browser. This panel can be collapsed for now -- it's use will be discussed in the Exploring Results section of this Manual.

This is a screeen shot of the MCODE Main Panel and Cluster Browser

• attachment:main_panel.gif attachment:cluster_browser.gif

Fine-Tuning Results

By default, MCODE analyzes networks using scoring and finding parameters that have been optimized to produce the best results for the average user and network. However, you may benefit greatly by familiarizing yourself with these parameters. Sometimes even slight customizations can produce considerable differences, reduce unwanted or false results, and increase relevance to your network. This is only an overview -- for a detailed insight into how these parameters function, it is best to consult the MCODE paper.

Scoring Parameters

1. Include Loops

   • If checked, MCODE will include loops (self-edges) in the neighbourhood density calculation. This is expected to make a small difference in the results.

2. Degree Cutoff

   • Sets the degree cutoff below which a node will not be scored. For example, nodes that share only one connection with one other node have a degree of 1. It is generally undesirable to include these in clusters and as such, the degree cutoff minimum value is 2. Increasing this parameter will filter out less interconnected nodes and will cause MCODE to return only the more interconnected clusters.

"Sets the degree cutoff below which a node will not be scored.\n" +

• "Nodes with a degree equal or higher to this value will be scored.\n" + "By default this is set to 2. Valid values are 2 or higher to prevent singly connected nodes\n" + "from getting an artificially high node score.";

Finding Parameters

1. Haircut

2. Fluff

   • Node Density Cutoff

"Sets the node score cutoff for expanding a cluster as a percentage from the seed node score.\n" +

• "This is the most important parameter to control the size of MCODE clusters,\n" + "with smaller values creating smaller clusters.";

1. Node Score Cutoff

2. K-Core

3. Max. Depth

New to Version 1.2: The user can now analyze a network as many times as desired by modifying the parameters. Each result set is stored sequentially for reference and comparison. Viewing different result sets will automatically rewrite the MCODE node attributes and revisualize the network. Note that MCODE can independently determine which portion of the algorithm needs to be conducted based on the user's parameter modifications. If the scoring parameters are altered, the given network will be rescored. If only the cluster finding parameters are altered, only the cluster finding portion will be conducted.

Exploring Results

Size Slider

Node Attribute Enumerator

Outputting Results

Tutorials (Coming Soon)

BiNGO Validation