December 11, 2017, Monday

From BIMIB

Jump to: navigation, search

CABERNET TUTORIAL

Contents

CABeRNET: Cytoscape app for Augmented Boolean models og gene Regulatory NETworks.

Welcome to the CABeRNET tutorial page.

The current version of CABeRNET has been designed for working only with Cytoscape 3.x.

How to install

  1. Download the file at this link or from the Cytoscape App Store.
  2. Open the Apps —> App Manager dialog.
  3. Import the CABERNET-1.0.jar file using the “Install from file” button in the “Install Apps” tab.
  4. You will find the item CABERNET in the Apps menu.
  5. Run the App.

Running the case study: T-Helper signaling network

In order to reproduce the case study the user has to:

  1. Open the CABeRNET wizard dialog from Apps -> CABERNET -> Wizard.
  2. Select the option “Augment the topology of…” and load the original T-Helper network from this grnml file: TCR file. Press the button Next.
  3. The next required parameter is the number of networks to obtain at the end of the simulation.
  4. You will be asked to define the structural features of the network. Choose the option “From input form” and fill the form with the case study parameters: 200 nodes, 400 edges 1.0 random functions and 0.0 for the other function types. Leave all the other parameters at their default values.
  5. Insert the GRN’s dynamics simulation parameter (in the case study partial sampling with 10000 initial conditions and flip perturbation with 200 randomly selected single/multiple node perturbations for each attractor state have been selected). Leave all the other parameters at their default values.
  6. Import the T-Helper lineage tree from the Cytoscape's view (sif file) or from file (Hemapoietic tree) and select the similarity function (in the case study perfect match has been selected). Note that the the maximum network to test value is mandatory. This parameter specifies a cutoff on the number of networks to test.
  7. Select the required outputs.

All the functions, like ATM and TES views, dynamical statistics, can be reached from the Apps menu. A particular TES network can be creates simply by defining the desired threshold in the ATM view. There are two different types of TES networks, for a detailed comparison check the user manual.

Once the simulation has been completed, it is possible to create the network, attractors and TES views and, after, set the appropriated style simply by selecting the desired one in the "style panel". In the following table all the CABERNET styles are grouped.

CABERNET Network The color of each node being related to the Boolean function bias and the size of each node proportional to its degree.
CABERNET Attractors For the visualization of the attractor graph network. Each state of every attractor is represented as a yellow square. 
CABERNET TES For the visualization of the ATN. There are two types of edges: one for the connecting two states in the same attractor and one for connecting two attractors. In the second type, the edge size is proportional to the transition probability.
CABERNET collapsed TES For the visualization of the ATN. Each attractor is represented as a unique diamond and the the edge size is proportional to the transition probability.


In order to reproduce the second part of the case study, i.e. single knockout experiments (KO), it is necessary to change manually each function in the grnml network file (matching network).

For example, in order to knock out the CD45 gene, the following changes must be performed.

Original:

<node id = "0" name = "CD45">   <function type = "random" input_number = "3">   <bias> 0.5 </bias>   <input_node >0</input_node>   <input_node >125</input_node>   <input_node >130</input_node>   <entry input = "110" output = "0"></entry>   <entry input = "011" output = "0"></entry>   <entry input = "000" output = "1"></entry>   <entry input = "111" output = "1"></entry>   <entry input = "100" output = "0"></entry>   <entry input = "001" output = "0"></entry>   <entry input = "101" output = "1"></entry>   <entry input = "010" output = "1"></entry>   </function>   </node>

Knock Out:

<node id = "0" name = "CD45">   <function type = "random" input_number = "3">   <bias> 0.0 </bias>   <input_node >0</input_node>   <input_node >125</input_node>   <input_node >130</input_node>   <entry input = "110" output = "0"></entry>   <entry input = "011" output = "0"></entry>   <entry input = "000" output = "0"></entry>   <entry input = "111" output = "0"></entry>   <entry input = "100" output = "0"></entry>   <entry input = "001" output = "0"></entry>   <entry input = "101" output = "0"></entry>   <entry input = "010" output = "0"></entry>   </function>   </node>

The full CD45_KO network can be downloaded here.

Once the GRNML file has been modified, the simulation can be started from the wizard menu as follows:

  1. Import the modified network and select "Generate networks completely defined via *.grnml file(s).
  2. Insert the GRN’s dynamics simulation parameter as defined before.
  3. Import the T-Helper lineage tree, select the desired similarty function and the maximum number of network to test (in the case study the histogram distance and only 1 network to test have been selected).
  4. Select the outputs.

This process has been repeated for all the 40 genes in the T-Helper signaling network and the obtained similarity values have been collected. 

Note: All the parameters that are not defined above have been left with the default value.

Documentation

For further information on using CABeRNET please check the complete manual.