|
Network Generation
A sample .gnl.txt file.
Three network growth models are available in GNLab. A random network
can be generated by the Erdos-Renyi model (-r), the Scale-free model (-f), or the
Charleston-Ho model (-g). The Charleston-Ho model is a newly proposed network growth model that is
based on well-know processes in genome evolution. This model is shown to capture the detailed topological
structure of the real GRNs (Ho and Charleston, in preparation).
The generated network is stored in a file with extension ".gnl.txt".
The file format contains information about the node, edges and edge weight.
Network Visualization
Visualization of the yeast GRN using GraphViz (left) and GEOMI (right).
GNLab can produce input files for GraphViz, GEOMI and Cytoscape. This functionality can be
invoked by the command-line option -v. For example, the command "GNLab -v yeast dot none direct"
would produce a file called "yeast.dot" from "yeast.gnl.txt".
Network Simulation
A sample .ma.txt file
A heatmap of a simulated microarray dataset (left) and a time-series
gene expression profile (right).
Using the Hill's kinetics, the gene expression pattern of a GRN can be simulated either
deterministically or stochastically. Data for the time-series gene expression profile can be generated
by invoking command-line option -t. The simulated data is stored in a text file with a ".data" extension.
Three types of microarray datasets can be simulated by GNLab:
time-series, gene perturbation, and condition-specific datasets. The command-line option -s is used to
invoked a simulation of microarray data. The microarray dataset is stored in a ".ma.txt" file.
Network Analysis
A sample .ana.txt file.
The static structure of a GRN can be quantified by a collection of network
topological features. A set of 11 topological features is calculated in GNLab. The distribution of
node in-, out- and total-degree, as well as other distributions can be obtained in a .ana.txt file.
The .ana.txt file can be obtained by analyzing a network with GNLab with the command-line option -a or
-n.
Network Comparison
The topological similarity between two networks with the same number of nodes can be calculated in GNLab.
A network comparison is invoked through the command-line option -c. A one-line summary of topological
differences between the two networks is outputed to the console.
Network Inference
GNLab does not perform network inference directly. However, it allows the microarray dataset it generates to be
converted into the input format of ARACNe and Banjo. This process is invoked by the command-line option -d.
|
