ConfeitoGUI: a tool for correlation network analysis

	ConfeitoGUI
a tool for correlation network analysis

★ Example 2

A dataset of Affymetrix Arabidopsis microarray experiments (the dataset is here.
Rows and columns represent probes and experiments, respectively.
9,442 experiments (a list of experiments is here).
22,746 probes (a list of probes is here).

The collection of genes co-expressed with a gene of interesst.
For that, the making of a correlation matrix between probes (genes).
The FPO analysis, followed by the FNI analysis.
A gene of interest here is “Myb28”, a transcription factor (probe name: 247549_at, element number: p002649, AGI code: AT5G61420).

Making a correlation matrix

Click launch.bat to open ConfeitoGUI.
Click Correlation Tool in the Init Tool tab to open the Correlation Tool window.
In the Input File section, click “Browse” and select “athgxmat.txt”.
In the Task section, select a type of correlation coefficient (select “Cosine”).
In the Output Folder section, select an output folder (make and select “cor”).
Push the Start button to execute the calculation.

Preprocessing for network analysis

Click Preprocessing Tool to open the Preprocessing Tool window.
In the Correlation Data Folder section, browse and select “cor”.
In the Output File section, browse this folder and name “athcor.flis”, and then push the select button.
Push the Start button to make the file “athcor.flis”.
Push “x” at the topright corner of the Preprocessing Tool window to close the window.
Click the Check Tool button to open the Check Tool window.
In the Correlation Data File section, browse and select “athcor.flis”.
Click the Start button to execute.

The FPO analysis

Click the Confeito Analysis tab and select Analysis Menu (select “FPO” in this example).
In the Correlation Data File section, browse and select “athcor.flis”.
In the Correlation Data Type section, select Cosine.
In the Output Folder section, make and select an output folder (make and select “fpo”).
Push the Start button to execute.

Preprocessing for the FNI analysis

Download and expand Perl programs for the preprocessing into a folder.
(“prefniprocess.pl” and “postfniprocess.pl” are made)
These tools are Perl scripts, and therefore, it is necessary that Perl commands are performable.
For a Windows user, install ActivePerl.
Input “perl -v” on the command prompt to check the performance of Perl commands.
For a Linux user, simply input “perl -v” on the terminal.
Move “prefniprocess.pl” into the “fpo” folder.
Open the Perl script by a text editor.
Input an element ID of interest within double quotations in the 5th line of the script (input “p002649” here).
Save and close the Perl script.
Doubleclick the file to execute (“analyzed_element_p002649.txt” is made).
Instead of using this Perl script, a user can make a list of element IDs in a column in a text file.

The FNI analysis

Click the Confeito Analysis tab and select Analysis Menu (select “FNI” in this example).
Correlation Data File: browse and select “athcor.flis”.
Correlation Data Type: automatically selected by the setting the “Correlation Data Type”.
Analyzed Elements File: select the file made in the previous section (select “analyzed_element_p002649.txt”)
FPO Analyzed File(*.fpo File): browse and select “result.fpo” in the “fpo” folder.
Push the Start button to execute (“result.net” is made).

Results

Open the output folder of the FNI analysis (“fni”).
Open the file “result.net” using a software for network analysis (such as Pajek).
Draw a network using the file.
Solid lines represents interconnections between elements detected by the FPO analysis.
Dashed lines represents interconnections from elements detected by the FNI analysis.
In the case of Example 2, a network module including Myb28, a transcription factor of interest, was obtained.
By the FNI analysis, Myb29 (AT5G07690, p005698), another transcription factor, was detected with the module.

Move “postfniprocess.pl”, downloaded in the preprocessing of the FNI analysis, with “prefniprocess.pl”.
Doubleclick the file to execute (“result.clu” is made).
Open “result.net” by Pajek and select “result.clu” in the Partition section.
Select “Draw” and then “Network + First Partition”.
The colors of vertices per pertition are changed by “Options” -> “Colors” -> “Partition Colors” -> “for Vertices”
In the module including Myb28, lighter and darker circles represent genes obtained by the FPO and FNI analyses, respectively.