Empty spaces in the dataset can be set to the average value in the dataloader.
Empty spaces in the dataset are displayed by a red "NULL" string.
Added a horizontal scrollbar in the zoomed dendrogram window
Empty spaces are marked by blue dots in the dendrogram (main and zoomed).
Problems in the SOM-visualisation with java 1.3 are fixed.
Ability to save gene identifiers and their expression values is added to
the "save" button in the gene graph.
CR characters are removed from the input data. These characters were visible
as squares and caused problems in the last data column.
Fixed the problem with empty spaces in the fist or last column in the
dataloader. Thanks to Philippe Marc for reporting this bug.
Fixed a bug in the plotting of the principal components in the PCA
module. Thanks to Martin Beibel for reporting this bug.
Installing and running
To run j-express, java runtime environment 1.2 or higher must be installed.
This can be downloaded from www.javasoft.com. The file jexpress.jar must
be unpacked either with jar, which is a standard java packing program and
a part of the java runtime environment, or with regular zip/winzip. When
this is done, it should be enough just standing in the directory with the
cluster.class file and type "java cluster". If the data we're analysing
is large, it can be necessary to use the -mx parameter to allocate more
memory to the application. For example if we want the program to be able
to use 100MB, we should run the program with "java -mx100M cluster". Remember
that when using hierarchical clustering we calculate a distance matrix.
If we analyse 6000 genes, we want to keep the distance between each gene
to all the other genes, requiring a triangular matrix of 6000*6000 elements.
If we allocate the memory to the matrix as integer values, the program
will demand: (((6000*6000)/2)*32 bits)/8 bytes = 72MB just for the distance
Unpack the jar file with: "jar -xvf jexress.jar" or "unzip jexpress.jar"
(The exe-file can be double-clicked from microsoft windows). Renaming the
jexpress.jar to jexpress.zip will make it a regular zip file.
Run the application with "java cluster" or "java -mx80MB cluster" (or another
amount of memory(type java -mx for details))
Start analysing. A testfile can be found in the dataset directory.
Running J-Express on a Mac:
From the MRJ SDK 2.2 folder, open the Tools folder
Then open the Application Builders folder.
Finally open the JBindery folder which contains the application JBindery.
Drag and drop the cluster.class file in the MRJ SDK 2.2 folder on top of
the JBindery icon.
Sollution to a common problem on macs: upgrading the Swing package:
This must be done when running J-Express results in this failure:
Go to the site http://www.ing.unitn.it/~luttero/javaonMac/
Find and download the swingall.jar file
unstuff it and drop the swingall.jar file into:
<:System Folder:>Extensions:MRJ Libraries:MRJ Classes (overwrite
if nessesary, but back up the old file)
Now try running jexpress again. It should work.
Send an email to firstname.lastname@example.org
if there is any trouble with this procedure, and I will try to help as
soon as I can.
you may also use this address to report bugs or contribute with new
See the help files for how to use the application
and how the methods work.