Prerequisite Software

To install the toolkit locally on your machine you will need the following software.

• Java Development Kit, the software was tested with jdk1.6.0_05.
• Tomcat Apache used as a container for the Data Access Web Application, the toolkit was tested with Tomcat 6.0. In the following we refer to the directory in which Tomcat is deployed by TOMCAT_HOME.
• The eXist database system for storing captured experiments, the toolkit was tested with eXist 1.2.0.
• The PEDRo data capture tool. In the following we use PEDRo_HOME to refer to the directory in which PEDRo is deployed.

Installing the Toolkit

The following presents the steps for installing the toolkit.

• Download the FuGEToolkit.zip file from the download web page.
• Unzip the file into a directory of your choice. You will find the following files as a result: proteomics.war, namespacemodifier.jar and a directory named PedroModels .
• Deploy the PEDRo models by putting the sub-directories of the PedroModels directory into PEDRo_HOME/models. The PedroModels directory contains two models: GelML and spML.
• Deploy the data access web application, proteomics.war, in the Tomcat container. This can be done by moving the proteomics.war file into the directory TOMCAT_HOME/webapps. The Tomcat container will automatically deploy the web application.
• Before you can use the data access web application, you will need to edit the file TOMCAT_HOME/webapps/proteomics/WEB-INF/web.xml to set information necessary to access the eXist database. Specifically the values of the following parameters may need to be edited according to your settings: dbDriver, the driver used for accessing the eXist database, you will not need to change the value of this parameter unless you opt for an XML database server other than eXist; dbServerName, this is the URL of the eXist server, e.g., exist://localhost:8680/exist/xmlrpc/; dbCollectionName the name of the collection that will be accessed, e.g., proteomics; username and password are respectively the user name and the password needed for accessing the collection at the eXist database server.

Capturing and Sharing GelML data

To capture GelML data you can use PEDRo data capture tool. For more information about the use of the PEDRo tool, users are referred to the PEDRo web site. It is worth noting that the Pedro models for GelML and spML come with a plug-in for facilitating the capture of references within experiments. Indeed, there may be several cross-references between the elements of a GelML or spML experiments, e.g., the attribute Software_ref of the element Provider has as a value the identifier of a Software element. Manually specifying these references can be time consuming: the user will have to fetch the referenced elements within the experiment. The plug-in we developed aims to facilitate this task; it automatically fetches the identifiers of the referenced elements within the GelML experiment and displays them as a drop down list. The user may then choose an item from the list to be used as a value. For example, Figure 1 illustrates the drop down list displayed by the plugin and which contains the items that can be used as a value for the Software_ref attribute of the Provider element.

Figure 1: PEDRo plug-in for editing cross references in a GelML experiment.

When using Pedro, it is possible to have an exception stating that the plugin classes were compiled with Java 5. To fix this, you can edit the Pedro batch files to use a locally installed Java of a sensible version (5 or 6).

Once you finished editing a GelML experiment you can export it as an XML document (A documentation on the use of the eXist database can be found here). Exported XML documents can be then stored within the eXist database. Before doing so, users will need to use the tool namespacemodifier.jar that come with the toolkit. This tool is used to convert the namespace of certain elements in the XML document exported by PEDRo to make it a valid FuGE XML document. The following command line shows how the tool can be used.

java -jar namespacemodifier.jar fuge.xsd experiment.xml fugeexperiment.xml

fuge.xsd is the XML Schema of the FuGE model, experiment.xml is the XML document exported using the PEDRo tool and the fugeexperiment is the XML Document output by this command and which is a FuGE compliant experiment.

Using the Data Access Web Application, the user can browse the experiments s/he captures. If your Tomcat is deployed locally on your machine on the port "8080", then you can visit the web page for browsing GelML experiments at http://localhost:8080/proteomics/index.jsp?application=Fuge%20Proteome%20Database&directory=Browse%20Gel%20Data&file=index.jsp. This web page contains a set of pre-defined canned queries (see Figure 2).

Figure 2: Canned queries provided for browsing GelML Experiments.

Using these queries users can retrieve: (i) general information regarding the conditions under which gel experiments were performed; (ii) information about gel locations (a.k.a. spots or bands), e.g., the user may request information about the gel locations that were detected in a given gel; and (iii) information about gels, e.g., Figure 3 shows the results returned when requesting information about available gels.

Figure 3: Screenshot of the web page used for browsing GelML Experiments.

Using the toolkit to capture and share data in your own FuGE extension (for advanced users)

You can use the toolkit for capturing and browsing experiments that are specified in your own FuGE extension, just like we did with GelML. For this, you will first need to specify the XML Schema of your FuGE extension. Using this XML Schema you can create a PEDRo model as described here. You then need to specify XQuery queries for browsing the experiments specified according to that model. The obtained XQuery are to be embed in JSP files. Examples of JSP files that include XQuery queries are supplied for GelML at TOMCAT_HOME/webapps/proteomics/Fuge Proteome Database/Browse Gel Data. You can use these files as templates for specifying you own canned queries.

If you have a query regarding the use of the toolkit then please feel free to contact Khalid Belhajjame.