• Postgraduate Programme (MSc)
  • BECRO

 

A WEB-BASED DATABASE SYSTEM TO SUPPORT HLA STUDIES
DHLAS (Database HLA Analysis System) is a database system for the management and analysis of HLA data. DHLAS has been developed using JAVA (Sun Microsystems), runs on a Java Virtual Machine.

Its user- interface is web-based powered by the servlet engine TOMCAT (Apache Foundation). It utilizes STRUTS (Apache Foundation), a Model-View-Controller framework and uses several GNU packages to perform several of its tasks.

The database engine it relies upon for fast access is MySQL (MySQL Inc), but others can be used a well. 

  

The system easily calculates all indices required for HLA population studies, such as

  • Hardy-Weinberg equilibrium calculated using both asymptotic and exact tests,
  • Genetic distances (estimated using different calculation approaches for distances of vectors (Euclidian, Nei) - phylogenetic trees are constructed using UPMGA (Unweighted Pair Group Method with Averages) and Neigbor-Joining method,
  • Linkage disequilibrium (pairwise and overall, including variance estimations) and
  • Haplotype frequencies estimated using the expectation and maximization (E-M) algorithm.
  • Canonical Variate Analysis.

DHLAS uses the R System for it's statistical analysis module.

DHLAS records numerous epidemiological data like: name, address, year of birth, disease the subject has etc. Additionally it records family data that can be used in family based studies. Low and high resolution HLA data are recorded for every subject. Each subject is categorized according to its ethnicity (population) and the study he has taken part. Data entry is done in a user-friendly way. The user can select predefined values from lovs (list of values) minimizing errors due to typing

  HLA Data Entry

DHLAS Internal Stucture
(large image)

DHLAS is organized internally into several conceptual modules. This helps separating the different functionality into coherent logical units. Figure 1 shows this internal structure. Each rectangle represents a component or group of cohesive components that perform certain functionality. Arrows represent dependency relationships among these components. DHLAS-MIS and DHLAS-STATS are the two main components. The first is responsible for all database activity, while the second performs all the statistical functions.  


Figure 2, is a diagram in standard UML (Unified Modelling Language) notation. Each figure represents a Java class and the different arrows represent relationships among those classes. For the sake of clarity, the classes don't have all of their fields shown.

 DHLAS class diagram
(large image) 

 

To install DHLAS you will need to download the following:

  • The Java SDK.
  • A servlet engine like TOMCAT (Apache Foundation).
  • The R System. DHLAS has been developed using version 2.0.1 which you can download here.
  • The following R packages MASS, gap, DBI, RMySQL all available at CRAN. You will also need RGen available here. Alternatively you can download all libraries as a single file here.
  • A database like MySQL, Oracle, etc. DHLAS currently uses MySQL which you can download from here. If you are going to use a database other than MySQL, make sure that your version of the R System supports it. You can download support for MySQL here.
  • Java to R connectivity support through the RServe. You can also download it from here
  • The DHLAS specific war file.

Install the above in your system following the vendor's instructions. To install the DHLAS specific files, follow your servlet engine's procedure. If you are going to use TOMCAT, copy the war file into the following directory tomcat istallation dir/webapps. THAT'S IT!

To install the database tables, download and run the following script. You will also need to locate hibernate.cfg.xml (under DHLAS war installation directory) and change the user name and password properties to your own. If you are using MySQL, a preconfigured database can be downloaded from here. In case your R support for MySQL uses a client version older than 4.1, use the OLD_PASSWORD command to convert to the old style passwords. For more details see this. If you are using a database other than MySQL you will need to install the appropriate jdbc driver. For more information and more advanced settings consult the Hibernate manual (pdf html).

To install R packages, open R and choose "Packages->Install package(s) from local zip files" and locate the library zip(s) file.

Please modify the genetics.properties file located under the DHLAS war installation directory, to point to your R installation directory

To install RServe consult its documentation page.

DON'T FORGET TO MODIFY hibernate.cfg.xml and genetics.properties settings corresponding to your username and password values. Both files can be found under your war installation.

Typical Installation

Our team has implemented an installation program, to help install DHLAS on your machine. It currently supports Windows© based installations. The installation program assumes that you hava already installed the Java SDK, the Tomcat Servlet engine, MySQL 4.1 and R System. You can download the installation program here. To start using it, type java -jar dhlas.jar.

If you have any problems with your installation, please e-mail to Paschalis Thriskos.