The KDDArT Knowledge Discovery System

We will present the KDDart platform which is designed to integrate high volumes of data from several sources. At the moment the core module collects field data (plant performance), but can be easily modified to collect any kind of phenotypic data from any breeding or natural plant and animal populations. Both phenotypic and environmental modules are connected to a molecular data module enabled to store marker data from high volume, high resolution genome profiling technologies (e.g. DArTseq). In addition the system is capable of storing inventory and pedigree/relatedness information as well as supporting synchronisation of data captured through mobile devices with the central database. KDDart's layered design contains 3 major components: 1.Database server; 2. Application server API (Data Access Layer) and 3. Client applications /software.

The application server API is designed as a RESTfull web service, through which the database can be accessed. DAL is primarily a set of generic methods to perform all possible database operations. Client applications layer is a collection of scripts, software applets, devices, etc. accessing the database through the DAL. This includes integration of phenotypic data collected in the field with KDSmart application which will be also presented. Because DAL exposes itself as a web service, any modern programming language can be used to write software on top of DAL. The collected data can be than analysed statistically or data-mined using KDCompute, which represents KDDart’s “plug-in”. Processed data/information can be viewed and further analysed and interpreted by project staff and other system users with access privileges.